Frequently Asked Questions

If you're backing a company that uses AI - calling LLM APIs or building features on top of foundation models - SaaS Due Diligence (Enhanced tier) covers it, including an applied-AI review. If the company builds AI - training or fine-tuning its own models, running its own evals, or claiming a model-based moat - Deep AI Due Diligence audits the full model lifecycle. Not sure which fits? Book a call and we'll point you to the right one.

Every engagement covers code quality and architecture, cloud and infrastructure, scalability, and security and compliance. SaaS DD adds an applied-AI review (the LLMs and harness in use, AI security, and evals); Deep AI DD goes deep on datasets, training, evaluation, and MLOps. You receive a board-ready scorecard, a prioritized red-flag list, and a remediation roadmap.

Not on its own - plenty of strong products are built on third-party models. But the label means little: roughly 40% of European "AI" startups have shown no evidence of material AI use. The real questions are defensibility and exposure - is there a proprietary data advantage, what's the switching cost, and how concentrated is the dependence on one provider? With the cost of model access falling over 280x in 18 months and AI gross margins (50-60%) trailing classic SaaS (70-90%), a thin wrapper rarely has a durable moat. We assess exactly that, and flag cost and lock-in risks.

Sources:(MMC Ventures, State of AI, 2019)(Stanford AI Index, 2025)(a16z, 2020)

Yes. We look at inference and token cost per request, how it scales with usage, and the impact on gross margin - a common blind spot in AI-heavy products, where gross margins (50-60%) often trail classic SaaS (70-90%). In Deep AI DD this extends to serving efficiency and the cost curve as the model and traffic grow.

Sources:(a16z, 2020)

Largely, yes. We cover modern web, mobile, and cloud stacks (AWS, GCP, Azure) and the common ML and AI tooling. If a target uses something unusual, we'll tell you up front whether it's in scope when we scope the engagement.

SaaS Due Diligence typically takes 7-10 calendar days from dataroom access. Deep AI Due Diligence is scoped per engagement, since model audits vary with the stack and stage - we'll give you a timeline when we scope it.

Yes - the service is built for seed-round pace. If you're up against a closing deadline, tell us and we'll confirm what's achievable; expedited turnarounds are often possible.

Typically read-only access to the code repository and cloud environment, access to the dataroom, and a short block of the founder or CTO's time. We work from least-privilege access and return or revoke it at the end of the engagement.

Light - usually a kickoff and one technical interview, plus answering a few follow-up questions. We design the process to be low-friction for the target's team.

A board- and IC-ready report: a risk scorecard, a prioritized red-flag list, and a remediation roadmap, followed by a live debrief call. It's written to drop straight into your investment memo.

We quantify severity and give you the context to decide - a red flag isn't automatically a deal-breaker. You get a clear remediation path and effort estimate, so you can price the risk, negotiate terms, or set post-investment conditions with full information.

That's your call - the report is yours. We conduct the founder interview professionally and frame findings as actionable, so if you do choose to share it, it reads as a roadmap rather than a hit list.

Yes. We're happy to sign your NDA or provide our own, and everything we see is treated as confidential and used only for your engagement.

We act as an independent fractional-CTO lens for the investor. Our incentive is an accurate, honest assessment - not selling downstream engineering work - so our findings stay objective.

With least-privilege, read-only access wherever possible, secure handling throughout, and revocation of access when the engagement ends. We don't retain source code, and data is held only as long as needed to deliver and support the report. See our Privacy Policy for full details.

We scope each engagement to the deal, so we don't publish fixed prices. Book a short call and we'll give you a clear quote based on the company's stage, stack, and the depth you need.

Yes - beyond a one-off review, we can provide continuous portfolio monitoring on a retainer, so you keep a live read on technical health across your portfolio. Ask us about it on your intro call.

Book a 15-minute intro call. We'll confirm which service fits, sign an NDA, and agree access to the dataroom and systems - then the clock starts.