Why does SAP need an AI community?
SAP is one of the largest software companies in the world: 120,000 employees in 130 countries, with a development center of more than 1,000 employees in Israel. As a company leading Business AI, it deals with a huge flood of AI information, tools, and innovation.
New models come out every day, requiring them to adapt their products, workflows, and development processes. The community was set up to make sure things don't get lost along the way, and that there's a clear place to learn and work together.
The main challenges that led to founding the guild:
- An unfiltered flood of information: new models and technology changes every week
- Different professional needs: every role needs a different approach to AI (developers, product managers, UX, managers)
- Regulatory challenges: the need to meet compliance requirements in a global company
- Preventing Shadow IT: employees using external AI tools without oversight
How did the community get started?
The guild started as a bottom-up initiative from the field. 3 people founded it: Daniel Nemet and Dina Winter (2 senior developers) and Asaf Bruner (the guest in this episode).
The founding process:
- Statement of intent: defining goals and areas of responsibility
- Internal call for participants: choosing 20 people out of a large pool of candidates ("cherry picking")
- Defining the structure: deciding on meeting frequency, content, and ways of working
- Getting buy-in: approval from leadership and from direct managers
We had so many questions about how to build this community. Honestly, something like 200 decisions we had to make along the way.
How is the community's activity structured?
The guild grew from 20 to 50 members and runs on a structure based on consensus and active involvement.
Management structure: there's no single leader. Decisions are made by consensus, and the guild leads come from several fields and different parts of the organization.
In-person meetings take place in Raanana and Tel Aviv every 3 weeks and last about 1.5 to 2 hours.
Examples of content:
- 6 topics of 15 minutes each: a move from long lectures to a fast format
- AI news updates: what happened, what's relevant
- Code examples: sharing best practices
- Tool showcases: internal and external
- Technical discussions: prompt engineering, context engineering
- Hands-on workshops for installing new tools
- Collaborations with other AI guilds across SAP worldwide
Practical tips for running it:
- Short, varied sessions (15 minutes per topic)
- In-person meetings for real connections
- Content that comes from the members themselves
- Frequent feedback, and changes as needed
What's the win-win of the community?
For the organization: faster AI adoption, preventing Shadow IT, more efficient employees, internal tool development, and integration into SAP's products.
For employees: professional development, a wide professional network, a chance to lead and present, and becoming the team's focal point for AI.
What challenges did you face?
Technological challenges:
- Staying up to date: "One of the biggest challenges in the AI world is keeping your finger on the pulse."
- Filtering out the noise: telling what's relevant from what isn't
- Technology clashes: different teams using different tools
Community management challenges:
- Member engagement: dealing with inactive members in an exclusive community
- Balancing content: across the different needs of developers, product managers, UX, and managers
- Deciding on size: not growing beyond 50 people, to stay effective
Tips for making it work
- Start with the right people: "Take people who have a passion for this field, who can lead and execute."
- Don't do it alone: "Without the other guild leads it wouldn't have been possible... this is something that takes more than one person."
- Ask for support from the top: make sure you have buy-in from leadership.