From Alert to Answer: How Tormach Used AI to Solve the Knowledge Gap in Technical Support

Tormach manufactures professional CNC machines sold to small-scale manufacturers, machinists, and fabricators who depend on fast, accurate support to keep their operations running. Their technical support team includes seasoned engineers who have spent years solving hard problems, and documenting what they learned. The knowledge is there. It lives in Jira tickets, Confluence pages, and years of resolved cases. The challenge was making it accessible at the moment a rep needed it, not just to the people who wrote it.

That challenge is common across manufacturers of complex technical equipment. Up to 70% of critical operational knowledge never makes it into a format that front-line support can act on quickly. Even when it does, generic AI tools return semantically similar results with no awareness of machine model, serial number range, or revision history. In CNC support, a vague answer that doesn't account for a specific controller generation wastes time and erodes customer trust, regardless of how good the underlying knowledge is.

This session is a case study in how Tormach deployed Strudel's Knowledge Scout to make their team's expertise accessible across every support interaction. Rather than replacing institutional knowledge, Knowledge Scout surfaces it, grounding each support interaction in Tormach's own ticket history, BOMs, and revision-aware documentation. The result was a system that could give a support rep not just general guidance on a CX-7 controller fault, but a direct reference to two prior tickets with matching symptoms that traced to a power brick outputting 10–12V instead of 24V, with the exact ticket numbers to prove it.

Speaker: Brian Kaufman, Co-Founder and CPO of Strudel, walks through what the implementation looked like, what didn't work at first, and what the outcomes were, including accuracy benchmarks, escalation reduction, and what it meant operationally for every rep to have the team's collective expertise within reach from day one.

Key takeaways:

• Why generic AI fails in technical manufacturing support and what "grounded" means in practice

• How revision-aware, model-filtered context amplifies the expertise your team has already built, making it available to every rep, on every ticket

• How ticket trend analysis enabled Tormach to get ahead of fleet-wide issues before they multiplied

• What the onboarding and integration process actually looked like (the honest version)

• Lessons applicable to any manufacturer whose support quality depends on deep institutional knowledge, and wants to make sure it doesn't walk out the door