The Retoba
AI Glossary

What is search marketing?

Search marketing is the practice of creating visibility and demand when people actively look for information, products, services or solutions. It includes organic search optimisation and paid search advertising, but its strategic role is broader: connecting a recognisable need with a credible next step.

Why it matters
Answers generated by AI do not eliminate search marketing. They change the surface on which discovery happens. Strong technical SEO, authoritative content and clear brand entities remain the foundation from which answer engines retrieve information.

What is search intent?

Search intent is the goal behind a query: the outcome a person is trying to reach rather than the literal words they type. Intent may be informational, comparative, navigational, transactional or local, and a single query can contain more than one of these needs.

Why it matters
A page that merely repeats a keyword can miss the actual decision. Conversational systems make intent more visible because they can ask additional questions and distinguish a casual enquiry from a complex business need.

What is marketing based on keywords?

Marketing based on keywords organises campaigns and content around the phrases people use in search. Keywords remain valuable signals of language, demand and competition, but they are an imperfect proxy for context: two people can use the same phrase while expecting very different answers.

Why it matters
Modern discovery requires a shift from exact phrase matching to topic depth, entity clarity and intent coverage. Keywords inform the system; they should not dictate unnatural copy or fragment one useful subject into dozens of thin pages.

What is AI search?

AI search is a discovery experience in which generative models retrieve, interpret and synthesise information into a contextual response. Instead of presenting only a ranked list of links, it may explain, compare, recommend and support further questions, while citing selected sources.

Why it matters
Visibility is no longer measured only by ranking position and clicks. Brands must also understand whether their information is retrieved, represented accurately and cited across Google AI experiences, Microsoft Copilot, ChatGPT, Perplexity and other answer systems.

What is the shift from search to conversation?

The shift from search to conversation is the movement from isolated queries and static result pages towards iterative dialogue that retains context. A person can refine the problem, add constraints, challenge an answer and continue without restating the entire situation.

Why it matters
The brand experience increasingly begins before a website visit and continues after the click. Organisations must therefore design discoverable knowledge and a useful conversational layer, rather than treating search, content and service as separate disciplines.

How Retoba applies it
Retoba designs connected journeys from query to dialogue in which discovery leads into an active, useful conversation instead of ending with a click.

What is conversational search?

Conversational search is the process of finding information through a sequence of questions and refinements in natural language. The system uses context from earlier turns to interpret pronouns, preferences, exclusions and changing priorities.

Why it matters
Content must answer not only the first broad question but the related questions that arise during evaluation. Clear definitions, comparisons, evidence and entity relationships make it easier for search and AI systems to retrieve the right passage for each step.

What is conversational discovery?

Conversational discovery occurs when dialogue helps a person recognise options, needs or possibilities they had not fully articulated at the start. It combines exploration with explanation: the interface narrows uncertainty while introducing relevant ideas at the right moment.

Why it matters
This is especially valuable for complex products and services where customers do not know the correct terminology. A carefully designed AI character can translate expertise into an accessible conversation without forcing the user through a rigid menu.

What is the Query to Dialogue Journey?

The Query to Dialogue Journey is a Retoba framework for designing the progression from an initial search or question to a useful brand conversation and a decision based on better information. It connects discoverability, answer readiness, conversational experience, insight capture and an appropriate business action.

Why it matters
The framework prevents teams from optimising isolated touchpoints. It asks whether a person can move coherently from “I need an answer” to “this brand understands my situation” while preserving transparency, consent and the option to reach a human.

What is conversational marketing?

Conversational marketing is a strategy that uses dialogue to understand people, provide relevant value and move them towards an appropriate next step. It can take place through AI interfaces, messaging, live chat, voice or a combination of human and automated touchpoints.

Why it matters
Traditional campaigns often infer intent from a click. A conversation can ask directly, respond to nuance and reveal the language people naturally use. It moves marketing from informing to dialogue and, when the brand genuinely helps someone decide, from advertising to advising. The result can be a more useful experience for the visitor and a richer source of insight for the organisation.

How Retoba applies it
Retoba combines a conversational experience designed for the brand with PulseDialog AI conversation intelligence, so dialogue supports both the visitor’s next step and the organisation’s next decision.

What is B2B conversational marketing?

B2B conversational marketing uses dialogue to help professional buyers investigate a complex need, evaluate options and coordinate an appropriate next step across a longer decision process. It can support several stakeholders, including business leaders, users, technical specialists, procurement and legal teams, each of whom may need different evidence.

Why it matters
B2B intent is rarely captured by one form, one visit or one decision maker. A useful conversational system can clarify the role, use case, constraints and stage of the buying process, answer technical and commercial questions, surface relevant cases and route the established context to the right specialist. Its purpose is to improve understanding and coordination, not to disguise automated qualification as personal advice.

What is a conversational customer journey?

A conversational customer journey is a sequence of connected interactions in which the experience adapts to what a person says, needs and has already established. Instead of forcing everyone through the same funnel, dialogue can support discovery, evaluation, purchase, onboarding, use and service.

Why it matters
The journey becomes less about moving users between pages and more about preserving useful context. Good design also defines where context should stop, when consent is required and when a human should take over.

How Retoba applies it
The Peugeot 2008 AI Ambassador case shows one conversational interface supporting discovery, purchase decisions and practical questions after the purchase.

What is a post-click conversation?

A post-click conversation begins when a visitor arrives from an advertisement, search result, social post, email or AI citation and can immediately discuss the need that prompted the visit. Campaign context can shape the opening without pretending to know more about the person than the available data supports.

Why it matters
Landing pages often lose relevance after the headline. A conversational layer can continue the promise of the source message, answer objections and direct a qualified visitor to content, a tool, a demo or a person.

What is a conversational landing experience?

A conversational landing experience combines the clarity of a strong landing page with an interface that lets visitors ask, explore and receive a relevant response. The page still needs a clear proposition, evidence and accessible navigation; conversation is an additional path, not a replacement for usable content.

Why it matters
Some visitors want to scan. Others need to explain their situation. Supporting both behaviours reduces friction and gives the brand a chance to answer questions that a fixed page cannot anticipate economically.

What is a customer journey shaped by dialogue?

A customer journey shaped by dialogue uses the logic of conversation to organise an experience across touchpoints: question, response, interpretation and the next best step. It may include AI and human interactions, but its defining feature is continuity of meaning rather than a particular channel.

Why it matters
This approach exposes gaps between marketing, sales and support. If a brand asks for information but cannot use it to improve the next interaction, the journey is collecting data rather than sustaining a dialogue.

What is conversational conversion?

Conversational conversion is the completion of a meaningful next step that was enabled or improved by dialogue. It can be a purchase or lead, but also a booked consultation, product match, completed onboarding task, resolved support issue or informed handover.

Why it matters
Counting conversations alone rewards volume, not value. Measurement should connect the exchange to an outcome while also monitoring answer quality, user effort, satisfaction, escalation and commercial impact.

What is Growth through Conversation?

