AI and the future of work

Jobs don't vanish whole.
Tasks move.

When first drafts, research and analysis get cheap, work doesn't disappear. It gets redistributed. Amit helps leaders work out which tasks move to AI, which stay human, and how teams and managers need to change around both.

25+Hours saved per week at Coinbase through agentic AI automation

What moves, and what stays human.

Moves to AI

The first pass.

  • Drafting and summarising
  • Research and sorting
  • First-pass analysis

Stays human

The call.

  • Deciding what matters
  • Weighing trade-offs
  • Owning the outcome

The useful unit isn't the role. It's the dozens of tasks inside it.

Bolting AI onto an old process doesn't transform it. It just makes the old process faster.

SR-001 calls this automation debt: quicker steps inside a workflow nobody redesigned.

Skills that grow in value.

  • Judgment

    Deciding what's worth doing, and what a good answer looks like before asking for one.

  • Verification

    Checking AI output against sources, context and common sense before it goes anywhere.

  • Context

    Giving AI the right background, constraints and examples so the output fits the real situation.

  • Better questions

    A leader's job shifts from having every answer to framing the questions a team and its tools work on.

How a team's work gets redesigned.

  1. Map tasks, not jobsList the work inside each role
  2. ReallocateDecide what AI drafts and what people decide
  3. Rebuild the manager's roleFrom answers to questions and review
  4. Govern the new workOversight in proportion to what AI does

Questions people ask.

Will AI replace jobs?

It replaces tasks more often than whole jobs. Most roles will keep their purpose but change their mix of work, with routine drafting and analysis moving to AI and judgment, relationships and accountability staying human.

What skills matter most as AI changes work?

Judgment, verification, giving AI good context, and asking better questions. These are the skills that decide whether AI output is useful or merely fluent.

How should managers change in an AI-augmented team?

Managers need to review AI-assisted work well, redesign workflows rather than add tools on top of them, and help people through the uncertainty the change creates. A team rarely adopts faster than its manager does.

Where this connects.

Redesigning work around AI?

Advisory, workshops and leadership sessions on how roles and teams change.