01 / The work

I build the system that doesn't exist yet.

For the bet everyone's excited about and nobody owns.

Bring me the gap ↗ Take the AI scorecard → 10 questions · free · no email

From the field.

>80%monthly active AI usage across 1,500 people ↗ 4agents in daily use ↗

Across classrooms & boardrooms

  • BITS Pilani
  • INSEAD
  • KPMG
  • EY
  • Tata
  • Novo Nordisk
  • Clinton Health Access Initiative
  • The Whole Truth
  • upGrad
  • Plum
  • LevelUp Labs
  • Viacom

One recurring gap · AI adoption

You bought the AI.
What got in the way?

10 questions about how your team actually works. You get a likely adoption pattern, two areas to investigate, and a next step you can use.

Take the AI scorecard → 10 questions · Free · No email required Already building? Find your first agent ↗
ADOPTION SCORECARDEXAMPLE RESULT

Your answers might suggest

Island of
excellence.

A few people get enormous value.
The practice isn't spreading.

ONE MOVE THIS WEEK

Give one champion protected time to help three colleagues complete a real task.

Illustrative example. Your result depends on your answers.

02 / How bets stall

  1. Mandate stays fuzzy.

    Everyone agrees it matters. Nobody agrees what it is.

    Shape the bet

    Market maps, customer truth, opportunity sizing, and a clear answer to where to play and how to win.

  2. Thing it needs doesn't exist.

    No function, no system, no route to market...

    Build the function

    First operating model, intelligence engine, GTM system, or strategy and analytics capability, from scratch.

  3. Works until the champion leaves.

    The pilot glows, then someone changes roles and off it goes.

    Make it hold

    Pilots, governance, measurement, and ownership until it runs without me.

03 / Selected work

Dispatches from the field.

See all projects →

04 / Detours

Bets I got wrong.

(You learn more from the detours.)

  1. What I believed
    A good dashboard shows everything.
    What actually happened
    Too many metrics. Adoption only went up once I cut it back to the essentials.
    What I do differently now
    I start from the decision an executive needs to make, and work backwards to the few numbers that answer it.
  2. What I believed
    Get the technology right on a “simple” AI tool, and people will use it.
    What actually happened
    I underestimated how much change management it needed.
    What I do differently now
    I plan adoption from day one: start with power users, appoint champions, scale once it’s proven.

The next interesting thing

Got a bet with no owner?

Bring the half-formed version. You'll leave the call with a clearer next step, whether or not we work together.

I take on 3 new bets a quarter. 1 open for Q4 2026.

Bring me the gap ↗

Prefer email? workwithakshatkharbanda@gmail.com

You scrolled all the way down. That's either diligence or procrastination. Either way, I respect it.