Growth through Conversation is a Retoba approach in which dialogue improves both customer experience and organisational learning. Conversations create value in the moment. When analysed lawfully and responsibly, they reveal recurring needs, objections, language and opportunities that can improve products, services, content and operations.

Why it matters
The compounding effect is more important than a single chatbot metric. A brand listens at scale, acts on what it learns and makes the next generation of conversations more useful.

What is a proactive conversation?

A proactive conversation is initiated by the brand when context suggests that timely help would be useful. Examples include offering guidance on a complex page, explaining an unfamiliar term or checking whether a customer needs assistance after repeated failed actions.

Why it matters
Proactivity can reduce effort, but poor timing feels intrusive. The trigger, wording, frequency and dismissal behaviour should respect the user’s attention and avoid manipulative pressure.

What is conversational commerce?

Conversational commerce uses dialogue to support product discovery, comparison, recommendation, purchase and service after the sale. The interface can translate preferences into criteria, explain advantages and limitations, and connect to inventory, pricing or transaction systems when those integrations are reliable and authorised.

Why it matters
The objective is not to push a product faster. It is to help a person make a suitable choice with less uncertainty. Recommendations must be grounded in current data and disclose material limitations or commercial influence.

What is conversational branding?

Conversational branding is the deliberate expression of a brand’s identity through interactive behaviour. It translates positioning, values, language and service principles into the way an AI or human interface asks questions, explains choices, handles emotion and guides action.

Why it matters
A static brand voice document cannot anticipate every exchange. Conversational branding defines principles and boundaries that remain recognisable across thousands of situations without forcing every answer into the same formula.

How Retoba applies it
Retoba turns brand strategy into a working system of knowledge, voice, character, boundaries and evaluation. See the Kotányi AI Chef approach.

What is an AI Ambassador?

An AI Ambassador is a branded conversational AI representative designed to express a defined role, expertise, voice and behaviour across customer interactions. It helps people discover, understand and use a brand’s products or services through natural dialogue, while remaining transparent that it is an AI system.

An AI Ambassador is more than a generic chatbot. It combines governed knowledge with conversational branding, character design, contextual guidance, defined boundaries and human handover. Depending on its purpose, it can support discovery, comparison, education, service and questions after purchase, while aggregated conversations can reveal recurring needs and opportunities for the organisation.

Why it matters
The ambassador becomes an active expression of the brand, so accuracy and personality must be designed together. The same principle must serve different industries without reducing every experience to the same character or script.

How Retoba applies it
Compare Retoba’s ambassadors designed for different roles: Peugeot 2008, Kotányi and GYM24.

What does an AI Communication Architect do?

An AI Communication Architect is an emerging specialist who designs how an organisation’s knowledge, identity and communication are represented across answers generated by AI and conversational systems. The role connects content strategy, brand voice, information architecture, structured data, prompt systems, governance and evaluation.

Why it matters
AI representation does not belong to one department. Reliable answers require communication, marketing, product, data, technology, legal and service teams to work from a coherent architecture. The title is not yet a universal professional standard, but it describes a capability organisations increasingly need.

What is brand voice in AI?

Brand voice in AI is the operational specification that governs how an AI system communicates on behalf of a brand. It covers vocabulary, rhythm, formality, humour, explanation style, sensitive situations, prohibited claims, uncertainty and escalation. It is more than a list of adjectives such as “friendly” or “bold”.

Why it matters
A generic model can produce fluent answers that sound unlike the organisation or create risk. A usable voice system turns identity into testable behaviours that content, design, technology, legal and service teams can evaluate together.

How Retoba applies it
Retoba translates brand voice into tested prompt rules, examples, boundaries and response patterns. The Kotányi AI Chef shows how proprietary knowledge and conversational character work together.

What is AI character design?

AI character design is the creation of a coherent role, personality and expressive system for an AI interface. It defines who the character is in relation to the user, what expertise it represents, how it speaks, what it never pretends to be and how visual, verbal and behavioural cues work together.

Why it matters
Cooking inspiration, automotive guidance and fitness motivation create different expectations. Character follows purpose; it should never be cosmetic theatre.

How Retoba applies it
Retoba designs each character around a specific job and context. Compare Kotányi AI Chef, the Peugeot AI Ambassador and the GYM24 AI Trainer.

What is AI persona architecture?

AI persona architecture is the structured system behind a conversational character. It connects role, audience, knowledge scope, communication rules, decision boundaries, memory, tools, risk controls and escalation so the persona remains useful in practical conditions.

Why it matters
A prompt can create an impression; architecture creates repeatability. It enables teams to test whether the persona behaves consistently when questions are ambiguous, emotional, adversarial or outside its knowledge.

What is conversational personality?

Conversational personality is the pattern of traits people infer from an interface over repeated exchanges. It emerges through word choice, pace, confidence, curiosity, humour, warmth and the way the system reacts when it cannot help.

Why it matters
Users judge personality from behaviour, not from the character biography. A credible design therefore specifies observable responses and tests how those responses change across languages, channels and demanding situations.

What is character consistency in conversational AI?

Character consistency is the degree to which an AI interface maintains its defined role, voice and boundaries across topics and interactions. Consistency does not mean repeating the same phrasing. A strong character adapts tone while preserving identity and rules.

Why it matters
Inconsistent behaviour weakens trust and exposes operational gaps. Evaluation should include routine questions, emotional exchanges, unsupported requests, attempts to override instructions and handovers to human teams.

What is conversational empathy?

Conversational empathy is the designed ability of an interface to recognise relevant emotional and situational cues and respond in a considerate, appropriate way. The AI does not feel human empathy. It can, however, acknowledge frustration, avoid insensitive language and adjust the response to the context.

Why it matters
False intimacy and exaggerated claims about machine feelings can be manipulative. Responsible empathy is transparent, restrained and focused on action: it helps the person, respects vulnerability and escalates when human judgement is needed.

What is emotional engagement in an AI conversation?

Emotional engagement is the attention, motivation or sense of connection created by how an interaction makes a person feel. It may come from encouragement, delight, confidence, relief or the feeling of being understood. It should never come from pretending the system is human.

Why it matters
Emotion can make information memorable and action easier, but it must serve the user’s goal. A fitness trainer can be motivating; a support interface handling a complaint should prioritise calm clarity over entertainment.

How Retoba applies it
Retoba evaluates emotional engagement as an observable conversation pattern, not as simulated humanity. See the measured engagement in the Kotányi AI Chef case.

What is tone adaptation?

Tone adaptation is the controlled adjustment of communication to the user’s context while preserving the brand’s core voice. The system may become more concise, explanatory, formal or reassuring based on the task, channel and signals in the conversation.

Why it matters
Unbounded mirroring can reproduce hostility, stereotypes or unsafe language. Adaptation therefore needs clear limits: the interface should meet the person where they are without abandoning brand standards or professional judgement.

What is relational UX?

