Randy Jones

Experimentation and adoption

When AI pilots stall, the reason is usually organizational rather than technical. People can’t get access to tools, trying something new takes months of approvals, and when a team does find something useful, nobody else hears about it.

These posts describe the system I’d put in place instead. Give everyone a small AI budget and a sandbox where trying new tools is safe and fast. Treat unapproved AI use as a sign of demand rather than something to stamp out. Keep a shared record so good solutions get reused instead of rebuilt. The post on why pilots stall is the short version of the whole argument.

5 posts, in suggested reading order

  1. Why AI pilots stall, and what to build instead

    Most AI pilots stall for organizational reasons. A budget, a sandbox and a shared record of what works do more than another round of pilots.

    4 min read · Updated September 2026

  2. An AI budget for every employee

    Give people a small monthly AI budget, no approval per experiment, inside a sandbox. The people closest to the work will find the use cases.

    4 min read · Updated September 2026

  3. Safe early access to new AI tools

    The bigger risk is often the year spent deciding whether a tool is safe. A sandbox lets people try it in weeks, with the risk contained.

    4 min read · Updated September 2026

  4. Shadow AI and the approved tools nobody uses

    Banning unapproved AI pushes it out of sight. Read it as a map of unmet needs, give people choice and a safe place to practise, and measure depth of use.

    6 min read · Updated September 2026

  5. Why teams rebuild what already exists

    Organizations pay for the same solution again and again because nobody can find the first one. AI agents will do it faster unless reuse gets easier.

    5 min read · Updated September 2026