What Clients Say After the Work Is Done

Situation

A third-party logistics operator was handling around 200 incoming enquiries per day across email and a web form. Staff spent significant time routing queries to the right department, often manually re-reading and categorising messages that followed predictable patterns.

What We Did

Over seven weeks, we designed and integrated a classification layer that read incoming messages, assigned a category from a defined taxonomy, and routed them into the relevant shared inbox. The taxonomy was built with the operations team, not by us alone. Handover included a short guide for amending the categories.

Outcome

After two months of operation, the client's estimate was that manual routing had reduced by roughly 70% for standard enquiry types. The remaining 30% were edge cases that the team considered appropriate for human review.

Situation

A mid-sized accounting practice had been approached by several software vendors offering AI tools. Management wanted a clear-headed assessment of which, if any, were worth pursuing — without relying on the vendors' own materials to make that decision.

What We Did

We ran a Threadwork Session with three senior staff and the practice manager, mapping how work actually moved through the firm. The written brief identified four candidates for AI integration, ranked by practical readiness, with notes on what data preparation each would require before any build was sensible.

Outcome

The client chose to pursue one of the four items identified, using the brief as the basis for an internal proposal to their management committee. They chose not to engage two of the vendors they had initially been considering, citing the dependency notes in the brief.

Situation

A consulting firm had adopted three AI tools over 18 months, partly through individual staff initiative and partly through management decisions. There was no clear picture of what each tool was doing, who was responsible for it, or how decisions about the tools had been made.

What We Did

We took on a six-month stewardship role, running monthly check-in meetings and maintaining a shared log. The first two months were largely spent documenting what was already in place. By month four, the client had a clear internal governance picture and had made two vendor decisions — one continuation and one discontinuation — on the basis of what the log showed.

Outcome

At the close of the engagement, the client had a documented inventory of their AI tools, a record of all decisions made during the period, and a closing brief with observations on what needed attention over the following six months. They chose not to renew, which we considered a good result.