Relational UX designs not only what a user can do, but the quality of the relationship created through repeated interactions. It considers trust, continuity, boundaries, repair after errors, memory, control and the ability to leave or reach a human.

Why it matters
Conversational products are judged socially as well as functionally. A fast answer can still damage the relationship if it is evasive, overconfident or difficult to correct. Relational quality should be designed and measured over time.

What is an AI conversational interface?

An AI conversational interface enables people to interact with a digital service using natural language while the system interprets context and generates or retrieves a response. It may use text, voice or multimodal input and can connect to content, tools and business systems.

Why it matters
Conversation can simplify complex choices, but fluency is not the same as usefulness. In a “talk to a website” experience, visitors can ask the site directly instead of navigating only through menus and pages. The interface still needs a defined job, trustworthy knowledge, accessible alternatives and honest communication about what it can and cannot do.

How Retoba applies it
Retoba makes conversation a visible part of the website rather than a chatbot hidden in a corner. See the GYM24 case study, talk directly to Mojster AVE and the Biomasa AI Ambassador, and explore the public Leapmotor implementation.

What is AI customer support?

AI customer support uses artificial intelligence to help customers find answers, diagnose issues, complete service tasks and reach the right human team. It can operate before, during or after a purchase, and should draw on approved knowledge and current account or product data where authorised.

Why it matters
The business case includes availability, faster resolution and lower repetitive workload. Success must also account for accuracy, customer effort, escalation quality and whether automation shifts hidden work onto the customer.

What is an AI support assistant?

An AI support assistant is a system designed for a defined set of service tasks for customers or employees. A customer assistant may resolve routine questions directly; a tool for human support agents can summarise history, retrieve procedures or draft responses for an advisor to review.

Why it matters
Separating the role clarifies accountability. A system that advises an employee can rely on human review, while one speaking directly to customers needs stronger controls over accuracy, claims and escalation.

How Retoba applies it
The Biomasa AI Support Agent helps users of Fröling heating systems and technical support teams with faster diagnostics, technical questions and remote support.

What is human handover?

Human handover is the controlled transfer of a conversation from an automated interface to a qualified person. A useful handover carries relevant context, explains what will happen next and avoids making the customer repeat information unnecessarily.

Why it matters
Escalation is part of the product, not evidence that the AI failed. Triggers can include explicit request, low confidence, sensitive subjects, repeated misunderstanding, legal risk or a task that requires human authority.

What does human-in-the-loop mean?

Human-in-the-loop describes a system in which people actively review, approve, correct or intervene in parts of an AI process. The human role may occur before release, during individual decisions, through sampled quality checks or when defined risk thresholds are reached.

Why it matters
Adding a person does not automatically create oversight. The reviewer needs sufficient information, competence, time and authority to challenge the output. The loop must be designed as an operational control, not a label.

What is omnichannel AI?

Omnichannel AI provides a coherent intelligence and service layer across channels such as web, app, messaging, social, voice and contact centres. It aims to preserve meaning and appropriate context while adapting interaction design to each channel.

Why it matters
Copying one bot everywhere is not omnichannel design. Channels differ in identity, privacy, urgency and available controls. Shared knowledge and governance should support experiences tailored to each channel rather than erase those differences.

What is knowledge base integration?

Knowledge base integration connects an AI interface to approved organisational information such as product documentation, policies, service procedures and frequently asked questions. The integration must address source ownership, permissions, freshness, versioning and traceability.

Why it matters
A language model’s general knowledge cannot substitute for current company truth. The quality of the conversation depends on selecting, cleaning, structuring and standardising the underlying knowledge, then assigning owners and an update process. A connection to poorly governed or stale material simply makes weak information easier to retrieve.

How Retoba applies it
Retoba first engineers the approved brand and product knowledge, then connects it to the assistant’s role, voice and safeguards. See the Kotányi AI Chef approach.

What is RAG for customer support?

Retrieval-augmented generation, or RAG, allows a support system to retrieve relevant approved information and use it as grounding when composing an answer. Retrieval narrows the evidence available to the model; generation turns that evidence into a response suited to the question.

Why it matters
RAG can improve accuracy and freshness, but it does not guarantee them. Teams must test retrieval quality, permissions, conflicting sources, citations, unsupported claims and what happens when no reliable evidence is found. Documented, incremental evaluation is essential because a change that improves personality or breadth can also weaken factual grounding.

What is intent routing?

Intent routing identifies what a person is trying to accomplish and directs the interaction to the appropriate answer, workflow, tool or human team. Routing can use rules, classifiers, language models or a controlled combination of methods.

Why it matters
A fluent response is unhelpful if it sends the issue to the wrong process. Effective routing includes confidence thresholds, fallbacks and monitoring for new or ambiguous intents that the existing taxonomy does not cover.

What is containment rate?

Containment rate is the proportion of automated support interactions completed without transfer to a human channel. The precise calculation depends on how an organisation defines a completed interaction, abandonment, repeated contact and time window.

Why it matters
A high rate is not automatically good. It can hide blocked escalations or incorrect answers. Interpret it alongside resolution on first contact, customer effort, satisfaction, accuracy, repeat contact and the severity of contained issues.

What is conversation data mining?

Conversation data mining is the systematic analysis of dialogue to identify recurring topics, intents, questions, language patterns, friction and opportunities. It can combine quantitative methods with qualitative review and should use data collected for a clear, lawful purpose.

Why it matters
Search analytics show what people typed; conversations can reveal why it mattered and what was still unclear. The resulting insight can improve products, services, support, content, campaigns and the conversational system itself.

How Retoba applies it
PulseDialog AI turns unstructured dialogue into structured signals for marketing, product and service teams. The GYM24 case shows the approach applied to 20,000 conversations.

What is conversation intelligence?

Conversation intelligence is the organisational capability to transform dialogue into reliable insight and action. It includes capture, classification, analysis, interpretation, governance and a feedback process through which teams decide what should change.

Why it matters
A dashboard of topics is not intelligence until someone can use it. Mature practice connects findings to owners in product, service, marketing, sales, knowledge management or risk. It also measures whether action improved the outcome.

How Retoba applies it
Retoba combines PulseDialog AI with human interpretation and connects the findings to business decisions. See how this extended the customer journey in the Peugeot case.

What is Voice of Customer?

Voice of Customer, or VoC, is the disciplined collection and interpretation of what customers need, expect, value and experience. Sources can include interviews, surveys, reviews, support contacts, sales calls, search behaviour and conversations with digital interfaces.

Why it matters
Conversational data adds spontaneous language and the context revealed by additional questions to a VoC programme. It should be balanced with other evidence because people who choose to converse may not represent the entire market.

What is intent mining?

Intent mining discovers and organises the goals people express across a body of conversations. Unlike a fixed intent taxonomy, it can reveal emerging tasks, combinations of needs and phrases that teams did not anticipate when the system was designed.

Why it matters
The findings can reshape navigation, training data, content and workflows. Human review remains important because the same sentence can signal different intentions depending on context, irony or what happened earlier in the dialogue.

