Check your readiness
Is your next project ready to build?
Find what needs attention before you commit time or budget. Answer six practical questions and leave with a brief to keep.
Is it clear what would make this project worthwhile?
You can name who benefits, what gets better and how you would tell.
An example
Prepare a useful client plan in one day, so the team can serve more clients without adding another review meeting.
Choose a first project that is useful on its own and builds something the next project can use.
Start with what the business needs to do
Imagine a firm whose clients need help long after their main project ends. Its experts cannot join every call or review every small change. The firm wants to offer ongoing help at a level more clients can use.
A first AI project could help clients turn an agreed strategy into their next plan of work. It would use the firm’s methods and the client’s goals. Experts would check the plan before it went to the client.
In a large business, the same idea could help regional teams use advice from a central team. The aim might be to open a new site or launch a service without starting from scratch each time.
These are possible uses, not results we are claiming. They give the assessment a clear purpose: find what stands between today’s work and that wider service.
Show how the first project leads to the next
For the service firm, the first project covers one type of client plan. It needs current client goals, agreed methods, access rules and a clear review step. Those are also things a wider service would need.
The table shows a made-up path. Each later step needs its own proof, budget and agreed plan.
| Step | What changes | What carries forward |
|---|---|---|
| First build | Prepare one type of plan for an expert to approve. | Agreed methods, up-to-date client records and a tested review process. |
| Next possible build | Help clients check progress and prepare a revised plan. | The approved plan, the same client records and the rules for asking an expert. |
| Wider service | Offer ongoing support to more suitable clients. | Work that has been tested, with clear costs, access and support. |
For a large business, the next step might be a second region. Check which rules stay the same and which need to change. A system that works in one team has not yet proved it can serve them all.
Check six things before committing to the build
Is it clear what would make this project worthwhile?
You can name who benefits, what gets better and how you would tell.
Have you agreed where the first piece of work starts and stops?
There is a small, useful job to change, with clear points where a person checks or decides.
Can the project use the information it actually needs?
The records and rules are current, someone can explain them, and permission to use them is clear.
Is it clear how this will fit into the working day?
You know where information comes from, where the result goes and who deals with a failed update.
Does someone own the work and have time to help it change?
A named person can make decisions. The people using and checking the work have time to try it.
Can you make a sensible decision about the cost?
Someone can approve the budget, including staff time and running costs. The main unknowns are visible.
Use this sheet to prepare a decision, not to launch a company-wide audit. Start with what you already know. Ask the right person to resolve a missing answer when it affects the decision.
You do not need to clean every file or map the whole business. Focus on what the first project needs and what a likely next project could share.
This is a planning sheet, not a certified score. A missing permission cannot be cancelled out by five other good answers.
Find the few gaps that could change the plan
In our example, the firm may have clear methods but keep them in senior staff’s heads. Ask those people to show how they handle real requests. Record the rules that change their advice.
It may also have several copies of each client plan. Someone must decide which version the system should use and keep that choice up to date.
In a large business, check early who can approve access across teams. Bring security, technology and buying teams into decisions that need them. Find out whether an existing system already does part of the job.
Give each serious gap one named person and a next action. That turns the assessment into work someone can finish.
Decide what is ready to move forward
You can plan a build when the goal is clear and someone is in charge of the work. Agree how to resolve the main unknowns within the project.
If the biggest doubt is whether clients want the service, speak to them before building it. If the doubt is whether AI can apply your rules, test that part first.
Keep later uses in view, but fund them when the evidence supports them. Reusing client records could reduce later work. It does not make a second service free.
How we help turn the ambition into a plan
Our AI Strategy Workshop helps your team compare ideas, work through the likely returns and choose where to begin. Bring the questions you still need help with. You do not need to fill in this sheet first.
We agree the build as a separate project. We lead the detailed mapping, data work, build and team onboarding for one agreed workflow in 90 days or less. That period starts when the build project begins.
Your team brings its knowledge, grants agreed access and makes the business decisions. We also discuss what the next project could reuse and the support needed after launch.
Further reading
- Lean AI: our book on choosing useful work for AI
- How we built a tool to sort journalist requests and draft replies
Written by Said Lopez, founder of LeverageAI and author of Lean AI. Our guides draw on the method in the book and our work with clients.