How to choose an AI consultant
Hiring someone to bring AI or automation into your business is easy to get wrong, because the work is hard to judge from the outside. These are the eight questions we think you should ask anyone you are considering, including us.
- 1
Can I see something like this working?
A slide deck proves nothing. Ask to see a real system doing a real job, or a demonstration built to show the same thing.
A good answer: A working demo or live system you can click through. Named case studies only where the client has agreed to them, so expect some to be anonymous.
- 2
What will change, in numbers?
“More efficient” is not a result. A good consultant agrees a starting point and a target before building, then reports against both.
A good answer: A measured baseline, such as hours a week, cost per job or enquiries answered in time, and a report against it once the system is live.
- 3
Who owns what you build, and my data?
If the work stops, you should still have what you paid for.
A good answer: Your data is yours. What was built specifically for you is yours once paid in full. The platforms it runs on belong to their providers.
- 4
Which tools will you use, and who pays for them?
Some consultancies resell a tool and call it a solution. Others hide the running costs until after launch.
A good answer: Every tool named in the proposal, with its cost and who holds the account.
- 5
What happens when it breaks?
Automation changes as software updates and your business moves on. Someone has to notice when it stops working.
A good answer: Monitoring and a named person to fix problems, and notes thorough enough that someone else could take over if needed.
- 6
Where does a person stay in charge?
AI makes mistakes. For anything involving money, compliance, safeguarding or a customer’s records, a mistake should be caught before it counts.
A good answer: The AI proposes, a fixed rule or a person confirms, and nothing is written to your records unchecked.
- 7
Where does my data go?
If a system touches customer, staff or pupil information, the law expects a written agreement about how it is handled.
A good answer: A data processing agreement signed before any personal data moves, and a plain list of where the data is stored and which services see it.
- 8
Can I start small, and what does it cost to stop?
A large contract before you have seen results shifts all the risk onto you.
A good answer: A small first step, a price you see before you commit, and clear notice terms if you want to stop.
Warning signs
- They start with a tool before they have asked about your process.
- They promise a specific saving or revenue figure before they have seen how you work.
- The same product is offered to every kind of business.
- You cannot find out what it will cost to run after launch.
- You would have to hand over broad access to your systems without a written agreement.
How we answer these
We start from the outcome you need, not a tool. The first session is free and ends with a written list of what to fix, yours whether or not you hire us. If something is worth building, you see a clear price before you commit to anything. Ask us any of the questions above on the call and we will answer them plainly.