What is topic detection?

Topic detection identifies the subjects discussed in individual messages or complete conversations. A system may assign known categories, discover clusters from the data or use both approaches to track volume and change over time.

Why it matters
Topic trends can reveal a launch issue, seasonal concern, content gap or rising service demand. Useful analysis distinguishes the topic from the customer’s intention and from whether the interaction ended well.

What is sentiment analysis?

Sentiment analysis estimates whether language expresses a positive, negative, neutral or more specific evaluative stance. More advanced approaches consider emotion, intensity, target and how sentiment changes during an interaction.

Why it matters
Sentiment is an uncertain interpretation, not a fact about a person. Sarcasm, culture, mixed feelings and short messages can mislead a model. It is best used as an aggregate signal or review aid, not as the sole basis for consequential treatment.

What is unmet needs detection?

Unmet needs detection looks for recurring situations in which people cannot achieve a desired outcome with the current product, service, information or process. Signals include repeated workarounds, missing features, unanswered questions, frustration and requests outside the existing offer.

Why it matters
Aggregated patterns can inform product and service development before they appear in formal research. Teams should validate the opportunity against strategic fit, frequency, value and broader market evidence rather than treating every request as a roadmap commitment.

What is customer language analysis?

Customer language analysis examines the words, metaphors, questions and distinctions people naturally use when describing a need or experience. It identifies differences between internal company terminology and the language audiences understand.

Why it matters
The findings can improve interface copy, campaign language, sales arguments, documentation and search coverage. Ethical use preserves meaning and avoids exploiting sensitive expressions or flattening diverse voices into a single “average customer”.

What is conversational knowledge gap analysis?

Conversational knowledge gap analysis identifies questions that an AI or organisation cannot answer reliably with its current approved sources. It examines failed retrieval, responses with low confidence, human escalations, contradictory content and unresolved questions.

Why it matters
Each gap is evidence for a knowledge backlog. Closing it may require new content, a policy decision, a system integration or a clearer boundary. A larger model alone is rarely the answer.

What is conversation clustering?

Conversation clustering groups exchanges according to semantic similarity without requiring every category to be defined in advance. Analysts can use it to discover recurring themes, unusual patterns or different ways people frame the same underlying need.

Why it matters
Clusters are analytical suggestions, not objective market segments. They require human interpretation, meaningful labels and checks for privacy, bias, stability and whether the grouping supports a legitimate business question.

What is first-party conversational data?

First-party conversational data is information an organisation collects directly through interactions it operates, such as a website assistant, service chat or branded messaging experience. It can include questions, responses, outcomes and operational metadata, subject to the stated purpose and applicable law.

Why it matters
Direct data can be highly relevant, but ownership does not remove responsibility. Organisations need transparent notice, data minimisation, retention rules, access controls and safeguards against using sensitive conversation content for unrelated purposes.

How Retoba applies it
Retoba uses PulseDialog AI to turn voluntary conversational signals into themes that can improve offers, services and communication. See the GYM24 insight drawn from its own conversational data.

What is zero-party data?

Zero-party data is information a person intentionally and proactively shares with a brand, such as preferences, goals or desired outcomes. In conversation, it can appear naturally when the interface asks a relevant question and explains how the answer will improve the experience.

Why it matters
Voluntary sharing can reduce guesswork, but the exchange must remain proportionate. A helpful question, clear purpose and genuine choice are more valuable than collecting extensive personal detail “just in case”.

What is SEO?

Search engine optimisation, or SEO, improves a website’s technical accessibility, relevance, authority and experience so its content can be discovered and presented in search. It includes crawling and indexing, information architecture, content, internal linking, performance, structured data and reputation.

Why it matters
Google states that its generative search features are rooted in core Search ranking and quality systems. If content cannot be crawled, indexed and trusted for search, AEO or GEO tactics cannot compensate for the missing foundation.

What is Answer Engine Optimization (AEO)?

Answer Engine Optimization improves content, evidence, entity structure and technical access so answer systems can retrieve and represent information accurately. It asks whether a source answers real questions clearly, supports its claims and remains useful when a passage is understood outside the full page.

Why it matters
AEO helps brands compete for visibility inside direct answers, summaries and comparisons. It is not a guaranteed placement mechanism or a replacement for SEO; it is a practical extension of strong search and knowledge work.

How Retoba applies it
Retoba combines an audit grounded in evidence, information and structure engineering, implementation priorities and ongoing guidance in its AEO programme.

What is Generative Engine Optimization (GEO)?

Generative Engine Optimization is the practice of improving a brand’s visibility and representation in answers generated by AI. It considers whether generative systems can find, understand, trust, cite and accurately describe the organisation’s information across different prompts and platforms.

Why it matters
The term is useful for measurement across platforms, but it should not encourage “AI hacks”. Durable work comes from distinctive expertise, corroborated facts, clear entities, strong public sources and technically accessible content.

How Retoba applies it
Retoba treats GEO as the layer of strong SEO and AEO that extends across platforms. First establish trustworthy, structured knowledge that machines can interpret, then test how major generative systems represent it. See the Retoba AEO and GEO framework.

How do SEO, AEO and GEO support complex B2B buying journeys?

In B2B markets, SEO creates discoverability during problem, category and supplier research; AEO makes expert knowledge usable in direct answers; and GEO improves how the organisation is represented in research and comparisons generated by AI. Together they connect the firm’s expertise, services, people and evidence across a buying journey that may involve many questions and stakeholders.

Why it matters
Buying groups verify claims over time and across sources. Useful B2B content therefore needs more than broad visibility: it should explain the problem, approach, fit, implementation, risks, proof, governance and commercial next step. Strong case studies, named expertise, client evidence and consistent external references help both people and machines evaluate credibility. Success should include accurate representation and qualified engagement, not only rankings or traffic.

What is AI visibility?

AI visibility is the extent to which a brand, product, person or source appears in relevant answers generated by AI and in agentic experiences. Visibility can include a citation, linked source, mention, recommendation, comparison or accurate use of the organisation’s information.

Why it matters
It must be measured across a representative prompt set and over time because answers vary by wording, location, model and date. Visibility without accuracy or relevance can create exposure but not business value.

How Retoba applies it
Retoba measures presence, citations, accuracy and competitive context across a repeatable set of prompts. A single vanity query is not a reliable measure. See how Retoba measures AI visibility.

What is an AI citation?

An AI citation is a visible attribution or link from a generated answer to a source used to support part of the response. Citation formats differ between systems and do not necessarily indicate that the source caused the entire answer or holds a fixed ranking position.

Why it matters
Citations can build authority and qualified traffic, while making claims easier for users to verify. Analyse which pages and topics earn citations, whether the surrounding answer is accurate and what visitors do after arriving.

What is entity clarity?

Entity clarity is the consistency with which a website and credible external sources identify what an organisation, product, person or concept is and how it relates to other entities. Names, descriptions, locations, ownership, expertise and references should not contradict one another.

