Use case

Project approval decisions

A project goes to a board as a business case, and the board's job is to find the part that does not hold. BrainPower AI builds the case so that the weak parts are visible before the meeting: the cost as a range rather than a figure, the slippage priced rather than hoped away, the option of deferring costed properly, and every assumption recorded with its source.
Run the live demo

Why this decision is hard

Because the case is written by the people who want the project. That is not dishonesty, it is gravity: the capital figure that gets typed is the one that makes the case work, and the slippage that has happened on every comparable build is left out because including it is an argument the sponsor would rather not have in writing.

Because deferring is treated as the safe option and never priced. Deferring costs years of the benefit and usually costs more capital when the money is finally spent. Until that is on the page, do it later is a way of avoiding the decision rather than making it.

And because the board asks what if. If the answer to what if the build runs six months late requires a week of rework, the board gets the version that does not answer it.

What goes in

  • The capital cost, as a range that includes the number the engineers actually said.
  • What the project saves or earns a year once it runs.
  • Months to build and commission.
  • Slippage on top of that, as a range, because this is the figure history is loudest about.
  • The deferral option, with what waiting does to the cost and to the remaining period.
  • The period you are judging over, and your discount rate if you use one.

What comes out

  • Approve now and defer, each with a low case, a middle outcome and a high case.
  • The chance each one comes out ahead.
  • Which figures the answer depends on, so the board can direct its questions at the ones that matter.
  • A full audit trail behind every figure: the inputs, the method, the calculation and the result, kept with the decision.
  • A record of the case as approved, which is what makes a review in two years a factual exercise.

An example

Illustrative example

Blue Quarry Cement, an invented company, with a build to approve

Approve the new line now, or defer it two years?

What goes in
Capital cost
$22m to $38m, most likely $27m
Saving a year once it runs
$4.2m to $9.6m, most likely $6.5m
Months to build and commission
18
Slippage on top of that
0 to 14 months, most likely 4
Cost of deferring, on the capital
6% more
Period judged over
8 years
What comes out
Approve now, low case
$1.5m
Approve now, middle outcome
$11.5m
Approve now, high case
$21.7m
Approve now, chance of coming out ahead
93%
Defer two years, low case
-$11.4m
Defer two years, middle outcome
-$3.8m
Defer two years, high case
$4.0m
Defer two years, chance of coming out ahead
27%

Approving now returns $11.5m on the middle outcome and is still positive in the low case at $1.5m. Deferring two years returns -$3.8m: a loss on the middle outcome, with a low case of -$11.4m. Approving comes out ahead 93% of the time against 27% for deferring. Waiting is the expensive option here and it is the one that would have been called prudent. The figures the answer depends on, in order: Saving a year once it runs, Capital cost, Slippage on top of that. Slippage is on that list, which tells the board where to press: not on whether the saving is real, but on whether the build can be delivered close to eighteen months.

The company and the figures above are invented for this page. The results are a real run of the BrainPower engine on exactly those figures, not a picture of one. Your own figures give your own answer.

Who it is for

Boards and investment committees receiving business cases. Chief executives and owners who sign the approval. Project sponsors who would rather bring the weak point themselves than have it found. Finance teams who have to stand behind the figures after the sponsor has moved on.

Questions people ask

What makes this different from the business case we write now?

Three things: the uncertain figures are entered as ranges instead of as single numbers, deferring is priced as a real option rather than assumed to be free, and the result carries an audit trail instead of a footnote. The arithmetic underneath is ordinary.

Can the board change an assumption in the meeting?

Yes. Change a range, run it again, and the new answer is on screen in the meeting rather than in a follow up email next week.

How is slippage handled?

As an input with a range, entered next to the build period, so it is charged against the years you are judging over. If slippage turns out to be one of the figures the answer depends on, that is the finding, and it is a better use of the board's time than a debate about the saving.

Does it produce a document for the pack?

It produces the figures, the audit trail behind them, and an export of the comparison you can paste into the pack. There is no ready made PDF report of a decision in the product today, and we would rather say that than imply one.

Can we review the decision afterwards?

Yes, and that is the point of the record. The same figures give the same answer, so a review two years later compares what was assumed against what happened, instead of comparing memories.

Keep reading

See it on your own numbers

The live demo runs in the browser with figures you set. A demo with us walks through project approval on a decision you are actually facing.

Run the live demo