Found by AI is a three-part Velo series on how B2B buyers now use AI to choose suppliers. Each sector is found in a different way. Manufacturers have to unlock what they already know. Technology firms have to prove what they claim. Professional services firms have to earn the right to be cited.
A design engineer needs a seal rated to 200°C that will survive food-grade washdown. She used to search, download five datasheets and ring a distributor. Now she asks an AI tool and gets three suppliers back in seconds. Your product could meet every requirement – but if you’re not in the AI shortlist, chances are you won’t make it onto hers either.
Industrial buying has already moved online. In industrial distribution, 57% of buyers now rank digital as their main purchasing channel, up from 33% in 2022. Buyers are also comfortable spending serious money remotely. McKinsey’s research on energy and materials buyers found that 75% are comfortable spending more than $50,000 through digital self-service or remote channels, and 42% more than $500,000. Gartner reports that 67% of B2B buyers prefer to buy without a sales rep.
AI answers here, but only if you feature in the LLMs like ChatGPT and Claude – they are deciding which products appear before a buyer visits any website.
Where manufacturers lose the answer
AI engines answer questions with hard limits by finding facts they can read and trust. Most manufacturers hold better facts than anyone else in their market. They lose the answer in three places:
- The facts are locked away. Specifications sit in gated PDF datasheets, scanned catalogues and configurators that AI crawlers can’t open. Much of the most valuable knowledge lives in senior engineers’ heads.
- The facts don’t match. Buyers now use an average of 10.2 channels in their buying journey, up from five in 2016. Your product appears on your site, on three distributor sites and in two trade directories, often with different numbers. When sources conflict, AI hedges or picks the wrong figure.
- Wrong facts are costly. 51% of B2B buyers say they’re more likely to encounter misleading information from GenAI. An out-of-date tolerance in an AI answer loses you more than a lead; in engineering, it becomes a liability.
The upside is that few manufacturers have fixed any of this yet. Demand is rising and competition is low.
Why the specification decides the shortlist
Around 95% of B2B buyers aren’t in-market at any given time (source: Ehrenberg-Bass Institute and LinkedIn B2B Institute). They’re forming impressions that will decide your place on the shortlist later. In manufacturing, a buyer usually moves for one of five reasons:
- A supplier fails.
- A new application comes up.
- A cost or compliance review begins.
- A regulation changes.
- An innovation makes the business case.
When one of those happens, the engineer writes a specification, and AI helps her decide who meets it. Procurement then checks the decision; procurement professionals are decision-makers in 53% of buying cycles. Your content has to satisfy the engineer’s technical questions and procurement’s commercial ones.
Get onto the shortlist early and you’re likely to stay there. 68% of buyers already have a front-runner in mind when they start, and that vendor wins 80% of the time.
What AI needs to see: the four Ps
- Pedigree. Accreditations, capacity, facilities and ESG credentials. AI uses these as filters: “AS9100-certified machining in the Midlands” is a prompt your accreditations page has to answer.
- Problems. Application guides and failure-mode explainers, written around the questions engineers bring to you every week.
- Proposition. Full specification pages as web pages, not PDFs: materials, ratings, tolerances, compatibility and part equivalents.
- Proof. Case studies that state the operating conditions and the outcome in numbers: hours of uptime, service life, total cost of ownership.
Steps to influence your appearance in LLMs: unlock, structure, align
Unlock it (architecture)
- Check your robots.txt file and hosting settings aren’t blocking AI crawlers. Fix crawl errors, broken links and duplicate pages.
- Turn your most-requested datasheets into web pages with specification tables in plain text.
- Improve page speed and mobile performance. Engineers often research from the shop floor.
Structure it (on-site content)
- Add structured product data (a standard code format AI and search engines can read) so specifications are machine-readable.
- Give every product family a page carrying the full specification, not a summary with a download button.
- Build an FAQ from the questions your technical sales team answers every week, in your customers’ words.
- Publish equivalence tables (“replaces part X”, “compatible with Y”).
- Name the engineer who checked each page, and show when you last revised it.
Align it (off-site signals)
- Audit distributor listings against your own data and correct every mismatch.
- Keep trade directory and industry body listings complete and current.
- Place technical articles in the trade press, since earned coverage carries weight with AI.
- Clean up your Google Business Profile.
Ready to make progress? This is our recommendation for making a difference in just 90 days
It starts with good SEO practices. This means many of the tools you use to optimise for traditional search can help you optimise for AI:
Days 1 to 30: Get access to Google and Bing Search Console then review the errors it reports. Run a prompt analysis: ask ChatGPT, Perplexity and Google’s AI Mode the 20 questions your engineers hear most, and note who gets named. Fix titles, headings, redirects and duplicate content. Publish specifications and case studies you already hold but haven’t put online. Expect errors to fall.
- Days 31 to 60: Build case studies with operating data and publish application guides. Tidy up directory listings. Improve pages with high bounce rates. Expect rankings and traffic on key terms to rise.
- Days 61 to 90: Build a citation and backlink plan with the trade press and your distributors, and form a digital PR plan. Rerun the prompt analysis and compare. Expect your AI appearance to grow.
AI visibility is a supporting device, not a lead-generation engine on its own. Its job is to put you on the list before a trigger fires.
Unlock what you already know
Your engineers know more about your products than any competitor, distributor or AI model. Getting found starts with releasing that knowledge in a form that machines can read and every channel repeats.
Also in this series: In Part 2, technology firms learn why AI rewards proof over promise. In Part 3, professional services firms learn why only genuine thought leadership gets cited.