Why it matters
Clear entities reduce ambiguity for people and machines. Useful practices include accurate organisation information, coherent author profiles, descriptive internal links, appropriate structured data and corroboration from independent sources.

What is external brand authority in search and answers generated by AI?

External brand authority is the body of credible evidence about an organisation, product or service that appears in sources the brand does not fully control. It can include genuine reviews and ratings, relevant media coverage, expert references, customer recommendations, professional directories, awards, citations and consistent public discussion.

Why it matters
Independent evidence can corroborate what a brand says about itself, reduce uncertainty and help people evaluate a recommendation. It may also support visibility in contexts where search or AI systems compare information across several sources. Quality, relevance, recency and consistency matter more than volume alone. No rating, mention or award guarantees a ranking, citation or recommendation, and manipulated reviews can weaken trust.

Sources and scope
Google Business Profile guidance states that more reviews and positive ratings can help local ranking. This is a local search statement, not a universal ranking rule. IAB research on shopping with AI found that clear sourcing and verified customer reviews were important trust signals for people evaluating recommendations.

What is structured data?

Structured data is markup that machines can read to identify entities and relationships on a page through a shared vocabulary such as Schema.org. It can identify an organisation, article, product, breadcrumb or, in this glossary, a defined term set.

Why it matters
Markup can reduce ambiguity and enable supported search features, but it is not a shortcut to AI inclusion. Google requires structured data to match visible content and says no special schema is required for its generative search features.

What is grounding in generative AI?

Grounding connects a generated response to specified evidence, data or tools rather than relying only on patterns learned during model training. Sources may be retrieved from the web, an enterprise knowledge base, a database or an authorised application.

Why it matters
Grounding can improve relevance, traceability and freshness. It does not remove the need to test retrieval, source quality, conflicting evidence and whether the final wording actually follows the evidence.

What is query fan-out?

Query fan-out is a technique in which an AI search system issues multiple related searches across subtopics and data sources to address a complex question. The system can retrieve evidence for different parts of the requested explanation, comparison or plan.

Why it matters
A page may answer one part of a broader question even when it does not match the original wording exactly. Brands should build coherent topic depth and expert evidence rather than manufacture a separate page for every imagined query variation.

What is answer-ready content?

Answer-ready content gives a clear response that stands on its own and then supplies the explanation, evidence and context needed to evaluate it. It uses descriptive headings, precise entities, readable paragraphs and accessible HTML without writing in a robotic template.

Why it matters
Good answer structure helps people scan and helps retrieval systems identify relevant passages. There is no universal word count: a definition should be as concise as possible and as complete as the subject requires.

How Retoba applies it
The Retoba AEO audit identifies questions with high business value, weak answer passages, missing evidence and structural barriers, then translates them into a prioritised implementation plan.

What is product explainability in discovery through AI?

Product explainability is the degree to which a product’s purpose, attributes, suitability, limitations, compromises and supporting evidence can be accurately understood and communicated by people and AI systems. It requires consistent product data, clear language, current commercial facts and explanations that remain meaningful outside an internal catalogue.

Why it matters
A product can be excellent yet absent or misrepresented in comparisons made with AI if its information is ambiguous, incomplete or contradictory. Brands therefore need to design products not only for use and purchase, but also for reliable explanation by search systems, conversational interfaces and purchasing agents.

How Retoba applies it
Retoba structures product knowledge for both machine understanding and natural dialogue. The Peugeot 2008 AI Ambassador demonstrates this in a complex product category.

What is a large language model (LLM)?

A large language model is created through machine learning and trained on extensive text or multimodal data to predict and generate sequences such as language. It can summarise, classify, translate, reason over supplied context and produce new text, but its outputs remain probabilistic.

Why it matters
An LLM is a component, not a complete business solution. Reliability depends on the task, model, instructions, context, integrations, evaluation and controls wrapped around it.

What is generative AI?

Generative AI is a class of artificial intelligence that produces new content or representations in response to input, including text, images, audio, video, code or structured output. The output is generated from learned patterns and supplied context rather than copied from a single stored answer.

Why it matters
Generation enables flexible experiences, but it introduces uncertainty, provenance and governance questions. Organisations need testing for each use case rather than assuming that fluent output is accurate, original or appropriate.

What is an AI agent?

An AI agent is a system that interprets a goal, plans or selects steps and uses tools to act in a digital or physical environment with a degree of autonomy. It may search, read, write, call services, update systems or coordinate specialised components.

Why it matters
The moment an AI can act, risk moves beyond the quality of an answer. Permissions, identity, confirmation, reversibility, audit trails, spending limits and human authority become central design requirements.

What is retrieval-augmented generation (RAG)?

Retrieval-augmented generation is an architecture that retrieves relevant material at request time and supplies it to a generative model as context for the response. Retrieval may use keyword search, semantic similarity, metadata filters, graph relationships or a hybrid approach.

Why it matters
RAG lets organisations use current, private or specialised knowledge without training a new foundation model. Its quality depends on content preparation, retrieval, permissions, context selection and evaluation. Storing documents in a vector database is not enough.

What is a vector embedding?

A vector embedding is a numerical representation of content designed so semantically related items are located near one another in a multidimensional space. Text, images and other data can be embedded to support similarity search, clustering and recommendation.

Why it matters
Embeddings help retrieve relevant material even when wording differs, but similarity is not truth or permission. Metadata, access controls, exact matching and evaluation for the relevant domain are often needed alongside vector search.

What is prompt engineering?

Prompt engineering is the design and testing of instructions, context and examples that shape how a generative model performs a task. In production, prompts work with data, tools, policies, output schemas, model settings and evaluation. They are not isolated verbal tricks.

Why it matters
Clear instructions improve consistency and maintainability, but prompts cannot guarantee factual accuracy or security. Critical constraints should also be enforced through architecture, permissions, validation and human processes.

How Retoba applies it
Retoba combines prompt engineering with proprietary knowledge, character design, safeguards and continuous testing. See how these layers work together in the Kotányi AI Chef.

What are AI guardrails?

AI guardrails are technical and organisational controls that keep a system within defined boundaries. They can restrict topics, data access, tools, claims, output formats and actions; detect risky input or output; require confirmation; and route cases to people.

Why it matters
No single filter is sufficient. Effective guardrails are layered, tested against realistic failure modes, monitored after launch and paired with clear accountability for incidents and exceptions.

What is an agent-ready website?

An agent-ready website is designed so authorised AI agents can understand its information, identify available actions and interact with important workflows safely. It uses accessible content, meaningful labels, stable structures, explicit states and secure action interfaces.

Why it matters
Agent readiness extends accessibility and technical SEO into action design. Consequential steps need authentication, clear prices and consequences, confirmation, error recovery and protection against unintended or automated abuse.

What is a personal AI twin?

A personal AI twin is a personalised AI system configured with a person’s authorised knowledge, work patterns, language, preferences and decision principles to support defined tasks on that person’s behalf. It can help prepare analyses, draft communication, organise knowledge, simulate options and use approved tools while preserving relevant professional context.

