IVA sees the job through the technician’s glasses or phone, listens to what they say, and fills in the work order in the system you already use. The technician checks it and signs. Nobody types a report in the van.
No custom hardware. Pilots open for telecom field operations and industrial maintenance teams.
Enterprise AI can only reason over what someone wrote down. In field work, writing it down happens at the end of the shift, from memory, on a phone, by someone who would rather be on the next job. The fault, the fix and the evidence get compressed into “issue resolved.”
The distance between the work done and the work recorded is what IVA closes. We call it the capture gap.
The technician wears commercially available AI glasses, or a phone in a chest mount. IVA captures photos, meter readings and their spoken notes while their hands stay on the work.
IVA matches what it saw to the asset, your fault codes and your procedure. When something is missing, a reading, a serial number, a part, it asks by voice before the technician leaves.
IVA drafts the completed record in your field-service, maintenance or trouble-ticketing system. The technician reviews it on site and signs. Then it’s submitted.
Tower, fibre and site technicians whose work orders feed network operations and trouble-ticketing.
Maintenance crews on plant floors, in warehouses and on robot fleets, where the record feeds the maintenance system and the next shift.
We run IVA with one crew at one site, against a baseline we measure together in the first two days. You keep the readout whether or not you continue.
Shadow the crew and measure how records get written today. Agree the target fields with your operations lead.
The crew works with IVA. Records go to a sandbox or your live system, your choice, with a person signing every one.
Before and after on every measure, failures included. Then you decide whether to continue.
What we commit to in every pilot. The full policy is on the trust page.
Enterprise AI is starving at the edge. The most valuable operational knowledge is created by people whose hands are busy, and it never reaches the system of record.
From The Capture Gap, a memo on why field work is where enterprise AI goes next. Draft 0.9, October 2026.
IVA is built by Shivashish Borah, founder of AI Humane Technologies. His career has been where software meets operations: telecom order and service systems, presales architecture, and selling autonomous robots into industrial sites. Each taught the same lesson. A system of record is only as good as what the field puts into it.
No. IVA is designed to run on commercially available AI glasses, or on a phone in a chest or helmet mount. We choose hardware for your site during the pilot.
Field-service management, maintenance management (CMMS) and trouble-ticketing systems that have an API. We build the connector for your system during the pilot, and that work carries over to every team on the same system.
No. IVA drafts the record and asks for anything missing. A named person signs every record before it reaches your system.
Into storage you control, for as long as you set, and nowhere else. The trust page covers retention, access and what happens to bystanders.
The original master plan from July 2026 is archived here. It described a long path from consumer glasses to an enterprise platform. We’ve narrowed the starting point to field work, where the value can be measured in two weeks. The memo explains why.