Exploratory conversation. Friction first, models second.
AI-assisted operational improvement
Who this is for
Operators and small organisations stuck in repetitive coordination, manual handoffs, or tool sprawl.
Problem
“Add AI” without redesigning the work usually creates another dashboard. The target is friction, not model fashion.
What a discovery conversation covers
- One named bottleneck you are willing to examine.
- A lightweight friction map (process vs software vs information stall).
- A leverage check: when automation or an agent surface helps, and when it does not.
- Engineering patterns we can show from products we ship (below), as building blocks, not your ROI.
- Data and secrets boundaries before any deeper access.
Engineering building blocks (from Trip and Move)
- MCP tools that call the same application actions as the UI. Agents cannot bypass tenancy or authorisation.
- Vision-assisted recognition for inventory contents with bring-your-own-key providers (Move).
- Hybrid search: full-text + trigram + vector embeddings (Move).
These show how we build agent-facing and AI-assisted software. They do not prove consulting delivery history or client outcomes.
See Trip and Move in the Showroom.
Example outputs (only if we continue under a separate agreement)
Friction map; prioritised interventions; an implemented workflow or application slice; operating notes.
Boundaries
- No invented ROI percentages or named client outcomes.
- A product demo is not your operational result.
- We do not replace regulated professional advice.
- We will not ask you to paste secrets into an unsolicited form.
Start a conversation
Email joel@vogonpartners.com with one repetitive bottleneck and your data-access constraints.
Subject suggestion: AI-ops: discovery
We will say honestly if software or AI is the wrong lever.