Latest / B2B SaaS Talks with Fexingo: Enterprise Software, Sales Cycles, and Procurement / Why Enterprise Buyers Now Mandate a Vendor AI Ethics Committee
Transcript
- Lucas: So we talk a lot on this show about the boxes enterprise buyers are now checking before they sign a software deal. We've covered SOC 2 Type II reports, cyber insurance floors, financial health scores. But there's one requirement that's quietly become a deal-breaker in the last twelve months, and I think it's worth digging into. Luna: Which one? Lucas: The vendor AI ethics committee. Not an AI ethics policy, not a set of principles on a website — a formal, named committee that meets regularly, has a charter, includes a diverse set of stakeholders, and has some real authority over product decisions. Luna: I've seen that pop up in RFPs more and more. I know a procurement lead at a large retailer who told me they walked away from a twelve million dollar deal earlier this year because the vendor couldn't point to a functioning ethics committee. Lucas: Do you know the details on that one? Luna: It was a customer data platform vendor. The retailer's AI governance policy requires that any vendor whose product touches customer data must have an AI ethics committee with at least one external member. The vendor had a policy document but no committee. They tried to argue the engineering team 'did ethics reviews informally.' The retailer didn't budge. Lucas: That's a stark example. And it's not just retailers. I've been tracking procurement language from financial services firms, healthcare systems, even some manufacturing companies. They're all adding variants of this clause. The typical wording sounds something like: 'Vendor must maintain a standing AI ethics committee that meets at least quarterly, with documented minutes and a process for escalating ethical concerns to executive leadership.' Luna: And that's a pretty detailed requirement. What does a 'good' AI ethics committee actually look like in practice? Lucas: Let me give you a concrete example. There's a mid-stage cybersecurity startup — I'll call them SecurEthos, not their real name — that won a competitive RFP earlier this year against two larger incumbents. One of the deciding factors was their AI ethics committee. They'd chartered it eighteen months before the RFP. The committee included the CTO, the head of product, a privacy attorney from outside the company, an ethicist from a local university, and a customer representative. They met monthly, reviewed every new AI feature before release, and had the power to delay or kill a feature if they flagged a risk. Luna: So it wasn't just a rubber stamp. Lucas: Exactly. And that's the key. Enterprise buyers are getting sophisticated at sniffing out paper tigers. If the committee only includes internal engineers and reports to the VP of engineering, it's not seen as credible. Buyers want to see independence. They want to see someone who can say no. The SecurEthos committee had already blocked one feature — a automated network scanning tool that could have inadvertently collected personal data without consent. That story came up during the RFP presentation. It proved the committee had real teeth. Luna: I imagine that kind of story is gold in a sales process. It's so much more powerful than a slide saying 'we value ethics.' Lucas: Absolutely. And it's not just about winning deals. There's a defensive angle too. I spoke to a general counsel at a mid-market SaaS company who told me that their AI ethics committee effectively saved them from a potential regulatory investigation last year. They had a feature that used AI to score customer churn risk. The committee reviewed the model and realized it was disproportionately flagging customers from certain demographic groups. They pulled the feature, retrained the model, and relaunched. If they hadn't had that committee, they could have been facing a discrimination complaint. Luna: So the committee is also a risk management tool. But I'm curious — how do buyers actually verify that a committee is real? They can't sit in on the meetings. Lucas: They ask for three things. One: the committee charter. Two: the meeting minutes from the last two or three meetings, with any sensitive info redacted. And three: a list of decisions the committee made — features approved, modified, or rejected. If the vendor can produce those, it signals a genuine process. If they say 'our minutes are confidential' or 'we don't keep formal minutes,' that's a red flag. Luna: And what about smaller vendors who can't afford to have a committee with external members? Is there a minimum viable version? Lucas: There is. For a startup with fewer than, say, fifty employees, a credible committee might include the CEO, the head of engineering, a customer success manager, and an outside advisor who can serve as the independent voice. The advisor could be a retired