Every industrial operator we work with owns more software than they use. A CMMS no one likes. A historian the reliability team tolerates. A document repository where inspection reports go to die. The software did what it was supposed to do — it captured data. What it didn't do was make the data answerable.
Claude changes the default answer to "who's going to open that tool and find out."
What "Claude is the UI" means in a plant
When a maintenance planner wants to know which pumps in the Angola field have run over 80% of their MTBF, they don't open the CMMS. They ask. Claude — wired to the CMMS, the historian, and the last three years of inspection PDFs through MCP — reads the question, walks the data, and hands back an answer the planner can act on.
The CMMS is still the system of record. Claude is the system of intelligence. Operators stop thinking in terms of tools; they think in terms of questions.
The software you already own doesn't go away
This is where most industrial AI pitches miss the point. The pitch is usually "replace your CMMS." Operators rightly ignore it. Swapping a CMMS is a multi-year, seven-figure project that the plant manager did not wake up wanting to run.
What operators will do is bolt on intelligence that respects the systems already in place. Kaysee — our reliability platform — does not replace the CMMS. It reads from it, writes to it, and lets the operator talk to it in plain English. The CMMS admin still has a CMMS. The planner now has a planning partner.
The implementation shape
Every Cortland engagement that ends with "Claude is the UI" starts with three questions:
- What data do your people want to ask questions of, and where does it live?
- What systems do they write decisions back into, and who owns them?
- What does "in the loop" look like for the people whose jobs this changes?
The first question sizes the MCP integration surface. The second sizes the action surface — the write path, not just the read path. The third is the one that actually determines whether the project ships.
What we ship in the first 90 days
A Walk engagement answers the three questions and delivers a roadmap. A Run engagement delivers a working Claude-backed surface wired to at least one system of record, and at least one human-in-the-loop review pattern. A Sprint engagement keeps that surface current as the MCP landscape evolves and the operator's data shifts underneath it.
None of this is a plug-and-play sale. Every industrial operator has a different set of legacy systems, a different tolerance for automation, and a different regulatory floor to clear. What travels between engagements is the thesis: Claude is the UI, the systems already in place are the plumbing, and our job is to connect them in a way the operator trusts.
That's the work.