A consulting practice was asked to assess fifty finance and HR processes across a manufacturing group's shared service centres in four countries. The previous assessment of the same scope had taken five consultants thirteen weeks and produced fifty slide decks nobody could compare. This time the team ran discovery on Pointee: eight weeks, three consultants, 120 consultant-days — and every process came out of it as a model ready to build.

The situation

Documentation was thin and uneven: 38 SOPs of varying age, 27 screen recordings, and process owners in four countries who each ran "the same" process differently. Nothing agreed on volumes, handling times or where the exceptions went.

The earlier assessment had cost 325 consultant-days and delivered fifty decks written in fifty different formats. Effort was estimated one way in Poland and another way in Portugal, so the numbers could not be added up, and the automation team could not tell which processes were worth building first.

What the team did with Pointee

Wk 1–2 Ingest. SOPs and screen recordings went in; AI-drafted process plans came back and were approved by the consultant — 50 first drafts in 6 working days.
Wk 2–5 Interview. 142 AI-run interviews by text, voice or screen-share, 60 of them as a meeting bot sitting in Teams calls. Notes were written live, so nothing waited on a consultant to type them up.
Wk 4–7 Quantify. Volumes, handling times and exception rates were entered once; effort, FTE, loaded cost, rework and straight-through processing were computed with one set of definitions across all fifty processes.
Wk 6–8 Deliver. Fifty branded as-is reports in three languages, twelve priced To-Be maps, and twelve draft processes generated straight into the designer.

The results

ResultsBeforeWith PointeeChange
Program duration13 wks8 wks−38%
Consultant-days325120−63%
Consultants assigned53−2
Stakeholders interviewed54142+163%
Findings cited to a source100%new
Hidden-work gaps found01423 of 84 FTE automatable
Handed to the build phaseslides12 drafts

What changed in the deliverable

  • One set of definitions. Effort, FTE, loaded cost, rework and STP were computed the same way in all four countries, so the fifty processes could finally be ranked against each other.
  • Every finding cited. 100% of findings trace back to a named source — an SOP, a recording or a specific interview — which is what let the client sign the business case without a second round of validation.
  • Hidden work surfaced. 14 gaps that no SOP described accounted for 23 of the 84 FTE in scope — and were automatable.
  • The model went into the build. Twelve processes left discovery as draft automations in the designer rather than as slides to be re-modelled from scratch.

What the practice took away

"We used to sell discovery on senior hours and deliver a deck. This time the deck was a by-product and the model went straight into the build. The client signed phase two off the same numbers."

Conclusion

Three consultants covered a scope that previously took five, interviewed nearly three times as many stakeholders, and finished five weeks earlier — with a comparable, cited, automation-ready model at the end of it instead of a slide library.