A thesis by Arya Shah · TechSquare Ventures
Deskless AI
The zero-onboarding enterprise layer for frontline workers.
Eighty percent of the global workforce does not sit at a desk. They keep planes moving, packages flowing, and warehouses stocked - and the software built for them is software they never open. The thesis: meet them where they already are. SMS and iMessage AI agents that read and write to the systems of record, with nothing to install and nothing to learn.
Enterprise software was built for the desk. The frontline got left behind.
Slack, Teams, and Salesforce assume a laptop, a login, a training session, and a quiet moment to navigate a dashboard. A baggage handler on a ramp, a driver mid-route, a supervisor on a loading dock has none of those. So adoption stalls, seats go unused, and the work that actually happens never makes it into the system of record. The gap is not a feature gap. It is an interface gap.
Apps the frontline never opens
- Requires a managed device, an install, and a login
- Onboarding and training before anyone is productive
- Per-seat licensing on workers who log in once a quarter
- Adoption stalls; real work stays on paper and in heads
An agent that lives in the text thread
- Works on any phone - the Messages app is the whole client
- Zero onboarding: if you can text, you can use it
- Reads and writes the systems of record in plain language
- Reliable offline-first delivery; nothing to forget to open
The whole product is a text message.
Pick a frontline scenario. The worker texts in plain language; the agent parses it, writes to the system of record, and texts back a confirmation - the same loop, whether it is a torn bag tag, a blocked route, or a bad pallet count.
Three properties the desk-bound stack can't match.
Zero onboarding friction
No app to install, no account to provision, no training to schedule. The worker texts a number and is productive on minute one. IT keeps its existing systems; the agent speaks to them.
Instant database reads & writes
The agent translates plain language into typed actions against the ERP, WMS, or routing system - and reads back the same way. The text thread becomes a live, two-way interface to the system of record.
Offline-first reliability
SMS degrades gracefully where data apps fail: a ramp, a basement dock, a rural route. Messages queue and deliver, so the report still lands when connectivity is thin.
What zero-friction logging is worth at scale.
Frontline teams lose real time to paperwork and double-entry. Estimate the time the agent reclaims by capturing the report at the moment it happens, in one text.
Assumes the agent removes 70% of daily logging time over 250 workdays, at a fully-loaded $32/hr. An estimate for discussion, not a quote.
Discuss this thesis
If the frontline is the next enterprise interface, let's talk.
Arya Shah is exploring this space with TechSquare Ventures - founders building for deskless workers, operators with the problem, and anyone with a sharp disagreement. Reach out.
Arya Shah · TechSquare Ventures