On your side of the table.
Five services. Each of them a judgment on something before you commit to it.
An agency, a chief technology officer, a software platform, a company you are acquiring: finding the candidates is rarely the difficulty. Judging them is. And nearly everyone equipped to help you judge — the vendor's solutions engineer, the search firm paid on placement, the integrator hoping to do the build — has an interest in which way you decide.
Our independence is a matter of structure rather than assurance. We are paid by the person who asked for the advice and by nobody else, and we do not compete for the work that follows from our own recommendation. We build software for ourselves alone: enough to keep the judgment current, never enough to make us a candidate.
We take few engagements and hold them for years. The ones that reach us tend to be the awkward ones — a decision with no obvious precedent, a business nobody else will examine properly, a question that fits no one's framework.
Most engagements begin with one and grow into two.
Five services.
01 — AI advisory
Whether to buy it, which one, and who is accountable when it is wrong.
Read more →
02 — Technical procurement
Deciding what to buy, before a vendor decides it for you.
Read more →
03 — Technical due diligence
What you are actually buying, in plain language.
Read more →
04 — Forward deployed engineering
We help you pick the right ones, and we are not one of them.
Read more →
05 — Technical leadership due diligence
The technical read on the person you are about to make CTO.
Read more →
Three movements, whichever service you start with.
How an engagement runs.
01
We learn what you are deciding
A week or two inside the business before we assess anything: what you are buying, what it must do, and what breaks if it is wrong. Engagements change shape at this stage more often than not, which is the point of doing it.
02
We examine the thing itself
We read the code, meet the candidate, and speak to the vendor's last three customers — including the one who left. Not the deck, and not the references the seller selected for you.
03
We give you a straight answer
Ranked and written, in language you can act on without an engineer to translate it, and including the questions we could not settle. Some engagements have run for years. One ended in three weeks, because the answer was no.
The record.
Since 2013 we have built, backed or advised 21 companies, 7 of which have exited. We have sat on the other side of this table as often as this one. The portfolio is public, including the ones that did not work — see the full portfolio.
Common questions.
- Do you build software for clients?
- No. We assess, we advise, and we sit in the room while you decide. The only software we build is our own products, which is what keeps the judgment current. It also means we are never a candidate for the work we just helped you scope.
- Do you take a fee from the vendors or providers you recommend?
- Never. No commission, no referral, no rebate, no reseller agreements. You pay us and nobody else does. A recommendation you have to discount for someone's incentives is not worth buying.
- How is this different from a consultancy?
- Most consultancies that assess your options also want the implementation. We have no implementation to sell. Thirteen years of building and operating is what we bring instead of a methodology.
- How many engagements do you take at once?
- Few. The work depends on senior people holding real context, and context does not survive being split five ways. When we are full we say so.
- Are we too small for this?
- We have worked with pre-seed founders and with companies processing billions in transaction volume. What matters is that a real decision is on the table and you can make it.
- Do you also invest?
- Yes, separately, at pre-seed and seed. We decline assessment work where we hold a position, and we tell you about a conflict rather than manage it quietly.
- What does it cost?
- Depends on the shape and the length. We would rather scope it properly than quote a range that means nothing. Tell us what you are deciding.
Where to next
Tell us what you are deciding.