The problem is not an AI problem
A cheap off-the-shelf tool, a fixed spreadsheet or one changed procedure will solve a surprising number of the problems we are called about. We will point you at it and charge you nothing.
How we work
Five steps, fixed prices, no mystery
Every stage produces something you can use on its own. You can stop after any of them and still be better off than when you started — and you will always know what the next stage costs before you commit to it.
How does an AI consulting engagement work?
A typical AI consulting engagement runs in five stages: a free scoping conversation to establish whether there is a real problem worth solving; a process audit that measures the current cost and produces a costed roadmap; a fixed-price pilot build of the highest-value automation; rollout with staff training and documentation; and then either a clean handover or an ongoing retainer. Each stage has a defined deliverable, and the client can stop between any two.
1
Free scoping call
free
30 minutes
A conversation, not a pitch. You describe what is painful; we ask the uncomfortable operational questions — volumes, exceptions, who really does the work, what breaks today. We will tell you plainly if we think AI is the wrong answer, and quite often it is.
- An honest opinion on whether this is worth doing
- A rough cost range so you can judge whether to continue
- A recommended first step, even if that step is not us
2
Process audit
From KES 75,000
1–2 weeks
We shadow the actual workflow rather than the documented one, measure the baseline in hours and shillings, and score every candidate automation on value, effort and risk. The deliverable is a decision document, not a description of AI.
- A measured baseline: time, cost, error rate today
- A scored shortlist of opportunities
- A costed, sequenced roadmap you own outright
- A clear list of what we recommend you do not do
3
Pilot build
From KES 250,000
2–6 weeks
We build the highest-value automation end to end and put it into production against your live systems. Fixed price agreed before we start. We run it in shadow mode alongside the manual process first, so you can see it agree with your team before you trust it alone.
- A working automation, live, integrated with your systems
- Human review queues wherever money or commitments are involved
- Full logging and an audit trail from day one
- A measured after figure to set against the baseline
4
Roll out & train
Included
1–2 weeks
Moving from “it works” to “the team runs it”. We train the people who will use it daily, write the runbook in plain language, and stay close for two weeks of hypercare while the first real edge cases surface.
- Hands-on training for the people who use it
- A plain-language runbook: what it does, how to pause it, who to call
- Two weeks of hypercare after go-live, included
- Handover of code, credentials and documentation to your accounts
5
Improve or leave
From KES 150,000/month
Your call
Either we hand over completely and step away, or we stay on a monthly retainer to monitor, tune and build the next automation. Reviewed every quarter. Thirty days’ notice either way, and nothing technical depends on us remaining.
- Monitoring, tuning and model updates
- A defined number of build days each month
- Quarterly review against agreed measures
- A clean exit whenever you want one
What we need from you
- Access to the systems, documents and people involved.
- A decision-maker who can approve without a committee.
- 2–4 hours a week of your team’s time during the build.
- Honesty about what really happens, including the workarounds.
Projects stall on access, not engineering. If getting a password takes three weeks, tell us on the first call and we will plan around it.
Our commitments
PricingFixed before work starts
OwnershipCode and accounts are yours
Lock-inNone. 30 days’ notice on retainers
ReportingWeekly, in writing, whether or not it is good news
Scope creepOur risk on fixed-price work, not yours
On confidentiality
We will sign an NDA before the first working session. We work inside your accounts where possible. Anything we must copy is agreed in writing, with a deletion date.
Honesty
When we say no
We turn down work regularly, and we would rather tell you at the start than three months in.
The data does not exist
If the information the system would need has never been recorded, no model can conjure it. That becomes a data project first, and we will say so before you spend money on the AI part.
Nobody owns it internally
Automation with no internal owner drifts and dies within a year. If nobody in your organisation is accountable for the process after we leave, we would rather not build it.
The expected value is not there
If a process takes two hours a month, automating it will not repay the build cost or the maintenance. We will tell you that in the audit rather than sell you the build.
It requires unexplainable decisions
Where a regulator or a customer is entitled to a reason, a system that cannot give one is a liability. We design for explainability or we decline the use case.
The timeline is not real
If a system must be live in ten days across four departments, we will say it cannot be done well. We would rather lose the work than deliver something that breaks in production.
Questions
How we work — answered
What happens on the free call?
You describe the problem; we ask awkward questions. How many times a week does this happen? Who actually does it, and what else are they meant to be doing? What happens today when it goes wrong? Which systems hold the data? By the end of thirty minutes you will have our honest opinion on whether this is worth automating, roughly what it would cost, and what we would do first. That opinion is yours whether or not you hire us.
Do we have to do the audit before you build anything?
No. If you already know exactly which process is hurting and you can give us access to the systems involved, we can go straight to a pilot build. The audit earns its place when there are several candidate processes and you need to defend the choice, or when a previous attempt failed and nobody is sure why.
What do you need from us?
Three things, and they are the three that determine whether a project succeeds. First, access — to systems, to documents, to the people who actually do the work. Second, a decision-maker who can say yes without convening a committee. Third, roughly two to four hours a week of your team’s time during the build. Projects that stall almost always stall on access, not on engineering.
What if the pilot does not work?
Then we tell you, and you stop. We build a small first project precisely so that failing is cheap and fast. In practice the more common outcome is that it works but reveals the real problem was somewhere else, in which case you have learned something valuable for a small fraction of a full programme budget.
Are we locked into a contract?
No. Fixed-scope projects end when they are delivered. Retainers run month to month with 30 days’ notice on either side and a quarterly review. You own the code, the accounts and the documentation from day one, so leaving is always practical — which is the only thing that makes staying meaningful.
How do you handle our confidential data during a project?
We sign an NDA before the first working session if you would like one. We work in your accounts wherever possible rather than copying data to ours. Where a copy is unavoidable — a sample of documents to build extraction against, for example — we agree in writing what is copied, where it lives, who can see it, and when it is deleted. If you have obligations under the Data Protection Act, 2019, we document the processing arrangement in a form your compliance team can file.