Why it matters
A carefully governed twin can extend individual capacity and help preserve institutional memory, but it is not the person and should never imply unrestricted authority. Its scope, data sources, permissions, review requirements, provenance and removal process must remain explicit.

What is a computer-using agent (CUA)?

A computer-using agent is an AI agent that operates graphical user interfaces through actions such as clicking, typing, navigating, reading screens and submitting forms. It can work with existing software in a way that resembles human interaction, including where a dedicated application programming interface is unavailable.

Why it matters
CUA can accelerate automation across familiar tools, but visual interfaces change and actions taken on screen can have real consequences. Grant only the minimum access required, use isolated environments where appropriate, confirm consequential actions, keep detailed logs and spending limits, and provide a recoverable path when the interface or agent behaves unexpectedly.

What is the agent economy?

The agent economy is an emerging environment in which AI agents discover information, compare options, coordinate services and execute tasks or transactions for people and organisations. Agents may interact with websites, applications, data services, other agents and the people responsible for decisions.

Why it matters
Brands will increasingly need to be understandable and actionable to machines as well as persuasive to people. Product information, authority, identity, permissions, commercial terms and transaction states must be clear enough for agents to interpret without removing meaningful human control.

What is augmented intelligence?

Augmented intelligence is an approach in which AI extends human perception, analysis, creativity and judgement rather than removing the responsible human role. The system can organise evidence, generate alternatives, test assumptions and reveal patterns; people define purpose, values and the decision that follows.

Why it matters
The strongest use of AI is not always full automation. In strategic, creative and consequential work, value often comes from a partnership in which AI increases the range and speed of thought while people challenge the output, supply context and remain accountable.

How Retoba applies it
REWORK by Retoba examines where AI should automate, where it should augment people and how roles can be redesigned without losing judgement, clarity or human responsibility.

What is an AI-first organisational culture?

An AI-first organisational culture treats responsible AI use as a standard capability in everyday work rather than an isolated technology experiment. It provides people with approved tools, practical learning, time to test, shared methods, clear guardrails and accountability for outcomes.

Why it matters
AI adoption is primarily an organisational challenge. Licences alone do not change decisions or workflows. Teams need leadership, psychological safety for controlled experimentation, evidence of value and clarity about how roles evolve. A useful pattern is to begin with one bounded problem, learn quickly and expand from proven practice.

How Retoba applies it
Retoba’s REWORK approach begins with real work, decisions and constraints, then redesigns roles, workflows and collaboration between people and AI around measurable value.

When does a prompt become an organisational asset?

A prompt becomes an organisational asset when it codifies reusable domain knowledge, decision criteria, desired behaviours and evaluation standards in a tested and maintainable system. Its value may include instructions, examples, tool rules, source requirements, output structures, failure handling and the documented evidence that it improves a defined task.

Why it matters
A valuable prompt system should be versioned, protected by access controls, evaluated and assigned to an owner like other operational knowledge. Calling it an asset does not mean every prompt automatically receives intellectual property protection; legal protection depends on its content, originality, confidentiality, contracts and applicable law.

What is conversational AI lifecycle management?

Conversational AI lifecycle management is the continuous ownership, testing, monitoring, updating and improvement of a conversational system from initial design through operation and eventual retirement. It covers knowledge, prompts, models, integrations, character behaviour, safety controls, analytics, feedback and change records.

Why it matters
A conversational AI system is a living service, not a single implementation. Products, policies, language, user behaviour and model capabilities change. Organisations need a named owner, update cadence, quality thresholds, incident process and a way to turn real conversation evidence into controlled improvements.

What is data sovereignty in AI?

Data sovereignty in AI is an organisation’s ability to retain meaningful control over how its data is stored, accessed, processed, shared, activated and moved across systems and providers. It combines legal jurisdiction with practical questions of ownership, portability, security, dependency and the ability to create value from information collected directly through the organisation’s own channels.

Why it matters
An organisation that cannot retrieve, govern or reuse its own data may become dependent on external platforms and pay repeatedly to access value it helped create. Sovereignty does not mean isolating all data; it means making deliberate, lawful choices about infrastructure, permissions, retention, providers and exit paths.

What is the EU AI Act?

The EU AI Act is the European Union’s legal framework for artificial intelligence, organised around levels of risk. It defines prohibited practices, obligations for high-risk systems, rules for general-purpose AI, transparency duties and responsibilities for actors across the AI value chain. It entered into force on 1 August 2024, with provisions applying in stages.

Why it matters
Most provisions became applicable on 2 August 2026, while amended timelines apply to specified high-risk categories. Classification depends on the system’s intended purpose and use: a marketing or support interface is not automatically high-risk merely because it uses AI.

Who is an AI provider under the EU AI Act?

An AI provider is the person or organisation that develops an AI system or general-purpose AI model, has one developed, and places it on the market or puts it into service under its own name or trademark. The exact role follows the legal definition and facts of the arrangement, not the everyday meaning of “vendor”.

Why it matters
A company integrating an external model into a branded service may have different obligations depending on how the system is developed, modified and supplied. Contracts should clarify roles without assuming that contractual wording overrides the law.

Who is an AI deployer under the EU AI Act?

An AI deployer is a person or organisation using an AI system under its authority in a professional context, except for personal activity outside a professional context. A brand operating an AI interface for customers may therefore be a deployer even when an external supplier provides the technology.

Why it matters
Deployers cannot outsource every responsibility to a technology provider. Depending on the system, they may need to follow instructions, maintain oversight, monitor operation, protect input data and cooperate with relevant obligations.

What is AI risk classification?

AI risk classification determines which legal and organisational controls apply based on the system’s purpose, context, affected people and potential harm. Under the EU AI Act, relevant categories include prohibited practices, high-risk systems, specified transparency cases and systems without additional obligations specific to the Act.

Why it matters
Classification should be documented before deployment and revisited when purpose, users, data or functionality changes. Organisations may also apply stricter internal tiers for privacy, brand, security or commercial risk.

What is an AI transparency obligation?

An AI transparency obligation requires specified information to be provided so people understand that AI is involved or can recognise material generated or manipulated by AI. Article 50 of the EU AI Act includes duties for systems intended to interact directly with people, unless the interaction with AI is obvious from the circumstances and context.

Why it matters
A branded AI character should identify its nature clearly and at the right moment. Disclosure must be accessible and understandable; burying it in terms and conditions does not create a transparent conversational experience.

What is AI literacy?

AI literacy is the knowledge and practical understanding that enables people to deploy, operate and interpret AI systems responsibly in their context. It includes awareness of capabilities, limitations, risks, expected controls and the consequences for affected people.

Why it matters
The EU AI Act requires providers and deployers to take measures, to their best extent, to ensure an appropriate level of AI literacy among relevant staff and others operating systems on their behalf. Generic training is insufficient when roles and risks differ.

What is human oversight of AI?

