AI Experimentation Jam
Turn protected team experimentation into shared, actionable AI use cases
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 7
- Confidence
- 97%
The AI Experimentation Jam creates a bounded period in which employees explore how AI could improve real work. Leaders begin with a problem, pitch, or theme, provide safe tools and learning resources, and divide participants into small teams. Each team experiments with a practical use case over an afternoon or several days, then presents its result and lessons. The event is not limited to developers: a nontechnical company can call it an AI jam session and focus on testing tools, redesigning workflows, or improving customer experiences. Protected time increases participation, demonstrations spread useful ideas, and shared enthusiasm helps make AI feel safe and constructive. Promising prototypes then enter a separate evaluation and implementation process.
Origin
Extracted from Marketing Against the Grain as Rachel Woods recommended hackathons and AI jam sessions as a first step for businesses beginning AI transformation.
Core principles
- 01Dedicated experimentation accelerates practical learning
- 02A clear problem or theme focuses creative exploration
- 03Small cross-functional teams generate varied solutions
- 04Demonstrations turn individual discoveries into shared knowledge
- 05The format works without coding or technical terminology
How to run it
- 1
Choose the focus
Define a business problem, customer group, workflow, or broad experimentation theme. Keep it focused enough to inspire relevant work while leaving room for different approaches.
Pro tip Use a theme such as finding one way AI can improve each team's recurring work.
Watch out A theme that is too broad can produce disconnected tool demonstrations.
- 2
Prepare a safe toolkit
Curate approved AI tools, basic resources, and clear rules about sensitive data. Ensure nontechnical participants can begin without setup barriers.
Pro tip Include examples of simple point solutions as well as general-purpose assistants.
Watch out Do not let participants place confidential information into unapproved services.
- 3
Protect the time
Reserve an afternoon or a multi-day block and treat attendance as real work. Remove normal meeting and delivery expectations during the session where possible.
Pro tip A short, focused event is better than an ambitious event that teams cannot attend.
Watch out Unprotected time will be consumed by ordinary work.
- 4
Form small teams
Combine people with different operational knowledge and perspectives. Ask each team to select a concrete task or experience to improve.
Pro tip Cross-functional groups can reveal handoff problems that individual departments overlook.
Watch out Avoid assigning one technical employee to build everyone else's ideas.
- 5
Experiment rapidly
Have teams test tools, prompts, workflows, or prototypes and compare them with the current way of working. Capture both successful and failed attempts.
Pro tip Focus on learning and evidence rather than polished presentation quality.
Watch out Do not expose prototypes to customers during the event.
- 6
Present and share
Bring everyone together to demonstrate findings, explain value, and name unresolved risks. Record reusable resources and insights in a shared space.
Pro tip Use a standard presentation structure: problem, experiment, result, risk, next step.
Watch out A demo without documentation will be forgotten quickly.
- 7
Select follow-up experiments
Identify ideas with meaningful value, feasible controls, and measurable outputs. Assign owners to test them more rigorously after the event.
Pro tip Advance only a small number of high-signal ideas.
Watch out The jam session discovers opportunities; it does not certify production readiness.
In the wild
A manufacturing company reserves Friday afternoon, provides links to approved AI tools, and asks mixed teams to explore one improvement to quoting, maintenance documentation, scheduling, or customer communication. Each team shares what it tried and what evidence would be needed next.
→ Employees discover relevant applications without needing developers or adopting a technology-heavy hackathon format.
Teams receive the prompt to find one way AI could deliver a better experience for a defined customer group. They build lightweight demonstrations over two days and present them to colleagues and leaders.
→ The event produces concrete use cases while increasing AI literacy and cultural momentum.
Common mistakes
Making the event developer-only
Useful AI experimentation can involve testing, workflow design, research, and operational judgment rather than conventional software development.
Skipping the final sharing session
Without demonstrations, individual learning does not become a reusable organizational asset.
Shipping prototypes immediately
Rapid experiments still require data review, evaluation, risk management, and accountable ownership before production use.
Is it for you?
Best for
It is best for teams that understand AI's importance but have not yet translated it into practical workplace use cases.
Not ideal for
It is not ideal for deploying sensitive or customer-facing systems without subsequent governance and testing.
From the transcript
“I mean I really love encouraging businesses to throw like a hackathon type day.”
“basically what you do is you, you know, establish a problem,”
“You know, it's a Friday afternoon AI jam session”
From the episode
AI Expert On How To Use Ai To Save Time & Grow Your Business (#149)