Less admin.
More selling.
Finding the bottleneck in a sales process — and turning it into a small, measurable AI pilot proposal.
- MY ROLE
- Initiator, analyst and presenter
- STAGE
- Approved to scope in principle
- PERIOD
- Autumn 2026
Where does a salesperson's time go?
A walkthrough of a real deal with the sales director exposed administrative work and handoffs around the sales process: compiling an opportunity form, waiting for an estimate and moving contract paperwork forward.
In that illustrative deal, preparing the opportunity form took about half a day; the estimate could wait up to five working days; and contract paperwork could stall for up to two weeks when the relevant person was unavailable. These were observations from one deal, not company-wide averages.
The problem wasn't how quickly AI could write a proposal. It was how to prepare a request that the next person could act on.
Automating the wrong step is still the wrong solution.
I compared a public sales-automation benchmark with IDX's actual process and existing AI-assisted RFP response capability. The bottleneck pointed to preparation and handoffs, not simply document generation.
So I narrowed the proposal: one controlled use case, tested alongside the manual process. AI can reduce preparation effort; it cannot manufacture estimate-team capacity or eliminate every wait.
One real problem. Two weeks.
The proposed assistant would transform approved deal information and a client conversation into an estimate-ready brief. A salesperson would check it and submit it using IDX's existing process.
Client conversation
Notes and approved deal details
Estimate-ready brief
Opportunity form, scope summary, open questions and estimate request
Review & submit
Salesperson checks the output and uses the existing workflow
AI suggests. People decide.
The proposal was designed with clear operational boundaries and success measures before any build commitment.
A handful of cases can demonstrate feasibility, but cannot establish a reliable return on investment.
From discovery to a decision.
I initiated the work, walked through a real sales case, analysed the bottlenecks, reviewed internal and external reference points, specified the narrow pilot, designed the measurement approach and presented it to Nordic management.
I used Claude for research, drafting and critique, while retaining responsibility for the framing, constraints and executive narrative. The decision deck was distilled from a longer explainer into nine main slides plus an appendix.
Approval to scope. Not a claimed saving.
Management gave the go-ahead to scope a two-week pilot in principle, conditional on feasibility, security and budget. The pilot has not yet run and no time-saving or ROI result is claimed.
Next steps are to write a one-page scoping note, baseline five to ten recent estimate requests, and confirm technical, security and budget ownership.