Ramanan KanakaratnamDelivery recovery. Toronto.

Free five day course · AI programme recovery

Your AI programme has stalled. The model is probably fine.

The programme around it is what needs fixing. Six lenses, one score, and something you can put in front of a leadership team by Friday.

CEO
COO
CTO
Head of delivery
50 to 1,500 people

HOW AN AI PROGRAMME GOES RED

Week 1The pilot demos well. Scale-up is approved.
Week 6Integration work nobody scoped starts eating the schedule.The team knows.
Week 12The data turns out to be two years from ready.Status is still green.
Week 20Production cost lands at five times the pilot.The report finally goes red.

Fourteen weeks between knowing and saying

The risk was visible in week six and written down in week twenty. Nobody lied. Everyone shaded.

The dates were set before you had delivery input.

The status has been managed rather than reported for two quarters.

The people closest to the work have stopped bringing you bad news.

Who it is for

Built for the person accountable for the AI programme

This is for you if

You are a CEO, COO, CTO or head of delivery at a company of 50 to 1,500 people
Your AI programme has stalled, overrun, or gone quiet
You need a defensible basis for changing scope without losing face

This is not for you if

Your AI programme is shipping on schedule
The benefit is already showing up in the accounts
You are looking for a maturity model or a vendor comparison

This is the conversation nobody has in the steering committee.
Programme Director, Financial Services

The week

One lens a day. One action a day.

Day 0

Start here

The six lenses, and the programme where I got this wrong.

Day 1

Green until week twenty

Why the gap between knowing and saying is structural.

Day 2

Value and data

The number nobody can name, and the readiness nobody funded.

Day 3

Acceptance and verification load

What good enough means, and who is checking on hours nobody costed.

Day 4

Economics and adoption

The cost at production volume, and the users quietly avoiding it.

Day 5

Count the reds

How to score all six and take the result into a room.

The framework

Six lenses, one business question each

None of the six is a model question. That is the point.

Value

What number was this supposed to move, and can anyone still name it

Data

Is the data ready for this use, and who is accountable for getting it there

Acceptance

What is good enough, who agreed to it, and how is it tested

Verification load

Who checks the output, how often, and is their time in the budget

Economics

What does this cost at production volume, not pilot volume

Adoption

Are the intended users actually using it, and if not, why not

You can also score it now. The four minute assessment gives you the same six lenses and your red count, before you sign up for anything.

What you leave with

Three things you did not have on Monday

A scored card

Six lenses, scored on artefacts rather than belief, with the rule for reading the result.

A red count

Programmes do not fail on averages. They fail on compounding reds, and you will know yours.

A sentence for the room

The five answers your next steering committee needs, in the form you found them.

Who is writing this

Fifteen years inside programmes that broke, and the ones that recovered

Programme and portfolio leadership across financial services, healthcare technology, transport, retail and education. The method was built inside a regulated environment on purpose, which makes every smaller setting a subtraction rather than a translation.

15years in programme and portfolio leadership

100+programmes delivered

92%on time across a data and analytics portfolio

$1M+annual savings from one platform redesign

PMP · CSM · CSPO · MEng Imperial College London · Based in Toronto

Score your AI programme this week

Day 0 arrives immediately. If the programme has gone quiet, this is the week to find out why.

Finally, someone who understands what it actually costs when the programme goes quiet.
VP of Project Management, Healthcare