Human oversight is the set of roles, information, controls and authority that enables people to understand, supervise, challenge, interrupt or override an AI system where appropriate. The form and intensity should match the task and risk.

Why it matters
A nominal reviewer is not enough. Effective oversight requires competence, time, accessible evidence, awareness of automation bias and the practical ability to act before harm occurs.

What is responsible AI?

Responsible AI is the practice of designing, developing, procuring and operating AI in ways that are lawful, purposeful, accountable and proportionate to potential impact. Common concerns include validity, safety, security, transparency, privacy, fairness, accessibility and meaningful human control.

Why it matters
Principles become credible only when they shape decisions, documentation, tests, monitoring and incident response. Responsible AI is an operating discipline across the lifecycle, not a statement added after launch.

How influential is AI in shopping decisions?

Among people who already use AI while shopping, it has become one of the most influential sources in the decision process. In a 2025 IAB and Talk Shoppe study, AI ranked second only to conventional search and ahead of retailer websites, apps, friends and family. 46 per cent used it during most or every shopping journey, and almost 90 per cent said it helped them discover products they might not otherwise have encountered.

What this means for brands
AI is becoming a new layer of discovery and consideration. Brands need information that AI systems can accurately understand, compare and support with evidence.

Source and scope
IAB, When AI Guides the Shopping Journey, 2025. The research included more than 450 observed shopping sessions involving AI and a survey of 600 US AI shoppers aged 18–64.

Does AI make the purchase journey shorter?

AI can make individual tasks faster while expanding the overall decision journey. In the IAB study, 95 per cent of participants took at least one additional action after consulting AI. The average number of shopping steps increased from 1.6 before the AI interaction to 3.8 afterwards, while visits to retailer and marketplace websites almost tripled.

What this means for brands
AI creates an additional discovery and validation layer rather than simply replacing search or the website. Claims, comparisons and information on landing pages must remain consistent across the complete journey.

Source and scope
IAB and Talk Shoppe research report, 2025.

What makes an AI shopping recommendation trustworthy?

People trust AI shopping recommendations when they can inspect the source, verify the evidence and understand why an option was recommended. Only 46 per cent of participants in the IAB study said they fully trusted AI recommendations, while 89 per cent verified them independently. Source transparency, verified reviews and explanations were the strongest reported trust signals.

What this means for brands
AI visibility without verifiable evidence is fragile. Reliable product facts, independent recognition, reviews, case studies and clear policies create the evidence architecture behind a recommendation.

Source and scope
IAB, When AI Guides the Shopping Journey, 2025.

How widely is AI already used across the buying journey?

AI is already influencing product research, review interpretation and deal discovery while physical retail remains highly relevant. An IBM Institute for Business Value and National Retail Federation study found that 45 per cent of surveyed people used AI during their buying journeys: 41 per cent for product research, 33 per cent to interpret reviews and 31 per cent to find promotions. At the same time, 72 per cent continued to shop in physical stores.

What this means for brands
The transition is towards decision processes shaped by AI across channels, not simply from physical to digital retail.

Source and scope
IBM Institute for Business Value and NRF, Agentic Commerce, 2026. The study included more than 18,000 people in 23 countries and 200 executives in retail, consumer products and ecommerce.

Are people ready to let an AI agent complete a purchase?

A significant share of people appear open to delegating the transaction after defining their preferences. In Adyen’s 2026 US retail study, 51 per cent said they would be willing to let AI manage the complete shopping process, including the purchase. This measures stated willingness, not the proportion of purchases already completed autonomously.

What this means for brands
Brands must prepare to be evaluated, selected, purchased, returned and supported through processes mediated by agents, while preserving explicit permission and accountability.

Source and scope
Adyen Retail Report 2026, United States. The research covered 2,000 US participants and 500 US retail merchants.

What makes people trust an AI agent with the final purchase?

Trust in an autonomous purchasing agent depends on transparency, accountability and control. Adyen found that participants wanted assurance of the best price or value, clear accountability for an incorrect purchase and an explanation of why an item was selected. An international study from 2026 similarly found that transparency strengthened cognitive trust, accountability strengthened affective trust and decision control reduced perceived risk.

What this means for brands
An agentic experience should explain decisions, identify responsibility, allow limits to be changed and provide a clear route for correction, return or human intervention.

Sources and scope
Adyen Retail Report 2026 and Letting the bot decide, Journal of Retailing and Consumer Services, 2026.

How could agentic shopping change brand loyalty?

Agentic shopping may weaken habitual loyalty when an agent chooses according to price, availability or functional criteria instead of brand preference. Adyen found that 27 per cent of retailers considered the possible loss of their direct customer relationship a barrier to agentic commerce, while 59 per cent of participants said loyalty benefits would make them more likely to continue shopping with a retailer.

What this means for brands
Quality, membership benefits, service history, delivery reliability and return conditions must become legible to both people and agents.

Source and scope
Adyen Retail Report 2026, United States.

How large could agentic commerce become?

Morgan Stanley estimates that AI agents could influence between USD 190 billion and USD 385 billion of US ecommerce spending by 2030. This would represent approximately 10–20 per cent of the market. Groceries and consumer packaged goods may be early growth categories because purchases are frequent, guided by clear criteria and often repetitive.

What this means for brands
Information that machines can interpret about ingredients, compatibility, pack size, availability, pricing, sustainability and replenishment can become commercially decisive.

Source and scope
Morgan Stanley Research, Agentic Commerce Market Impact Outlook, 2025. These figures are forecasts, not observed future results.

What creates trust in an AI chatbot?

Trust in an AI chatbot is created primarily through demonstrated ability and integrity. A 2025 study involving more than 1,300 participants found perceived ability and integrity to be significant foundations of trust. Greater trust was associated with stronger intentions to use the chatbot and greater willingness to disclose information.

What this means for brands
Character and conversational style matter, but competence comes first. The system must give accurate answers, respect its role, protect information and acknowledge uncertainty.

Source and scope
When the bot walks the talk, Journal of Experimental Psychology: General, 2025. The research comprised two pilot studies and a main study with 1,001 participants.

Does personalisation automatically increase trust in an AI chatbot?

No. Personalisation can increase relevance and willingness to interact without automatically increasing trust. In experimental research, a personalised chatbot was perceived as more capable, benevolent and human in its behaviour and generated stronger usage intentions, but personalisation did not directly increase trust. Separate retail research found that generative AI increased perceived usefulness and familiarity while trust remained largely unchanged and privacy concerns increased.

What this means for brands
Personalisation needs transparent data use, permission boundaries and a way to correct or remove remembered preferences.

Sources and scope
When the bot walks the talk, 2025 and From familiarity to acceptance, Journal of Retailing and Consumer Services, 2025.

Can empathetic communication close the gap between AI and human service?

Carefully designed empathetic communication can significantly improve how people evaluate an AI service agent. Across several studies, lower evaluations of chatbots were explained by lower perceived empathy. When the chatbot communicated more empathetically, evaluations improved and in one study reached a level comparable with a human agent. The improvement came from communication rather than an avatar designed to look human.

