Now everyone can build everything
AI:CLEAR is a Neuroblitz programme on leading with AI. Departments write their own automations, teams train their own assistants, and some people already work with agents. What remains open is whether all of it still fits together. AI:CLEAR brings leaders and teams together to settle that question.
In our booklet, this programme is called KI:KLAR.
Signs that alignment is missing
Every team has its own solution
Nobody has an overview of what has been built, what exists twice and where data ends up.
Some are far ahead, others keep their distance
Between enthusiasm, scepticism and quiet worry, there is no shared language.
Results come quickly, responsibility gets blurred
Who stands behind a text, an analysis or a process that an AI helped to produce?
The strategy does not reach everyday work
There are papers and policies, and teams still don't know what these mean for their own work.
What matters more when building gets easy
Technical hurdles are falling away. The work moves to questions that no tool will answer for you.
Metacognition
How do we think when AI thinks along?
People who notice when their own judgement holds and when it is merely going along use AI differently. This can be practised.
AI architecture
What is connected to what?
Tools, agents, data and processes need a shared picture that people without an IT background can read.
Governance
Who decides what?
Rules work when the people affected have helped to write them. So we develop them while working on specific tasks.
From trying things out to shared practice
Participants experience first and talk about it afterwards. Here are four formats from our workshops.
Train of Thoughts
Human intelligence competes live against artificial intelligence, and you can feel how differently each arrives at its answers.
Human, AI or both
Three groups solve the same task under different rules and compare the results.
Centaurs and cyborgs
Teams sort tasks from their own areas and decide where they want to keep human and AI work apart and where they want to merge it.
Hackathon
Mixed groups work with AI on a question from the organisation and decide who takes on which role in the team.
Centaur or cyborg?
Ethan Mollick describes two ways of working with AI. Centaurs keep a clear division of who does what. Cyborgs work so closely with it that the line blurs. Where do you stand?
A prompt for reflection, not a validated test. Your answers stay in your browser and are not stored.
How the programme works
Understand
Conversations with sponsors, IT and business units. What already exists, which rules apply, where are things stuck?
Focus
Use cases, resistance and tensions are distilled into clear fields of action.
Try out
In workshops and hackathons, participants work on their own tasks with their own tools.
Anchor
Planned experiments and shared rules carry the topic into everyday work, with our support.
Building blocks
The mix depends on your starting point.
What you have afterwards
A map of your use cases
Tasks from your areas, ordered by their importance to the organisation.
A shared picture of your AI landscape
What is already running, what belongs together and who is responsible for what.
A Leadership Code
Five principles for working with AI, written by your own leaders.
Planned experiments
Concrete next steps for everyday work, each with human sign-off.
How one workshop became a partnership
For the leaders of an energy company, we developed a two-day workshop on psychology and leadership in the age of AI. They used the company's internal AI tool on their own tasks and collected the places where AI could make the most difference.
Since then, we have supported the company with webinars, hackathons, workshops and talks.
What the research says
In a study with BCG consultants, some kept their work clearly separate from the AI, while others wove the two together. This is where the terms centaur and cyborg come from.
A meta-analysis of 106 experiments found that human-AI combinations on average did worse than the stronger of the two alone. Creative tasks showed gains.
Dell'Acqua et al. (2023), Harvard Business School Working Paper. Vaccaro, Almaatouq & Malone (2024), Nature Human Behaviour.
Frequently asked questions
Will participants also learn how to use the tools?
Yes. They work on their own use cases and get to know different AI tools. We always link this to the question of how the work fits together in the team.
Which tools do we work with?
Mainly those available to you. Where useful, we show what other tools can already do.
Who is the programme for?
Leaders and teams who already use AI or are introducing it. No prior knowledge is needed, and mixed groups are welcome.
How do you handle confidential data?
We agree together in advance which data and tools may be used in the programme.
Who is behind it
Dr. Ben Hartwig founded Neuroblitz. At the Max Planck Institute, Ben researched molecular genetics and epigenetics. Add to that over 20 years of improv theatre in five groups, six of them with Springmaus, with more than 1,000 performances, festivals and tours around the world.
Dr. Ben Hartwig has worked with language models since 2019. In the theatre project Improbotics, Ben performed on stage with an early model: some performers spoke for themselves, others repeated what the AI fed them through an earpiece. The audience had to work out who was who, and sometimes could no longer tell.
Contact us
info@neuroblitz.de