Giving staff an AI tool without making room to learn it leaves experimentation competing with the day job. A Barracuda product chief says his company’s hackathon worked better when normal work was set aside and the tools, access and guardrails were ready before the clock started.
Watch Desk analysis
What happened
In a Forbes Technology Council article, Neal Bradbury describes lessons from a two-day AI hackathon at Barracuda. He says teams were given time away from sprint work, with API access, integrations, licences and model capacity prepared in advance. Participants worked in sandbox environments using synthetic rather than customer data, with least-privilege access and human approval for AI-generated changes.
Bradbury says teams tackled business-relevant problems, including speeding up software delivery and automating familiar processes. Projects were scored partly on measured impact, and he describes examples including cutting one workflow from three weeks to under 24 hours. Those are company-reported results from his account, not a general guarantee that hackathons produce the same gains.
Why it matters
The practical lesson is that “try AI” is not much of a plan if employees must squeeze it between meetings and deadlines, then spend their scarce time asking for access. Protected time and a working, secure setup make experimentation possible; choosing real business problems and measuring before-and-after results make it easier to tell useful work from a polished demo.
There is also a way to stop the learning evaporating when the event ends: Bradbury recommends collecting reusable tools and techniques in a shared playbook and giving promising projects a route forward. Otherwise, the best output may be a very impressive demo and a familiar backlog.
Our read
This is a practical argument, not proof that every organisation needs a hackathon. The more transferable advice is to remove the predictable friction: give people time, prepare access, set security boundaries, and reward evidence of impact rather than stage presence. If a company wants staff to adopt AI, it should make learning part of the work, not a hobby performed after it.
What to watch
- Whether teams carry useful agents, integrations and prompt patterns into a shared playbook.
- Whether claimed time savings are measured against a clear baseline.
- Whether promising experiments get a path into ordinary work, with appropriate review.
Discussion spark: Should employers protect dedicated time for AI experimentation, or should new tools prove their value within existing workloads first?
Sources and evidence
- Want Employees To Adopt AI? Give Them Time To Learn – Forbes (4 November 2026, 14:15 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.