Choose and prepare the work
Confirm two recurring tasks, permitted sources and tools, quality criteria, and staff time.
DBX / service offerAI Communities of Practice
For public agencies and mission-driven organizations
Turn AI learning into better everyday work.
For teams that have introduced AI tools or training and want to make them useful in daily work.
DBX helps your team put approved AI tools to work on real tasks, learn together, and build methods colleagues can reuse.
02 - The ideaSmall group / related work / reusable methods
Practice is the bridge between training and work.
A community of practice is a small staff group that practices on related work, reviews results, and shares useful methods.
The group works on agreed tasks in approved tools, reviews real outputs, and shares methods colleagues can use again.
Progress means staff can apply the method to a real task and repeat it with less support.
03 - Briefing processApproved source material
Illustrative example
A briefing from approved source material.
A staff member prepares a briefing for a colleague or decision-maker. The group tests the approach, reviews the output, and saves a method others can reuse.
- 01
Choose
Choose permitted sources and the purpose of the briefing.
- 02
Draft
Draft in an approved AI tool using the agreed source material.
- 03
Check
Check facts against sources, omissions, tone, and suitability.
- 04
Improve
Improve with peer feedback and make the reasoning visible.
- 05
Review
A designated human reviews the briefing before use.
The outputShort briefing with source links
Editable guideso another colleague can repeat the method
Review checklistfor facts, omissions, tone, and suitability
Source permission does not guarantee accuracy. Designated people remain responsible for checking and approving work before use.
04 - The engagementSet up / facilitate / document / hand over
A focused launch, with ownership visible from the start.
DBX helps your staff configure an agreed practice space and templates in approved existing tools where practical. We run practice together, document two task guides and worked examples, review applied work, and coach two internal hosts to take over the routine.
Practice, review, improve
Apply the method to real work, use peer feedback, and get help during scheduled sessions.
Show repeat use
Inspect applied work and document what staff can repeat with less support.
Prepare to continue
Two internal hosts rehearse welcoming members, guiding discussion and keeping reviewed guidance findable, with DBX support before taking over.
Managers agree staff participation time.
Internal hosts organize sessions, welcome colleagues, and keep guidance current.
Workflow owners retain approval responsibility for work used by the organization.
What your team keeps
Two task guides, worked examples, a shared practice space, evidence of applying the method, two internal hosts, and a continuation plan.
Team-led operation is the design goal, with an internal owner and agreed staff time. Optional ongoing DBX facilitation and support are quoted separately; technical operations remain distinct.
05 - ProgressAgree what useful practice looks like
Look for practice that holds up outside the session.
Progress criteria are agreed before the work begins:
- Agree quality criteria upfront.
- Inspect work samples.
- Check repeat use with less support.
- Compare task time or review effort only when comparable.
Attendance shows participation; work samples and repeat use show applied practice.
At close
You receive a clear account of what staff demonstrated, what still needs support, and which tasks are worth developing further.
06 - Who leads itPractical experience, clearly described
Led by Denis Bundi, with the DBX team
Learning design meets implementation reality.
Denis connects learning design, software development, and hands-on AI implementation. His personal experience informs how DBX frames practice, review, handover, and responsible use.
Explore our learning systems expertise07 - Buyer questionsShort, practical answers
Before you decide.
Use the fit conversation to test the work, people, and tools - not to commit to a platform.
What work is suitable?
Recurring work with a clear owner, a reviewable output, and enough repetition for a small staff group to learn together.
Do we need a new platform?
Agreed existing tools may be used where practical. A dedicated platform is chosen only if it is useful and justified by the scoped work.
How much staff time is needed?
Managers and DBX agree participation time and cadence upfront, after understanding the task and constraints.
What should not go into the tools?
Do not use sensitive or unapproved content in AI tools. The organization’s permissions, policy, and workflow owners govern what is permitted.
How does the fee work?
The engagement is fixed-fee once scoped. Platform, integration, and ongoing costs are made clear before commitment. There is no public price.
08 - Start with one taskRequest a fit conversation
Bring one recurring task.
Start with the work you want people to do better.
A 30-minute fit conversation explores your work, people, and approved tools before DBX recommends a focused launch and scoped proposal.
Denis Bundi, Founder & Principal
denis@dbxai.com