professor, a lawyer with AI expertise, or even a customer who's willing to participate. The key is that it's not just the founding team patting themselves on the back. And it meets regularly — at least quarterly — with an agenda and minutes. Luna: That's helpful. I think a lot of founders listening might be thinking, 'Okay, I need to set this up. But what's the biggest pitfall to avoid?' Lucas: The biggest pitfall is treating it as a checkbox exercise. I've seen vendors create a committee, meet once, and never meet again. Or they appoint people who have no relevant expertise. One vendor I know put their marketing director on the committee because 'she's good with people.' That doesn't fly. Buyers will ask around. They'll talk to references. If the committee isn't real, it'll come out. Luna: And once it's out, it's worse than not having one at all, because it signals that you're trying to game the system. Lucas: Exactly. It becomes evidence of a culture problem. On the other hand, a genuinely functioning AI ethics committee can be a powerful differentiator. In a market where every vendor claims to be 'responsible' and 'trustworthy,' the committee is a structural proof point. You can point to it and say, 'We don't just talk about ethics. We have a process.' Luna: Let's take a quick step back. I know this is a topic we're both passionate about, and honestly, if this episode was worth a coffee to you, there's a link — buy me a coffee dot com slash fexingo. It keeps the show ad-free and helps us keep digging into these procurement trends. Lucas: Yeah, listener support makes a real difference. We're able to stay independent and focus on what actually moves the needle for operators and builders. So if you find value here, that's the place. Luna: Alright, back to it. So we've established that the committee is becoming a requirement. But I want to push on one thing: what about vendors that don't use AI in any significant way? Do they still need one? Lucas: That's a fair question. If your product has zero AI — no machine learning models, no generative features, no algorithmic decision-making — then some buyers might waive it. But more and more, enterprise buyers are taking a broad view. If your product handles data that could be used to train an AI in the future, or if you have any kind of automation that makes decisions, they'll want the committee. And honestly, it's not a bad practice to have anyway. It shows you're thinking ahead. Luna: So it's becoming table stakes. What's the timeline for a founder who's listening and realizes they need to set this up? How fast can you spin up something credible? Lucas: I'd say you can put together a credible committee in four to six weeks. You need to identify the members, draft a charter, schedule the first meeting, and produce minutes. The charter should define the committee's purpose, scope, membership, meeting frequency, decision-making authority, and escalation path. There are templates available from groups like the IEEE and the Partnership on AI. But the real work is not the paperwork — it's building the habit of actually deliberating on ethical questions. That takes time and practice. Luna: And I'd add: start before you need it. Don't wait until you're in the middle of an RFP with a Fortune 500 customer. By then, it's too late to look authentic. Lucas: Great advice. One more thing I want to mention: the composition of the committee matters beyond just independence. Enterprise buyers are increasingly looking for diversity — not just demographic diversity, but disciplinary diversity. They want to see legal, product, engineering, customer-facing, and ideally an external voice. That mix is what produces robust debate. If everyone in the room thinks the same way, the committee is less likely to catch blind spots. Luna: That's a good point. And it ties back to something we've said before: procurement is becoming a mirror of governance. Buyers are essentially auditing the vendor's internal governance structures. An ethics committee is one of those structures. Lucas: Exactly. And I think we're only going to see this trend accelerate. As AI regulation matures — whether it's the EU AI Act, state-level laws in the US, or sector-specific rules — having a committee in place will become not just a sales advantage but a compliance necessity. The vendors that start now will be ahead. Luna: Let's leave listeners with a concrete takeaway. If you're a B2B SaaS founder, what's the one thing you should do this week? Lucas: Draft a one-page charter for an AI ethics committee. Even if you don't have the members yet, get the structure down. Decide who should be on it, how often it will meet, and what authority it will have. Then start recruiting. That document will be the foundation of your governance story. And when the next RFP asks, 'Do you have an AI ethics committee?' — you'll be able to say yes, and show them the proof.