Small models for industry, starting with MSME quotations
Beyond GTM
The seams between revenue teams aren’t the only place a narrow, repetitive, high-volume decision quietly costs money. The same pattern shows up wherever a business makes the same call by hand, hundreds of times a month, from scattered tools and memory.
Loglens Labs takes one approach to all of them: small, purpose-built models that run on the business’s own data, inside its own infrastructure, priced for its own volumes. Each one is trained for a single operational job, not built as a general assistant. We ship one at a time, starting where the money leaks fastest.
Starting with quotations for Indian MSMEs
For an MSME manufacturer, the quote is where revenue is won or lost. Requests arrive by email and WhatsApp, mixed in with everything else. Someone has to spot them, read the spec, cost the job by hand and write the reply, and that takes hours while the buyer is asking competitors too. Then the quote often goes out once and is never followed up.
We are building a quoting decision layer for that job:
- Intelligent email. Picks out quote requests from the inbox, or from WhatsApp.
- Quote request. Reads the request and extracts the job spec.
- Quoting decision layer. Small models cost the job from the shop’s machine rates, live material prices, catalogue and past similar jobs. The owner adjusts and approves, and every correction, with its reason, calibrates the next quote.
- Automated send. Sends the approved quote and follows it up until the customer answers.
It fits the tools MSMEs already use (email, WhatsApp and Excel) with no rip-and-replace, and it runs on the business’s own quote and email history. It is in build now with a prospective design partner, an MSME lubrication-systems manufacturer.
On the research bench
The next operational jobs we are looking at training small models for:
Receivables and collections
A model that watches invoice ageing and payment history could sequence the follow-ups, knowing which customer needs a gentle reminder and which needs a call, without a full ERP.
Repeat-order retention
Most MSME revenue is repeat business. A model over past order cadence could flag the customer who is overdue to reorder, before they quietly move to another supplier.
Compliance and GST operations
High-volume, deadline-driven filing and reconciliation work that is sensitive to where data goes: a bounded job a small model can run cheaply on the business’s own data.
Machine uptime
For manufacturers whose machines are instrumented, a small model over telemetry (vibration, temperature, current draw, cycle counts, error codes) could flag a failure before it stops the line. It would run on-premise, so the machine data never leaves the factory floor.
Have a job like this in your industry?
We choose what to build next from the problems businesses bring us: the repetitive, high-volume decision that quietly costs deals or money. If that’s a job in your business, tell us what it is.