AI Solutions · Tailored Builds
When nothing off the shelf matches how you actually work.
Governed AI is the category, and a tailored build is one of five ways an outcome inside it gets delivered: a scoped design and build, running in your own environment, for a problem no product solves. You own the result. It is also the slowest and most expensive of the five options — so this page is as much about when not to do it as when to.
The boundary
We are Microsoft-centric about your environment and deliberately open about what we put to work inside it.
A tailored build is the fifth of five ways an outcome gets delivered with governed AI, and it is the right one only when nothing off the shelf matches how the business actually works. That is a real condition, not a sales qualifier. If the work lives in documents, mail and meetings, option one is cheaper and faster. If the problem is a process with steps and handoffs, option four is. We would rather tell you that during the assessment than six weeks into a build.
All five are governed AI. The option is chosen against the shape of the problem, never against a vendor relationship.
- 01Copilot
- 02Company AI
Accounts - 03Connected
Assistants - 04Workflow
Platforms - 05Tailored
Builds
The category
A category, not a single product.
A quoting tool, an operations portal, a knowledge system — all tailored builds, each scoped against the business that needs it. One of them we have packaged, and it is described below. The rest are scoped and quoted one at a time.
Whichever it is, the shape of the engagement does not change: one outcome delivered with governed AI, a baseline measured before anything is switched on, a named owner, and the result published.
When it is right
When how the business delivers is genuinely unusual, and the gap between what you need and what you can buy is worth paying to close.
When it is not
When one of the first four gets you most of the way. This is the slowest and most expensive of the five, and the honest answer is usually one of the others.
The honest part
We will tell you during the assessment if we think you are reaching for this when option one would do. Six weeks into a build is a bad time to find out.
Who sells what
Scoped design and build, priced on the outcome rather than on seats.
The gap between what you can buy and what the business needs
The first four options cover most of the distance. What remains is a gap, and a tailored build is only worth it when that gap is wide enough to pay to close.
- What the first four reachMost of the distance, most of the time
- The gapOnly worth paying to close when it is wide
- How the business actually deliversThe thing you were trying to reach
A build answers a gap, not an ambition
If one of the first four gets you most of the way, the gap is not worth a build. We would rather establish that in the assessment than six weeks in.
The one packaged build
Answer Desk
The problem is not that nobody knows. It is that nobody can find out.
In most firms the answer exists. It is in a procedure nobody can locate, a project from four years ago, or the head of someone who is in a meeting. Answer Desk is one place where a question gets an approved answer, or reaches the person who has it — and where that answer is captured so the next person does not have to ask.
Built for organizations where knowledge lives in people’s heads and finding it costs real money — professional services, engineering, architecture, inspection, anywhere the expensive people are the bottleneck.
How it works
One question. Six steps. Then it never costs that much again.
1Ask
Someone asks in the place they already work. No new habit, no new tab.
2Answer, or route
If the answer is in the approved material, they get it with its source. If it is not, the question goes to the named expert for that subject.
3The expert responds
Once. In a queue they can see and clear, not buried in a thread.
4Review
The answer is approved before it becomes canonical. Sensitive subjects can require a second reviewer.
5Published
The approved answer joins the knowledge base. The next person asking gets it immediately.
6Measured
What was asked, what went unanswered, where the gaps are, and what it cost.
The expensive step happens once. Today, the same question reaches the same expert four times a year and is answered from scratch each time. The loop is worth building because step three stops repeating.
Where it lives
It meets people where they already are
People will not adopt a destination. They will adopt an answer that arrives where they are already working, which for most organizations means Teams.
The intelligence runs underneath in Microsoft Foundry, inside your own Azure tenant. What people see is the place they already open every morning. Where Copilot is already in use as a personal assistant, it stays exactly that — this is not a replacement for it, and the two do not compete for the same job.
The experts need somewhere to answer
The people who hold the answers are usually the busiest people in the building, and the design question that matters most is how little of their time this takes.
