Technical Advice & Architecture Review
Decide first. Then choose a builder.
One decision, framed in writing, with the recommendation, the constraint that forced it, and the signal that would reverse it. No slides, no retainer, and no project attached.
You leave with the decision framed and the evidence that would settle it.
The decision memo, heading by heading
One decision stated as something either done or not done, every option considered including doing nothing, one recommendation, and the observable signal that would make it wrong.
Advice with no reversal signal cannot be argued with, only believed.
- The decisionOne sentence, in the words your board would use, stated as something that is either done or not done. Half of all stuck decisions are stuck because nobody wrote down what is being decided.
- The options consideredEvery option genuinely on the table, including the ones I would not be paid to deliver, and including doing nothing. An options list with one entry is a proposal wearing a memo’s clothes.
- The recommendationOne option, chosen, with the reasoning that chose it. Not a ranked shortlist, which is how an adviser hands the decision back while still invoicing for it.
- The constraint that forced itThe binding one: a date, a budget, a regulation, a system you cannot change, a skill you do not have. If the recommendation would flip under a different constraint, that flip is written here.
- The signal that would reverse itA named, observable thing that, if you see it, means this recommendation has become the wrong one. Not “we will review quarterly”, which is a refusal dressed as a process. This is the line the rest of the category does not write.
- The ownerThe person inside your company who owns the outcome, by name. A decision with no owner is a document, and documents do not ship.
- The dateWhen it was decided and on what evidence, so that in nine months nobody reconstructs the reasoning from memory, or worse, from whoever is still in the room.
What the memo arrives as
The container changes with the question. The seven headings do not.
Architecture review
A standalone independent read on a system or a plan, with each default stated alongside the condition that would overturn it. No build attached, and none proposed.
Technical due diligence
Someone else’s asset: an acquisition target, an incoming vendor, an investment. Deal-stoppers separated from fixable findings, with a cost of fixing attached to the second list.
Second opinion
A proposal you cannot judge, a build-versus-buy call, or a direction your own team is split on. Includes the case for neither of them yet, which is a real answer and sometimes the right one.
- A written decision memo in the seven headings above
- The options that were rejected, and the reason each one lost
- The named signal that would reverse the recommendation
- A conflict declaration written into the memo itself wherever the recommendation points at work ViitorCloud could win
- The questions to put to any vendor before you sign with them
What I am commercially interested in
Every adviser in this category has an interest. The useful signal is not the absence of one, it is whether it was declared before you found it.
I am co-founder and CTO of ViitorCloud, which builds software for money and could be paid to build yours. That work is sold in two named shapes: Embedded Engineer and Scoped Delivery. When a recommendation points at building something, you can weigh what it would be worth to me instead of guessing.
What is not on that list is my own time. My hours are not a unit anybody buys, so no recommendation ever resolves to more of me.
Three rules contain the rest. The memo is written to be read by an engineer who has never spoken to me, so two other firms can quote against it. Delivery is a separate decision you make afterwards. And where the conflict is not containable, the answer is that I am the wrong adviser — and you hear it on the first call rather than in month two.
One limit, stated rather than hidden. I cannot show you a filled-in client memo, because client work is not published on this site. On that axis a competitor with a named client and a published case study beats this page, and pretending otherwise would fail the test I just set.
Questions people ask first
Is this a sales call for your development company?
It is a call with the CTO of a company that builds software, so treat the question as live rather than settled. What contains it is a rule published on this page rather than a reassurance: the recommendation is written down and portable, delivery is a separate decision you take afterwards, and the conflict is declared in writing inside any memo whose recommendation points at work ViitorCloud could win. Where the conflict is not containable, the correct answer is that I am the wrong adviser, and you hear it on the first call.
What is a fractional CTO, and is that what I need?
A part-time senior technical leader, typically one to three days a week over months. It is a model built for a stream of decisions. Most people who arrive here have one decision with a date on it, and for that a bounded piece of work with an artefact at the end is cheaper, faster and portable. If there really is a continuing stream of consequential decisions and nobody inside who can own them, then the retainer is the right shape and I will say so.
What do I actually receive at the end?
A written decision memo in the structure published on this page. Depending on the question it arrives as an architecture review, a due diligence report or a proposal evaluation. The container changes; the seven headings do not. It is written for an engineer who has never spoken to me, because a document that only works while I am in the room is not advice.
Do you take equity instead of fees?
Not for advisory work. Somebody holding equity in your company has a stake in which direction you choose and in how much you spend getting there, which is precisely the interest an independent read is supposed to be free of. If what you actually need is a technical co-founder rather than an adviser, that is a different arrangement with different incentives.
Bring one decision
Not the whole roadmap. The single call you are stuck on, with a date attached and a cost if it goes the wrong way. You leave with it framed, and with the evidence that would settle it.
In the age of AI
The advantage was never the model. It's knowing what to build with it — and having a team that can actually ship it.
That's the part I help with: finding where AI genuinely makes your business faster, deciding what's worth building, and standing behind it once it's live.
Four offices, one very full passport
Every dot on this map is a conversation I still remember.
- Where I've spoken
- Office
