Build or buy AI: how to decide
When an off-the-shelf AI product is the better choice, when custom AI software is, and how to settle the question for a specific problem.
AI strategy
Deciding where AI should go first
Most companies aren't short of AI ideas. What they lack is a clear view of which idea is worth building first and what it would take. Our strategy work answers that. We spend time with the teams involved, look at where hours and money go, and come back with one well-defined project: what it should do, what it needs, roughly what it's worth, and how we'd build it.
When this makes sense
- 01
The leadership team wants an AI plan, and nobody agrees on where to begin.
- 02
There's a long spreadsheet of AI ideas and no sensible way to rank them.
- 03
You've tried an off-the-shelf AI tool and it didn't match how your people work.
- 04
One idea looks promising, but nobody knows if the data exists or what it would cost.
What's involved
-
Time with your teams
We sit in on the work itself, interview the people who do it, and look at the systems they use. Workshops are useful, but people describe a process differently from how they run it.
-
Where the time goes
We map the delays, rework and manual steps we find and estimate what each one costs. Then we judge how well software could deal with each.
-
A short list
We rank the candidates by value and by how hard they'd be to build, including whether the data is there, and agree the first one with you.
-
Build or buy
If a product on the market already does the job, we'll tell you. Building only makes sense when nothing available fits.
-
A plan to build from
For the chosen project we write the requirements, sketch the design, list the data and systems involved, and set out cost and timing.
What you get
Every project follows the same five stages, and each stage ends with something you can review.
- Map of where time is lost
- Ranked list of opportunities
- Requirements for the first project
- Design, budget and plan
AI strategy: common questions
What do we get at the end of an AI strategy engagement?
A ranked view of where software could help, a first project chosen with you, written requirements for it, a design and a business case. The aim is a plan detailed enough to start building from straight away.
How do you pick the first AI use case?
We weigh what the problem costs today against how hard it would be to solve. Data matters a lot here: a valuable idea with no usable data behind it usually isn't the right place to start.
Will you always recommend building something?
No. Plenty of problems are already solved well by existing products, and in those cases buying is cheaper and faster. We'd rather say so up front.
Further reading
How to scope an AI use case
A practical way to scope an AI project: watch the work, choose one problem, write requirements and success measures, then test a prototype.
Not sure where to start?
Tell us what you're working on. We reply within two business days.
Talk to us