What DSA does and when it’s the right fit

‍

Most AI work doesn't begin with a clean technical specification. It starts with a request: Can we use AI here? Can we automate this? Should we build an agent for that?

Very quickly, the question stops being purely technical. Someone needs to understand the problem, look at how the work happens today, bring business and technical context together, and decide what AI should do, what people should continue to do, and whether anything needs to be built at all.

No single person usually holds enough context to make those calls alone.

That is the space we work in.

What Design Sprint Academy actually does

AI can do the work. Judgement is still human.

The more AI can do, the more important some human decisions become: which problem is worth solving, where AI can create meaningful value, how a workflow should change, and what is worth building.

These are usually cross-functional decisions. The people who understand the business problem, the workflow, the users, the technology and the organisational constraints are rarely the same people.

Someone has to bring that context together and help the group make a good call.

We call that capability AI Facilitation.

It combines structured problem discovery, critical thinking, cross-functional collaboration and decision-making around AI. The goal isn't simply a productive workshop. It is a decision the group understands, can defend and can act on.

AI Facilitation is a capability, not necessarily a job title. It can sit with an AI Champion, an AI Enablement lead, a Forward Deployed Engineer, a solution architect, an IT business partner or someone else whose job increasingly requires them to work across business and technology.

How organisations work with us

There are three main ways.

We facilitate the work with you. We use methods such as AI Problem Framing and the AI Workflow Sprint around a real challenge, bringing the relevant people together and guiding the decisions that need to be made.

We build the capability inside your organisation. If you already have people who should lead this work, we train them to use the methods themselves, in the context of the situations they actually face.

We help you make the capability repeatable. The AI Lab is an operating model that can combine facilitated work, training, coaching and ongoing support when you need a system rather than a one-off intervention.

These are different ways of applying the same underlying capability.

Where this capability becomes useful

The context changes, but the core problem is similar: important AI decisions depend on knowledge spread across several people.

Moving from AI adoption to workflow-level impact

Many organisations have already invested in AI literacy, tools and experimentation. Eventually the question changes from Are people using AI? to What is actually changing because of it?

A prompt course can help someone perform an existing task faster. It doesn't tell a team whether the workflow itself should change or which opportunities matter enough to pursue.

That is where AI Enablement and transformation teams need people who can identify meaningful opportunities, bring the relevant domain and technical perspectives together, and help teams redesign work rather than simply add AI to one step.

Developing the consulting side of Forward Deployed Engineers

Forward Deployed Engineers bring deep technical capability, but the work before and around the build also requires them to explore ambiguous business problems, work with domain experts and help decide what should be built.

Structured facilitation gives FDEs a repeatable way to run that discovery and alignment work. It is not generic soft-skills training. It is methodology for the non-technical work that makes the technical work possible.

Helping AI vendors and IT teams work further upstream

AI requests rarely arrive build-ready. A team asks for an agent or an automation, but the problem, workflow, constraints and decision points may still be unclear.

AI Facilitation gives internal IT teams, solution architects, implementation partners and professional-services teams a structured way to work through that ambiguity with the business before engineering starts.

The aim is simple: turn a vague request into enough shared understanding to make a good decision about what happens next.

The methods give the group a way through the ambiguity

Our roots are in Design Sprints and Problem Framing: structured methods for bringing different perspectives together, making assumptions visible and moving toward a concrete decision.

AI changed the nature of some of those decisions. The question is no longer always about a new product or feature. Often it is about how existing work should change once AI can perform parts of it.

That led us to develop AI Problem Framing and the AI Workflow Sprint.

AI Problem Framing helps a group understand the problem and identify where AI could create meaningful value before jumping to solutions.

The AI Workflow Sprint takes the workflow itself as the unit of change. The team examines how the work happens, where judgement sits, what AI could take on and how the workflow should operate differently.

The methods are important because they make difficult conversations easier to navigate. But the method is not the end product. The decision is.

The workshop is only the visible part

A good session depends heavily on what happens before people enter the room.

We need to understand what has already happened, what evidence exists, who holds relevant context, where the disagreements are likely to sit and who can make the final call. If research or existing artifacts already exist, we use them. If the group needs something specific to understand the problem, we prepare it.

This is also why we don't automatically recommend a workshop.

If the relevant information is already known, the trade-offs are manageable and the right people can make the decision in a normal meeting, do that.

Structured facilitation earns its keep when the problem is ambiguous, context is distributed across people, important assumptions are hidden or the group keeps reopening the same decision.

In those situations, decision quality matters more than the illusion of certainty.

You may not know whether the decision will turn out to be right. But you can ask whether the right people contributed, whether the relevant evidence was considered, whether assumptions and trade-offs were surfaced, and whether everyone understands what was decided and why.

That standard matters even more when AI makes building faster. Easier execution is useful, but it also makes it easier to build the wrong thing quickly.

We complement the teams already doing the work

Most organisations we work with already have capable people: AI teams, internal facilitators, transformation teams, consultants, implementation partners or product and technology teams.

We don't need to replace them.

Often our role is to help at the points where those groups need to make a decision together, or to give them a method they can use without us next time.

There are two reasons an external facilitator can still be useful.

The first is neutrality. Someone inside the organisation may have the facilitation skill but also report to a person in the room, own part of the initiative or have a preferred outcome. An external facilitator has less history and less attachment to the answer.

The second is specialisation. Your internal team has another job to do. Ours is designing and facilitating these decision processes, refining the methods and teaching others to use them.

For an everyday decision, that difference may not matter. For a difficult cross-functional decision with real consequences, it can.

The longer-term goal is still internal capability. If this kind of work matters repeatedly, your people should eventually be able to lead it themselves.

When we are probably the wrong fit

You don't need us for every AI conversation.

If you need a team-bonding day, there are facilitators who specialise in that. If the decision has already been made and the workshop exists mainly to communicate it, you don't need a decision method. If you need a partner to build and operate a large AI platform, you need technical delivery capability. If you need someone to run your entire transformation program for several years, that is a different kind of engagement.

And if your own team already has the capability, neutrality and capacity to do this work well, use them.

We are most useful when there is a meaningful AI question in front of you, the answer is not obvious, and several people need to combine what they know before the organisation can move.

That is the work AI Facilitation is designed for.