How to brainstorm in the age of AI?

Do you remember the days when conducting a brainstorm to collect great ideas was a considerable effort? Gather a core group of people in a room, discuss and debate the problem for a few hours, and stay there until everyone agrees on the outcome.

Advances in AI have put paid to that – we now have access to tons of ideas. The problem is, we have far too many.

Today, we are under pressure to think faster, deliver quicker and add ever more value. It makes sense that people are turning to AI to reshape how this is done. Where we once came together collectively to solve a challenge, now one person can do much of the legwork on their own and in half the time.

One person, hundreds of ideas – whatever could go wrong?

From shared thinking to solo thinking

On the surface this new approach looks like progress. However, this subtle shift, away from collective wisdom to solo exploration, changes more than just speed. It also changes the ownership of ideas from the group to the individual.

Here’s an example. We recently developed a new product and needed to define the value proposition that would shape how it was communicated to clients. Instead of spending a few hours together debating ideas and refining language, we asked just one team member to create a first draft with the help of AI. The plan was to circulate it and gather feedback from the wider group afterwards.

The approach started well, a great first draft was developed and feedback was shared. However, what we had not bargained on was the challenge of alignment.

The person who developed the draft and spent hours thinking deeply about the problem was understandably quite attached to their ideas. But no one else had been taken through the thought process and been able to challenge and debate along the way. As a result, no one fully understood how the final proposition had been reached or felt true ownership of it.

What followed was weeks of emails, each person suggesting changes or offering new ideas. What should have been a quicker process actually took longer.

Brainstorming is never just about ideas

What we learnt from this process is that the value of brainstorming is never just about the ideas that get produced at the end. The value is in the process itself.

The act of debate and discussion means that people feel they have some agency in the room and their voices and opinions are heard. After a brainstorm everyone can tell you what came out of it and how they got there.

With siloed thinking, where one person works with AI alone, only one person gets that benefit. This is a problem if several people are then expected to explain and stand behind the output.

Designing a new way of brainstorming

So, what is the answer? The answer is not to go back to our old methods, but to move forward with a new way of working designed for the world of hybrid thinking that we now live in. One that uses AI as part of the team rather than in place of it.

In practice this could look like the following:

1. Develop a bespoke AI Agent

Built once and shared with the team to ensure it helps rather than hinders:

• Fully briefed on the problem/ challenge
• Trained in any technical knowledge needed
• Provided with background documents and context
• Given examples of what good looks like from similar challenges

2. Run a pre-task sprint

Each person explores the challenge individually using the shared AI agent, with clear objectives:

• To help clarify the brief and core business question
• To conduct desk research to deepen understanding
• To help people arrive with some early ideas to get the creative juices flowing
• The idea is NOT to come with a solution or polished thinking at this stage

3. Hold a hybrid brainstorm

Involving the core group and bespoke AI Agent, in a room together:

• A regular human brainstorm takes place
• Whenever thinking gets stuck or ideas run dry, the group turns to their bespoke AI Agent to help unlock new angles
• A series of human / AI sprints occur until the group feels 90% there
• Everything is captured on an electronic whiteboard or audio recorded

4. Complete the last mile

This is when one person takes the near-final idea and creates the polished version:

• The brainstorm outputs are added to the AI Agent to inform it of the latest thinking
• One person runs a focused session to refine language and resolve final details
• The final version is circulated for review

This may seem like a longer process than simply asking one person to draft something with AI. However, we have learned that humans and AI together are more powerful than each on their own.

The results from a process such as this will benefit from:

• The collective wisdom and creativity of the group
• The processing and writing strengths of AI
• A shared understanding of how the solution was reached – making it more meaningful and memorable to all involved
• A clear beginning and end, reducing endless feedback loops

Hybrid thinking is designed, not accidental

One of the biggest lessons from this shift is that hybrid thinking does not just happen on its own. Simply adding AI into the mix does not automatically create better collaboration. If anything, it can pull thinking apart and create misalignment if not used intentionally.

Before your next workshop, sprint or co-creation session, it is worth asking a simple question: are we using AI to think instead of each other, or to think better together? The first may feel efficient in the short term but the second is what gives ideas the momentum to truly move forward within an organisation.