← Presentations
Bonsai Labs

Operational AI Discovery Read-out

The headline

The model held up everywhere

BONSAI LABS

What we confirmed

  • Every team independently converged on the patterns we proposed: speed bumps over hard blocks, human-in-the-loop, exception-only processing.
  • Service and Back Office were confirmed day one as the priority and absorbed nearly all onsite time.
  • Billing confirmed its lower-yield, do-ops-first ranking.
  • We are at nineteen use cases with a solid handle on most; the work is shifting from discovery to prioritization.

Adoption is a workstream

Glia already auto-summarizes calls and reps rate it highly, yet after-call work is flat and every interviewee discovered the Glia-to-Salesforce gap during our sessions. Part of the ROI is unlocking what is already paid for.

Fix intake first, concretely

The form-completeness check is the pattern-setter: same-system, self-contained, no PII, no writes. It directly attacks the 13 percent rework rate and sets the template for everything after.

The constraint is data lineage

Trading ops, OCIO billing, and commissions are the same problem: people stitching data off-platform in Excel when the source often already feeds Databricks. The question is automate the human versus fix upstream.