Imagine you’re choosing between a handful of AI products. The demos look impressive and the websites make similar promises, but you’re still unsure who to trust. Then someone whose judgment you respect tells you which one they use, and why.

“Distribution is everything” has become a meme among AI founders. But a launch video or billboard can only do so much when buyers can’t tell who to believe.

Lulu Cheng Meservey captured the shift: first distribution was scarce, then attention. As feeds filled with bait and slop, credibility became scarce.

AI makes the appearance of a serious company easier to produce, just as customers are being asked to entrust software with more consequential work. When polished websites and persuasive copy tell us less, the judgment of people we trust becomes more useful.

For a new startup, endorsements from credible sources give buyers a reason to take a chance before you’ve built a track record of your own.

A customer logo suggests another organisation has evaluated you and found a reason to buy. A reference call goes further, letting the buyer hear how the product works from someone who uses it.

Relevance matters more than fame. A customer facing the same constraints as your prospect may carry more weight than a household name solving a different problem.

Ambience Healthcare is a good example. Cleveland Clinic announced its rollout after evaluating AI documentation tools across more than 80 specialties and subspecialties. For another health system considering Ambience, that provides evidence from a peer with demanding requirements.

Industry advisers can confer similar credibility. If you’re selling security software, a practitioner respected by CISOs may help you get a hearing that a large general audience never would. Within weeks of launching in 2022, Island announced investor-advisers including Silicon Valley CISO Investments, a group of practising security leaders. The announcement featured the CISOs of SoFi and Chipotle explaining why they believed in its enterprise browser.

When shortlisting influencers, look for someone whose advice your customers seek before buying, however small their public following. If they’ve worked with your product and can explain why they believe in it, their endorsement has substance. It also carries a reputational cost if they’re wrong.

Before spending more to reach your next customer, ask who they already trust, and what would persuade that person to put their name behind you.

Links

Designing effective evals is critical to scaling industry-focused applications and services firms, particularly in domains with traditionally “unverifiable” work. Really enjoyed this case study of how Crosby is doing it with contracts:

He’s back:

Very exciting to see this level of collaboration to pull together the data needed to make AI-driven predictions in biology: