01
An initial conversation
A free call to understand the problem and tell you honestly whether it’s worth pursuing, and what the next step would be.
Book a callAI consultancy · UK
AI consultancy.Find the right problem to solve.
Before you build anything, get clear on where AI could make a real difference, what it would take and whether a custom build, an existing product or no AI at all is the better answer. Practical advice from the engineer who would build it.
Free first call · Fixed price after scoping · Sometimes the answer is a simpler tool
Where could AI make a real difference to our operations team?
- Review processes and volumes
- Check data and systems
- Compare buying with building
- Estimate cost, value and risk
Three opportunities ranked. One recommended first version.
Costed scope and risks documented for your decision.
Advice from the person who would build itHonest about when not to buildA costed plan you can act on
Is it right for you?
Where consultancy
earns its keep.
Consultancy is useful when the right next step isn’t clear yet. If it already is, I can go straight to scoping a build.
A good fit when
- You believe AI could help but aren’t sure where to start
- A tool or pilot has stalled and you want to understand why
- You need to compare an existing product with a custom build
- You need a costed plan before committing budget
Probably not yet when
- You want a general presentation on AI strategy rather than a practical plan
- You already have a clearly defined build. I’d go straight to scoping
How I can help
From a first conversation
to ongoing engineering.
Start where it makes sense. Each step is agreed separately, so you only commit to what’s useful.
02
Discovery and feasibility
A defined, paid piece of work for when the uncertainty is real. I inspect the data and systems, test the hard parts where needed and agree success measures.
03
Build or buy assessment
A clear comparison of existing products against a custom build for your situation: fit, cost, risk and what each would need to succeed.
04
Ongoing AI engineering
An experienced AI engineer on an ongoing basis, from two days a month: support, monitoring and evaluation (often called LLMOps), improvements and new capabilities for the systems you rely on.
The level of cover, response expectations and responsibilities are agreed in writing.
What discovery produces
Leave with decisions,
not a slide deck.
Discovery is written up so you can act on it, with me or with anyone else.
The problem
A clear definition of the problem and how success will be measured.
The data
What data and systems are involved, their quality and how they can be accessed.
The options
Build, buy, combine or wait, with the reasoning behind the recommendation.
The scope
A costed first version with acceptance criteria and estimated running costs.
The risks
What could go wrong, and how testing, review and permissions will manage it.
The kind of advice you get
Practical, specific
and written down.
Advice is only useful if it survives contact with your real data and systems.
Engineering-led
Recommendations come from the person who would build the system, grounded in what’s feasible with your data and tools.
Honest about “no”
Sometimes the right answer is an existing product, a process change or waiting. I’ll say so.
Evidence over enthusiasm
Where feasibility is uncertain, I test the hard part with your real examples before recommending a build.
Yours to keep
Findings are documented clearly enough to use with any supplier, not only with me.
Ownership and running costs
Clear scope.
Clear price.
Deliverables
What you receive from discovery
A written record you can use to make the decision.
- Problem definition and success measures
- Data and integration assessment
- Build, buy or wait recommendation
- Costed scope for a first version
- Risk assessment and mitigation
Costs
How it’s priced
Each stage is agreed before it starts.
- Initial conversation: free
- Discovery: a fixed price, agreed from its scope
- Build: a fixed price for the defined first version
- Ongoing engineering: scoped separately, if you want it
How a project runs
Map. Shadow. Own.
A focused first version. Tested in the real world. A clear path to what comes next.
01
Map.
I talk through the business, the problem and the systems involved, then agree whether discovery is needed and what it should cover.
- Initial conversation
- Agree the questions
- Scope discovery
02
Shadow.
I study the process as it really runs, inspect the data and test anything uncertain against real examples.
- Process review
- Data inspection
- Feasibility tests
03
Own.
You receive a written recommendation and costed plan. What happens next is your decision.
- Written findings
- Costed first version
- Your decision
Do I need my data in order before we talk?
No. Bring what you have. Part of the first conversation is working out what the data needs to look like, and whether tidying it is worth doing before any build.
What does a feasibility study give me?
Decisions, not a slide deck: what is feasible, what is risky, what a sensible first version looks like and what it would cost to run. It is written so you can act on it with any supplier, not only with me.
Do we need discovery before a build?
Not always. If the problem, data and systems are clear, I can go straight to a costed scope. Discovery is for when there’s real uncertainty: unknown data quality, a difficult integration or an unproven approach.
Is the first conversation free?
Yes. It’s a chance to understand the problem and for me to be honest about whether it’s worth pursuing. Book a call.
What if you conclude we shouldn't build anything?
Then I’ll say so and explain why. The findings are still yours, and a clear “not yet” can save far more than it costs.
Can you help us choose between off-the-shelf AI tools?
Yes. A build-or-buy assessment compares products against your requirements, including what it would take to make each one work in practice.
Do you offer ongoing or fractional AI engineering?
Yes. I can provide ongoing engineering for AI systems you rely on, from two days a month, covering support, evaluation, improvements and new capabilities. The level of cover and responsibilities are agreed in writing.
Do you only work with businesses in Hull?
No. I’m based in Hull and work with businesses across Yorkshire and the UK, mostly remotely.
Related
Other ways
I can help.
A useful first conversation
Not sure where AI fits yet?
Bring the problem, the idea or the process. I’ll talk through what’s possible, where the risks are and a sensible first step, even if that step isn’t a build.
Discuss your AI project Read the free RAG guide Or email hello@niroxai.comIndependent AI engineering. Based in Hull & East Yorkshire, working across Yorkshire and the UK.