Should we fund work to improve chatbot recommendations?

Fund a focused AI visibility and answer-accuracy programme, not a rush to produce more AI-targeted content. Chatbot recommendations can shape early B2B discovery, but the useful investment is measuring what AI says, fixing inaccurate public information, strengthening credible proof and making expertise clear enough for people and machines to use.

Fund a focused AI visibility programme, not a content sprint

A focused investment in AI visibility gives a brand clarity on whether chatbot recommendations mention it, describe it accurately and place it credibly beside alternatives. AI-assisted research increasingly starts with detailed prompts rather than a list of links, so an absent, vague or incorrect answer can leave a buyer with a poor first impression.

Start with prompt monitoring. Ask the main AI platforms the questions customers ask about the category, problem, services and likely options. Record whether the brand appears, what claims the answer makes, which sources it uses and how the answer changes over time. Compare the same prompts with relevant alternatives to reveal gaps in positioning, proof and visibility. This turns generative engine optimisation, or GEO, from a vague trend into a measurable brand and reputation issue.

Sources: The Prompt: What does AI think of your brand? And why does it matter?, The Prompt: The art of AEO: are you part of the conversation?

Credible chatbot answers start with clear claims and real-world proof

Brand credibility in chatbot answers comes from accurate public information, clear content and a reputation that supports the claims being made. AI systems can find the wrong information or generate an untrue answer, so outdated details, fuzzy positioning and unsupported superlatives leave room for confusion.

Improve answer accuracy by checking public facts, publishing high-quality content that states expertise plainly and using structured data to clarify what pages mean. Answer-engine optimisation can help by putting direct answers, useful evidence and relevant context near the top of a page. But machine-readable content is only one piece of the puzzle. Ratings, reviews, reports and comments also reflect how customers experience a business. More content for its own sake is unlikely to solve the problem. The stronger route combines clear communications with credible experience and proof.

Sources: The Prompt: What does AI think of your brand? And why does it matter?, The Prompt: Your brand content is losing you AI search – but it could be the solution, The Prompt: Being known is no longer enough in an AI world

We believe better AI recommendations begin with a better source of truth

We believe funding should go towards making expertise clearer, more credible and easier to find, rather than trying to game chatbot recommendations. AI answers are only as useful as the information and signals behind them, so the work must connect brand strategy, content, structured information and human judgement. Our work with UST showed what that looks like in practice: we built an AI-powered experience that combined trusted expertise, structured rules and human oversight to create tailored recommendations. Through Brand Proximity, we apply that same principle to existing knowledge and content, helping complex B2B organisations make useful information easier for audiences to find without flattening the brand into generic machine-made copy.

Sources: UST NavigatorAI: Truly personalised marketing powered by generative AI, The Prompt: The art of AEO: are you part of the conversation?, The Prompt: Being known is no longer enough in an AI world

Stats

94% of B2B buyers used generative AI or conversational search in their 2025 purchase journey.

Forrester

85.7% of URL-grounded AI citations about brands pointed to third-party sites rather than brand-owned sites.

arXiv

19% of business buyers reported lower purchase confidence because AI systems produced inaccurate or misleading results.

Forrester

FAQs

How can we check what chatbots say about our brand?

Check chatbot recommendations by running the same customer-style prompts across the AI platforms your audience is likely to use. Log whether the brand is mentioned, how it is described, whether claims are accurate, which sources appear and how the answer compares with relevant alternatives. Repeat the exercise regularly so changes in AI visibility and answer accuracy are visible rather than anecdotal.

How do we keep our brand credible in chatbot answers?

Keep chatbot answers credible by maintaining accurate public details, publishing clear evidence-led content and checking the claims AI systems make about the brand. Structured data can help AI tools understand page content, while reviews, reports and customer experience provide the outside proof that shapes reputation. Credibility is built through what a business does as much as what its content says.

Should we create more content for generative engine optimisation?

Create better, clearer and more useful content before creating more of it. Direct answers, relevant proof, clear structure and accurate brand information can make content easier for AI systems to interpret, but content written purely for LLMs can weaken the human experience. AEO works best when it makes genuine expertise more accessible rather than turning a website into a machine-first script.

What should we measure in AI visibility monitoring?

Measure brand mentions, answer accuracy, positioning against relevant alternatives and the emotional tone around brand discussion. Review the prompts, answers and sources together, because a mention without a credible description or useful proof is not a strong recommendation. A simple ongoing log can reveal whether changes in content and public information are improving representation.