AI Discovery Sprint

AI Discovery Sprint

Identify and prioritise applications of AI in your business

An AI Discovery Sprint is a (typically) 2-day workshop with the aim to identify opportunities for applying AI to improve your product, company or department.

With AI becoming increasingly available to companies of all sizes and impacting all aspects of business, from simple cost reductions through automation to powering completely new business models, it is essential to understand how it can be applied in your context.

In the AI Discovery Sprint we merge problem mapping tools from Design Thinking with inspiration checklists for AI products to generate concepts in a fast and structured way.

We are now witnessing a new shift in computing: The move from a mobile-first to an AI-first world

- Sundar Pichai, CEO, Google

Deliverables & impact

  • A structured and prioritised list of potential use cases for applying AI / Machine learning within your company
  • An initial assessment on feasibility and value
  • A clear vision on how AI will impact your company or department in the short to medium term

When to run an AI Discovery Sprint?

Simply put: now.

Unless you already have a clear vision of how to apply AI and Machine Learning to your product, department or company then you are likely to be at best missing opportunities and at worst risking being left behind.

An AI Discovery Sprint is the most effective process we have found to gain that vision.

Amazon’s six-page memos are famous. Executives must write one every year, laying out their business plan. Less well known is that these missives must always answer one question in particular: how are you planning to use machine learning? Responses like “not much” are, according to Amazon managers, discouraged

- - The Economist

Who should take part?

Around 5 participants plus a coach, in particular:

  • Business and domain experts
  • Product owners
  • AI Engineers
  • An AI Sprint Coach to run the training and guide the team.

The participants need to have a good understanding of the capabilities and limitations of AI.. If need be, they can be trained ahead of the workshop.


2 days is typical for this exercise, but this can vary a great deal depending on the amount of processes you want to cover and how complex they are. If you are unsure, we recommend starting by narrowing the scope down to a process that fits within those two days.

Day #1

  • Step 1: Define focus and map
    E.g. automate business processes, optimize existing products or invent a new product or service

Day #2

  • Step 2: Ideate
    Identify “AI-Friendly” problems and ideate solutions using checklists and inspiration triggers
  • Step 3: Evaluate
    Sanity check, initial feasibility evaluation and impact assessment
  • Step 4: Decide and prioritise
    Prioritizing next steps: Buy, build, monitor or park

And then ...

Projects you decide to implement yourself need to be analysed more precisely for business / user value and technical feasibility, either by working on a technical proof of concept or by running an AI Design Sprint.

Products you decide to buy in need assessing, though in that case it doesn’t differ from other IT procurement processes.

Want to find out more?

Read our blog post on how to run an AI Discovery Sprint.

Ready to run an AI Discovery Sprint?

Get in touch

We can run your AI Discovery Sprint in English, German or French