A queue they can see and clear is faster to build and cheaper to run. Routing everything through chat is possible, and it costs more time and more money to deliver. That choice is made during discovery, with the experts in the room, before anything is priced — because getting it wrong is the difference between a system people use and a system people resent.
The commitments
What is settled before the first line is written.
It lives in your tenant
The build runs in your Azure environment. Your data does not move somewhere else to be useful.
You own what we build
The intellectual property is yours. If you want to run it without us in two years, you can, and we will hand over the documentation to do it.
The answers are approved, not scraped
It draws on a curated set of sources that someone signed off, not on every working file and draft version in the organization.
Priced on usage, not per seat
You pay for what people actually use rather than a license for everyone who might.
A provider who becomes indispensable has failed at the job. The handover documentation is written during the build, not promised for later.
How we start
The first version is for learning, not for launching.
Two ways to start a build
Switching the whole vision on at once leaves you unable to tell which part was wrong. One thing for one group answers a single question, and informs what comes next.
Build the whole vision, switch it on
Nobody can tell which part was wrong
Everything is live at once, so nothing is attributable
One thing, one group, one question
What comes next is informed, not guessed
Do people actually use this? The answer shapes the next build
If people use it, the shape of the next build is known. If they do not, you learned that early — and that is a good outcome too.
The temptation with a system like this is to build the whole vision and switch it on. The reason not to is simple: when everything is live at once and people do not use it, you cannot tell which part was wrong.
So the first build does one thing for one group, and it answers one question — do people actually use this? If they do, the shape of what comes next is informed rather than guessed. If they do not, you have learned that for a fraction of the cost, and that is a good outcome too.
Scope, sequence and the measure of success are set in discovery, with a baseline taken before anything is switched on. No pilot advances on enthusiasm alone.
Begin with the truth
What we know, and what we must prove.
What we know
- The architecture is built and demonstrated: the question loop, the expert queue, the approval step, the admin view, and the usage and cost reporting.
- It runs in the client’s own Azure tenant, and the client owns the result.
- It reaches people inside Microsoft Teams, where they already work.
- Usage-based pricing is how Azure consumption works, not a claim we are making about it.
What we must prove
- That people choose it over asking a colleague. We believe they will when the answer is faster; we have not yet measured it in production.
- That subject matter experts answer through it consistently once the novelty passes.
- What the approval step costs in practice, in reviewer hours per month.
- The total cost of ownership over a full year, not a month.
- We have no published result yet. When we have one, it will appear here with a number attached.
Where we come in
Advice, then the work, then the handover.
Whichever of the five is chosen, the shape of the engagement is the same: one outcome, a baseline measured before anything is switched on, a named owner, and the result published. What changes is the scope and the price.
- 01
Decide whether it is actually your option
The assessment establishes what you already own and pay for, what your people are already using, and which single outcome is worth doing first — with the baseline taken and the owner named before anything is switched on. If one of the first four gets you most of the way, we say so here. Six weeks into a build is a bad time to find out.
- 02
Do the work
One thing, for one group, scoped in discovery and built in your own environment. The first version answers one question — do people actually use it? Measured against the baseline, because an outcome nobody measured did not happen.
- 03
Hand it over
The result is published, and the documentation to run it without us is written during the work rather than promised for later. A provider who becomes indispensable has failed at the job.
If you are a technology partner
You own the client relationship, the contract and the invoice. We are the specialist capacity behind it — the method, the assessment, and the delivery depth you would otherwise have to hire. You decide whether we are named in front of your client, and whether we are in the room at all.
If you are a business
Where a technology partner supports you, we work through them. Where one does not, we work with you directly and say so plainly.
Where to start
Start with the evidence, not the build.
An assessment establishes what you already own, how identity, devices and permissions are configured today, read-only against your own tenant, what AI is already in use, and which single outcome is worth doing first. A tailored build is the right answer sometimes. The assessment is how you find out whether this is one of those times.
Technology partners begin with the MSP Secure AI Practice Assessment. Businesses begin with the Secure AI Operations Assessment, offered through your technology partner where one is in place.