5 Questions to Answer Before Your Organisation Touches AI

For years, the goal has been a single source of truth.
AI changes that.

• Marios Chrysanthou
• Managing Director
• Bluelight

Every week, our consultants are asked some version of the same question: What should we be doing with AI?

It is a fair question. It is also, in my experience, rarely the right place to start. AI is not a tool you can hand to someone and expect them to create magic. It is a cultural and mindset shift for an organisation, and it only delivers when the thinking has been done first.

That thinking comes down to five questions. If you can answer them clearly, there is a conversation worth having. If you cannot, the most responsible thing we can do is say so.

1 - What problem are you trying to solve?

This sounds obvious, but it rarely is.

In discovery sessions, we regularly hear requests that are more aspiration than problem statement.

"Can we have a chatbot?" is not a problem. 

"We receive a thousand enquiries a day by email, and we cannot keep up" is a problem with a shape, a scale, and a measurable outcome.

The distinction matters. When you can describe the challenge in a single sentence, with a scale and an outcome attached, there is something concrete to work with. When the request is closer to an aspiration than a problem statement, it usually means the thinking has not happened yet. That is not a criticism. It is just a sign of where the conversation needs to start.

2 - Is there value in this for your organisation or its members?

Not for the individual. Not for the team that is excited about the technology. For your organisation, or the people you serve.

This is where enthusiasm and strategy part company. AI can do impressive things in a demo. But when we dig into what you are hoping to achieve, the answer sometimes comes back to something general: "It would be nice to have." Compare that with a conversation where the answer is "our members are waiting three weeks for a response to a straightforward query, and we are losing renewals because of it." The second conversation has somewhere to go. The first does not, at least not yet.

3 - Do you have priority use cases?

"We’re not 100% sure, but surely everyone would benefit" is a sentence I have heard more than once. It is well-intentioned. It is also a red flag.

If you are ready, you can point to specific areas where you believe AI could add value. Four or five initial use cases, grounded in real operational pain, where the investment of time and resources has a clear return. If those use cases do not exist yet, the work is to identify them before anything else moves forward.

Across our discovery work, the challenges organisations describe are remarkably consistent: manual workarounds that rely on spreadsheets; data spread across multiple platforms with no single view; and reporting that takes days to compile and still does not answer the question leadership is asking. These are real use cases worth solving. They are also notably operational foundations rather than AI projects. Getting them right is where AI readiness begins.

4 - How clean and accessible are your data sources?

"We want to point it at everything in our organisation" is the answer that tells me the thinking has not been done yet.

What we see in practice is straightforward: where systems are fragmented and records are inconsistent, AI amplifies the mess rather than clearing it up. It will work with whatever is given, and if what is given is incomplete or contradictory, the outputs reflect that. The organisations that end up frustrated with AI are rarely frustrated with the technology itself. They are frustrated because the foundations were not there.

The organisations that are ready sound different. They say something like: "Our CRM data has been cleaned, as part of a recent project, and we want to explore how engaged our members are." That is a start for a defined scope, a clean source, and a question worth asking.

5 - Who will run this programme?

"Our IT team have played around with AI, and they are very excited to have a go." That is not readiness. Enthusiasm is valuable. Governance is essential.

What we look for is senior-level buy-in, a clear owner, and a framework for how decisions will be made, how outputs will be validated, and how the programme will be evaluated. Without that structure, what tends to happen is pockets of experimentation that never connect to anything you can learn from. People do interesting work in isolation, and none of it scales.

A note on reporting

There is one area where AI introduces a challenge that most organisations have not yet considered.
For years, the goal has been a single source of truth: one report, one set of numbers, one shared understanding of what is happening. The report might be wrong, but at least everyone is looking at the same version of wrong.

AI changes that. Unless it is structured and instructed very specifically, it may give different answers to the same question depending on who asks and how they phrase it. Greater flexibility, certainly. But also less certainty, and a real risk that different people in the same organisation are making decisions based on different interpretations of the same data.

That is a solvable challenge. But only if you know it exists before you start.

What if you cannot answer these questions yet?

That is not a failure. It is information.

If your organisation is not ready to start an AI programme, the most useful thing a partner can do is tell you that clearly and help you understand where the gaps are.

The organisations that get the most from AI are not the ones that started first. They are the ones who started ready.

Get in touch today to discuss how we can support your next project.

  Find us on LinkedIn and connect!