Why is the memo free?

Because it is the cheapest honest way to show you how we think, and because the portfolio work afterwards is where the commercial relationship sits. Roughly one in four memos leads to a further conversation. The other three cost us two days and gave you something useful, which is a trade we are happy with.

Does the company we are evaluating find out?

Only if you tell them or if we need to ask them something, and we ask you before contacting anyone. Where the memo is written from public artefacts alone it says so on the first page, because a partner reading it needs to know how much weight it carries.

How is this different from a diligence firm?

Scope and speed, deliberately. A diligence firm gives you thorough coverage in two to four weeks for a fee. This gives you the technical questions that change a decision, in two days, for nothing, with the gaps marked. They are different instruments and a fund running a competitive process usually needs this one first.

Can you sit on a technical advisory basis across the portfolio?

Yes, and several funds prefer that shape: a standing arrangement covering memos on live processes plus a set number of portfolio engagements per year. It is scoped per fund rather than priced on a page.

Do you work with accelerators as well as funds?

Yes. The accelerator shape is usually a cohort session plus short engagements with the two or three companies where AI is central to the product, rather than one memo at a time.

VCs and accelerators

The deck says the platform is proprietary. Somebody has to open the repository and check.

For funds and programmes that need a technical read before the cheque, and hands-on help for the portfolio afterwards.

Work delivered for

Singapore Government project

$100Mroughly in annual operational value at DHL scale, from the programme AISynq contributed to. A figure that size is never one person’s work.Read the DHL case study

The problem

Technical diligence sits in an awkward place in a fund. It matters most on the deals moving fastest, it requires a skill nobody on the investment team has to keep sharp, and the partner who could evaluate it properly is on three other processes this month.

So it gets handled two ways, both unsatisfying. Either a friendly CTO from the portfolio takes a look as a favour, which produces a genuine opinion with no time behind it and no accountability for being wrong. Or a full diligence firm is engaged, which produces a thorough report two weeks after the round closed.

The questions that matter are rarely the ones a checklist covers. Whether the impressive-sounding model is a thin wrapper around an API anyone could call. Whether the two engineers who built everything have written anything down. Whether the infrastructure bill grows linearly with users or worse. Those are answerable quickly by someone who has built the same things, and slowly by anyone else.

This is for you if

You are evaluating a company where the technical claim is material to the valuation.
You need a read in days rather than weeks, because that is the shape of the process.
You would rather have a short honest memo with its limits stated than a long one that implies more coverage than it has.
You have portfolio companies who need hands-on AI help after the round, not only a diligence opinion before it.

This is not for you if

You need a full security audit or a compliance attestation. That is a different discipline and a different firm.
You want a report that supports a decision already taken. We write what we find.
The target will not permit any technical contact at all. There is a version of this from public artefacts alone, and it is considerably weaker, which the memo would say.
You are evaluating deep hardware, biotech or quantitative trading infrastructure. Outside what we can read properly in 48 hours or at all.

The method

The method, in your situation

01IdentifyThe memo itself. What the technical claim actually is, what the codebase and the team support, where risk sits that the model has not priced, and what the first six months of engineering will cost in headcount and infrastructure. Written to be read on a phone before a partner meeting.
02BuildAfter the round, for portfolio companies. The same engagement described on the startup page: find the AI worth building inside the company, build it into their stack, hand it to their engineers. Funds often want this for two or three companies in a cohort rather than one.
03ProveA baseline taken before the build and read again afterwards, which is as useful to you as it is to the founder. A portfolio company that can show a measured change in activation or unit cost is easier to mark up honestly and easier to introduce to the next round.

What you get

What you end up holding

Artefacts rather than activities. Each of these is a thing that exists after we leave.

The 48-hour memo
Three to six pages. What the claim is, what supports it, what does not, what it will cost to build the next six months, and a clearly marked section on what 48 hours could not cover.
A question list for the founder
The five or six questions worth asking directly, phrased so the answers are checkable. Several partners have said this is the part they use most.
A cost projection
Engineering headcount and infrastructure spend for the next two quarters, against the growth assumptions in the deck rather than against ours.
Portfolio engagements
For companies you have backed: the opportunity note, the build, and the result note, exactly as described on the startups page.

The entry offer

A technical due diligence memo in 48 hours, on a company you are evaluating right now.

In 48 hours we can tell you whether the codebase matches the story, where the team is carrying risk it has not priced, and what the first six months of engineering will actually cost.

48 hours is not enough to audit security properly, so if that is your concern, say so on the form and we will scope it separately. The memo marks its own gaps on the first page rather than letting a partner assume coverage it does not have.

No fee and no obligation. The clock starts when the form arrives, including at the weekend, because that is when a competitive process usually needs it.

Getting started

How engagements start

01You send the company name and whatever access you have, using the form on this page. The 48 hours start when it arrives, not on the next working day.
02If we need read access to a repository or 20 minutes with their CTO to answer something properly, we ask within the first few hours rather than at the deadline.
03The memo lands by the deadline. There is no fee and no obligation, and roughly one in four leads to nothing further, which is a fine outcome.

If one of these is you

Your cohort is full of AI companies. How many of them have built anything?
For accelerators and programmes running an AI-heavy intake, where the partners are technical enough to know they cannot check twenty companies themselves.

What 48 hours cannot do

It is not enough time to audit security properly. It is not enough time to assess a codebase over about 200,000 lines with any confidence. It is not enough to evaluate a machine learning system whose quality depends on proprietary training data we cannot see. If your concern is any of those, say so and we will scope it separately rather than delivering a memo that quietly implies coverage it does not have. The honesty about scope is the persuasion here, not a weakness in it.

FAQ

Questions from people in your position

Next step

Send the company you are looking at now. The memo lands in 48 hours, with its own limits printed on it.

Request the memo

No fee. No obligation. Roughly one in four leads to anything further.