~/triam/about.essay · an argument in six paragraphs · bengaluru, in

Every founder, accountant and CFO in India has the same hidden tax: reconciling between tools that don’t talk to each other. Tally for the books. Zoho or Razorpay or Shopify for the AR. An Excel sheet for inventory because the warehouse manager won’t use anything else. A second Excel sheet for FP&A. WhatsApp for approvals. Email for the bills.

On a good month, this works. On a bad month — a closing month, an audit month, a fundraising month — you spend more time arguing about which number is true than improving the business. The accounting industry papered over the problem with the word “integration”.

sprawl.mapfig 01
tally.erpthe bookszoho / razorpaythe ARinventory.xlsxthe warehousefpna.xlsxthe modelwhatsappthe approvalsemailthe billstriam.ledgerone truth
We don’t think integration is the answer. We think the answer is owning the ledger.

So we built a different shape: one double-entry GL that AR, AP, banking, inventory, payroll and FP&A all post into directly. Not via webhook. Not via overnight ETL. Same database, same audit trail, same RBAC. When your warehouse manager moves stock, the GL knows. When payroll runs, the GL knows. When Trex closes the month, it’s pressing the same button your accountant would.

ledger.hubfig 02
general ledgerdouble-entry

We picked India first on purpose. Indian businesses live in a particular reality — GST, TDS, multi-state tax, Tally lock-in, CA workflows, statutory close calendars. Building for that, not adapting an American product to it, changes a lot of small decisions. ₹ formatting. HSN codes. Three-level COA. Indian banking statement formats. Tally migration as the wedge feature, not a checklist item.

india.primitivesfig 03
first-class primitives
GSTTDSHSN codesmulti-state tax₹ en-IN formattingthree-level COACA workflowsstatutory close calendarIndian bank statementsTally migration
not a settings page · not a regional override · the spine

We picked AI second, also on purpose. The agent space is loud right now and most of it is theatre — a chatbot reading the API of an accounting system. We took the other path: build the ledger first, then build an agent that sits inside it. Trex can close the books because Trex has the same database access as your accountant. Approvals still go through humans. Audit trail still holds. But the keystroke labour in between — that’s the agent’s job now.

trex.sessionfig 04
trex.sessionproposing
$ trex close --month may
→ journal proposals drafted
→ routed for human approval
✓ audit trail holds

three convictions · defended at length

What we believe.

Integration is a tax. Ownership is the fix.

Every “integration” is a place your books can drift. We own the modules instead of integrating with them. Yes, that’s more software to build. Yes, it’s the only way the close-the-books promise is real.

Agents work when they have skin in the database.

Bolting a chatbot on an integrations layer gets you a worse version of Excel. Embedding an agent in the ledger, with the same RBAC and the same audit trail your humans have, gets you something useful.

India is the product, not the localisation layer.

We built GST, HSN, TDS, three-level COA, multi-state tax, and Tally migration as first-class primitives. Not as a settings page. Not as a regional override. As the spine.

the ecosystem

Part of the Triam Platform

Triam Accounting is the core of the Triam platform. Inventory, billing and every Trex action flow into the same double-entry ledger — reconciled automatically.

Triam Accounting

coreyou are here

AI-native double-entry accounting — the core platform.

Triam Tangible

module

Agentic inventory management for product companies.

Antichurn

module

SaaS billing & churn prevention — recurring revenue, posted to the GL.

one ledger · every transaction reconciled automatically
argue.sessionOPEN

We’re in Bengaluru. Reachable. Happy to be wrong about specifics, less happy to be wrong about the shape.

status: reachablelocation: bengaluru, inchannel: open