What this means for brands
Designed empathy means recognising the context, acknowledging the concern and helping the person move forward without implying that the machine experiences human feelings.

Source and scope
Consumer reactions to chatbot versus human service, Journal of Retailing and Consumer Services, 2024. The research included 714 participants across three vignette studies and a further study involving direct interaction.

Can an AI character strengthen brand trust?

An appropriately designed AI character can make a brand more recognisable, relatable and engaging. Research into branded voice assistants found positive relationships between brand anthropomorphism, brand trust and emotional and behavioural engagement. The result does not mean that every interface should imitate a human as closely as possible.

What this means for brands
A useful character requires a defined role, personality, vocabulary, emotional range and behavioural boundaries. Role fit and trustworthy performance are more important than maximum resemblance to a human.

Source and scope
Hey Google, I trust you!, Journal of Retailing and Consumer Services, 2024. The study focused on young adults and branded voice assistants.

How should a brand disclose that a conversational agent is AI?

A brand should disclose an agent’s artificial identity clearly, naturally and early enough for the person to understand the interaction. Experimental research suggests disclosure can reduce trust in some contexts, but strong social presence, competent communication and appropriate timing can reduce the effect. This evidence must never be used to justify concealment.

What this means for brands
The introduction should explain that the interface is AI, what it can do, how it uses information, where its limitations lie and when a human can take over. Applicable transparency obligations, including Article 50 of the EU AI Act, must also be considered.

Source and scope
The impact of providing cues that reveal a nonhuman identity, European Journal of Marketing, 2025.

Can chatbot service quality influence brand loyalty?

Chatbot quality can influence the wider brand relationship because people interpret the interface as evidence of the organisation’s competence and reliability. A 2026 banking study found positive relationships between information quality, system quality and trust; trust was associated with satisfaction, while satisfaction was associated with loyalty and positive electronic word of mouth.

What this means for brands
Incorrect information, broken conversational paths and poor escalation can damage the brand. Reliable answers, continuity and effective resolution can strengthen satisfaction.

Source and scope
Trust in the digital age: Chatbot service quality, Telematics and Informatics Reports, 2026. This was a correlational study of 362 banking customers in Peru and does not establish universal causality.

How much autonomy should an AI agent have?

The appropriate level of autonomy depends on the consequences, the person’s preferences and the ability to reverse an action. A Wharton evidence review suggests that moderate autonomy is often more acceptable than either minimal assistance or complete automation. People value setting boundaries, reviewing important decisions and intervening when necessary.

What this means for brands
Autonomy should increase gradually and remain based on explicit permission: recommend first, prepare an action next and execute only under defined authority. Consequential actions should remain visible, reviewable and reversible.

Sources and scope
Wharton, Blueprint for AI Agent Adoption, 2026 and Letting the bot decide, 2026.

What technical foundations are emerging for agentic commerce?

Agentic commerce requires shared protocols for exchanging product, identity, permission and transaction information securely. OpenAI’s Agentic Commerce Protocol, Google’s Universal Commerce Protocol, Visa’s Trusted Agent Protocol and Mastercard Agent Pay represent different parts of that emerging infrastructure. They do not yet form one universal standard.

What this means for brands
Preparing for agents requires reliable product data, current inventory and pricing, clear merchant identity, policies that machines can interpret, secure permissions, payment handling and operational processes for returns and support.

Primary sources
OpenAI Agentic Commerce Protocol, Google Universal Commerce Protocol, Visa Trusted Agent Protocol and Mastercard Agent Pay.

How much are enterprises investing in conversational AI?

Juniper Research estimated that global revenue generated by enterprise spending on conversational AI would rise from USD 14.6 billion in 2025 to more than USD 23 billion in 2027. Across the three years from 2025 to 2027, conversational AI services were forecast to generate approximately USD 57 billion in global revenue.

What this means for brands
Conversational AI is moving from isolated experimentation into operational infrastructure spanning messaging, voice, service, sales and increasingly agentic workflows.

Source and scope
Juniper Research, Global Conversational AI Market 2025–2029. The forecast measures revenue originating from enterprise spending on conversational AI platforms, not sales completed through them.

How fast could advertising in generative search grow?

WPP Media forecasts global advertising revenue from generative search of USD 5.1 billion in 2026, approximately USD 32 billion in 2028 and more than USD 100 billion by 2030. Its category includes paid placements and impressions within discovery environments shaped by AI, including search experiences generated by AI and standalone conversational AI products.

What this means for brands
AEO and GEO are preparation for an emerging media and commerce channel in which brands may compete through both earned inclusion and paid visibility inside generated answers.

Source and scope
WPP Media, This Year Next Year: 2026 Midyear Global Advertising Forecast. These are forecasts for a newly defined advertising category and should not be treated as confirmed future expenditure.

Are advertisers already investing in answers generated by AI?

Yes. In a 2026 McKinsey survey, more than half of participating advertisers said they were investing in advertisements embedded in answers generated by AI. Almost 75 per cent expected AI to increase total media spending during the following 12 months, while one in three expected AI to improve return on advertising spend by more than 10 per cent.

What this means for brands
AI is changing both campaign execution and the environments in which people discover, compare and select brands. Being surfaced and recommended is becoming a distinct commercial objective.

Source and scope
McKinsey, The agentic advertising economy, 2026. The survey covered 182 agency and marketing leaders from the United States, representing organisations with annual marketing spending between USD 5 million and more than USD 5 billion.

Are advertising budgets moving from search towards conversational AI?

Early evidence indicates that part of the budget is moving from conventional search and the open web into discovery shaped by AI. McKinsey reports that roughly 40 per cent of expenditure reallocated to formats driven by AI is shifting away from traditional search and the open web. WPP Media expects generative search to grow rapidly and forecasts that conventional search could begin to decline from 2029, while remaining the larger channel for some time.

What this means for brands
In search, a brand competes for ranking and clicks. In conversational discovery, it also competes to enter the answer, comparison set and recommendation. SEO, AEO, GEO and product information designed for agents must work together.

Sources and scope
McKinsey, 2026 and WPP Media, 2026. These are projections of an emerging transition, not a recommendation to abandon conventional search.

How quickly are brands investing in agentic and conversational marketing?

Agentic and conversational marketing are moving rapidly from awareness to practical investment. In the IAB 2026 Outlook Study, 78 per cent of advertising buyers expected greater focus on generative AI in media campaigns, 73 per cent on optimising content for answers generated by AI and 66 per cent on agentic AI for ad buying and campaign execution. Gartner separately predicts that 60 per cent of brands will use agentic AI for streamlined individual interactions by 2028.

What this means for brands
Investment needs to extend beyond technology licences to knowledge management, data governance, conversational design, brand voice, measurement, integration and human oversight.

Sources and scope
IAB 2026 Outlook Study, based on 205 advertising buyers, and Gartner’s 2028 forecast for agentic marketing.