Final-round presale
Winning back the last call after two meetings had failed
A long-standing account was one bad meeting from closing. I rebuilt the pitch around the client's own operating numbers and a working prototype, and cut every claim we could not source.
- Granted on condition of proving business understanding
- One call leftGranted on condition of proving business understanding
- Sourced claims that survived independent verification
- 21 of 25Sourced claims that survived independent verification
- Slides, rebuilt with the demo deliberately last
- 15 → 12Slides, rebuilt with the demo deliberately last
The client
A large commercial bank in the Middle East, corporate and wholesale division. Already running a mature AI programme in production with its own data function — so the sale was never about whether AI works, only about whether we understood their business.
The engagement
Final-round presale recovery: one deck, one prototype and one meeting, prepared in the week before a decisive call.
A working demo of this build exists.
It is not public yet — ask for access, and where the NDA allows I will send a link or walk you through it on a call.
The problem
Two business meetings had failed to show an independent grasp of how the bank actually makes money, which was the one thing that side of the table cared about. A third and final call was granted on condition of a new face and real domain understanding. The existing deck was written in the past tense — other organisations' benchmarks presented as though they were ours — which is precisely what a sceptical business audience punishes.
What I did
I separated claim from proof and treated that as the entire brief. Every number in the story was re-derived from a primary source; four headline figures were refuted outright and circulated internally as a do-not-use list, which matters more than the twenty-one that survived. The deck was rebuilt on three acts — your world, what changes, how and the proof — with the demo deliberately last and every capability traced back to a metric named earlier in the deck. The prototype was re-cut to the functional pillars of the technical proposal, with everything speculative moved into a visible roadmap section rather than quietly implied as delivered, and external benchmarks relabelled as reference ranges to confirm in discovery rather than as promises.
What was built
A single-narrative prototype of an AI advisor for relationship managers — portfolio health, next best action, credit submission support, pricing intelligence, market insights and an embedded assistant — with every insight carrying a citation, a confidence level and a human approval step, delivered bilingually and mapped screen by screen to the pillars of the platform proposal.
On the table at the end
- Twelve-slide business deck on a three-act arc
- Working bilingual prototype aligned to the technical proposal
- Verified-claims list and an explicit do-not-use list of refuted figures
- Objection-to-artefact map from the two failed meetings
What it changed
Kept a multi-year account and a tier-one platform opportunity alive after two failed meetings, and replaced a benchmark-driven pitch with one where every capability traced to a metric the bank's own relationship managers are measured on.
How it ran
- 01
Read the room, not the org chart
The data-side leadership was already convinced; the business leadership was not. Every recorded objection from the earlier meetings was assigned to a specific screen or slide.
- 02
Evidence pass
Structured research on the market and operating model. Twenty-one claims confirmed against primary sources, four refuted and struck from every artefact.
- 03
Rebuild around their scoreboard
Three acts instead of a capability tour, opening on the client's own published results and the metrics their front line is measured on.
- 04
Realign the prototype to the proposal
Navigation re-cut to the pillars of the technical proposal, with out-of-scope ideas moved to a roadmap section instead of blurring into the demo.
- 05
Honesty pass
Benchmarks relabelled as external reference ranges, and the architectural constraints the demo could not honour stated as constraints.
Other work
All case studies →- Regional banking
One prototype, two markets, no budget in the room
A meeting with no project, no budget and no agenda. I brought a day-in-the-life prototype in the client's own language rather than a capability deck, then shipped their feedback as working features.
- Private banking
A validated proof of concept in ninety hours
An Asian bank wanted to know whether an AI copilot would help its relationship managers. The first iteration took nineteen hours; the whole validated proof of concept took ninety and scored four and a half out of five on acceptance.
- Capital markets
An agentic core with a human gate on every state change
A thirty-year financial house wanted autonomous AI in a regulated business. I designed the agents to propose and people to approve, then priced it against a payback date instead of a benefit narrative.
Something similar on your plate?
Thirty minutes, no deck. I will tell you whether it is worth doing at all.