Independent proof of what your AI and technology spend is delivering

You need a board that makes decisions, and you keep being asked from the top which of your AI spend is worth it.

This bundle is both kits.

• Decision-First Programme Board Pack. A 12-slide board deck that opens with the decisions, a worked example, a controls workbook (tolerances, RAID, decisions, benefits) and board terms of reference.
• AI Portfolio Keep / Fix / Stop Kit. A register that gives every AI line a Keep, Fix or Stop call weighted by evidence, a deep-dive workbook for the close calls, and a board paper with a filled-in example.

Ten files, including two guides, with worked examples throughout. £99, against £108 bought separately.

If your board pack opens with twenty slides of status and RAG ratings, the decision you need tends to come at the end, when the chair is already watching the clock.

The boards I've seen work all started with the decision. I built this pack that way.

• A 12-slide board deck in PowerPoint. The first slide after the cover holds the decisions needed today: what you are asking for, your recommendation, the cost, and what happens if the board does not decide. Every slide has speaker notes saying what goes on it.
• The same deck filled in for a fictional finance systems programme, so you can see the length and tone a board actually reads.
• A Programme Controls Workbook in Excel: tolerances with automatic RAG, a scored RAID log, a decision log that flags overdue decisions, a benefits tracker that keeps forecast and booked apart, and a dashboard that writes the headline for your pack.
• Programme Board Terms of Reference in Word, in UK programme language: SRO, Senior User, Senior Supplier, tolerances, exceptions, quorum.
• A four-page guide to running it in about an hour a week.

It is five files that work together, built around one worked example.

They came in as licences and pilots at different times, and each sponsor will tell you theirs is working.

Your board is starting to ask which ones are worth the money. In my experience most organisations can't answer that yet. MIT's NANDA research in 2025 found that 95% of the organisations it looked at had no measurable return from their generative AI pilots six months on. It was a small study, mostly of larger firms, and the authors say so.

I don't think a cleverer ROI calculator is the answer. What works is the same few tests applied to every line, and someone willing to write down Keep, Fix or Stop.

• AI Portfolio Register (Excel). Every AI tool and initiative on one sheet. It weights each claimed benefit by the evidence behind it, from opinion (0%) to money in a budget (100%), checks adoption, ownership, kill criteria and data clearance, and gives every line a Keep, Fix or Stop call with a plain reason. A Board Summary tab writes the sentences for your paper.
• AI Initiative Evidence Workbook (Excel). For the one line that needs a closer look: full cost to Year 3 including internal time, benefits as claimed and as weighted side by side, payback, a downside case, and a ten-question gate score with hard stops.
• Board paper template (Word) that puts the decision first, and the same paper filled in from the worked example, so you can see the length a board will read.
• A six-page guide to running the review in two weeks.

Everything ships with a worked example: a fictional 600-person business with eight AI lines.

Work out what is really wrong with a troubled programme and put a clear Keep, Reset or Stop recommendation to your sponsor.

Everything ships with a worked example: a fictional housing system replacement, followed from health check to sponsor paper, so you can see how the finished files read before you put your own programme in.