BREAKING Madhu Tera builds an AI command center for finance teams /// TERA reports 1,200+ companies onboard /// From State Street back office to founder's chair /// Andhra University → Kansas State → Harvard ALM /// HQ in Acton, Massachusetts - not Silicon Valley /// Meet Finpilot, TERA's in-product AI for spend /// BREAKING Madhu Tera builds an AI command center for finance teams /// TERA reports 1,200+ companies onboard /// From State Street back office to founder's chair /// Andhra University → Kansas State → Harvard ALM /// HQ in Acton, Massachusetts - not Silicon Valley /// Meet Finpilot, TERA's in-product AI for spend ///
Founder // Solutions Architect // Reformer

Madhu
Tera

He spent years wiring the back-end systems of Fidelity and State Street. Then he walked out and started building the bank those institutions never quite got around to making.

Founder & CEO, TERA AI Finance Spend Management Acton, MA
EST. 2018 MT tera.cloud
1,200+
Companies on TERA
2018
Year Founded
3
Universities, 3 Countries
~$7M
Reported Revenue, 2024
The Dispatch

The finance team's spreadsheet has a rival, and its name is Finpilot.

Open TERA and the first thing that greets you is not a ledger. It is a chatbot for money. Madhu Tera named it Finpilot, and the joke is only half a joke: he wants it to fly the back office the way an autopilot flies a plane. Cards, expenses, invoices, vendor payments, approvals, fraud checks - the unglamorous plumbing that keeps a company solvent - all funneled into one screen that talks back.

This is the pitch of TERA, the company Madhu founded in 2018 and still runs as CEO from an office in Acton, Massachusetts. The product calls itself "the ultimate AI command center for finance teams." Strip the marketing and the ambition is plainer and bolder: take the dozen tools a growing business uses to move money and collapse them into a single, intelligent layer. Corporate and prepaid cards with spending controls. A UPI wallet for business payments. Expense management. Accounts payable. Procurement. All of it watched over by AI that is supposed to notice the duplicate invoice before a human ever does.

By the company's own account, more than 1,200 companies now run on it - hospitals and manufacturers, e-commerce shops and agritech outfits, logistics firms and financial-services players. Names like Spinny, Medicover, Wheelocity and Scoops India sit on the customer roster. For a startup based two hours from the nearest venture-capital cluster, that is a quietly stubborn kind of traction.

"Over 91% of small businesses feel underserved by their banks." - that single statistic, lifted from his own essays, is the closest thing TERA has to a founding grievance.

What makes Madhu interesting is not that he is building fintech. It is where he is building it from. Most founders who attack banking are outsiders throwing rocks at a fortress. Madhu spent years inside the fortress, holding the blueprints.

"An AI-first approach means designing intelligence into the foundation, not bolting it on later."

- Madhu Tera, on building fintech systems

Origins

An architect learns where the money gets stuck.

Before TERA, Madhu was a solutions architect - the person companies hire to figure out how the pieces of a giant system fit together without falling over. His resume reads like a tour of American finance's machine room. He was an Assistant Vice President and Solutions Architect at State Street, one of the oldest custodian banks in the country. He worked as a Consultant Solutions Architect at Fidelity Investments. Earlier still, he wrote software at Oat Systems, a division of Checkpoint Systems.

That kind of work teaches you something most fintech founders learn too late: the problem is rarely the front end. It is the seams. The handoffs between systems that do not talk to each other. The reconciliation that takes three days because the data lives in four places. Madhu spent a career staring at those seams, and TERA is, in a sense, his argument about how to stitch them shut.

He thinks in writing. On Medium he has published essays with titles only an architect would love - "Low-code Architecture for Any Application," "Framework for User Configurable Business Workflow Automation," "Building Blocks for Creating Low-code BI Dashboards." Read them and you can see the TERA product taking shape years before most customers had heard the name: configurable workflows, intelligence baked into the foundation, dashboards a non-engineer can bend to their will.

In one 2021 essay about corporate cards, he does something a pure engineer would never do. He reaches past the spreadsheets to Edward Bellamy's 1887 utopian novel "Looking Backward," which imagined a universal credit card decades before the plastic version existed. It is a small tell. The man building payment rails likes to know where the idea of credit came from in the first place.

The Map

Three countries, one through-line.

Start

Andhra University, India

Earns a Bachelor of Technology - the engineering grounding.

Then

Kansas State University, USA

A Master of Science. The map jumps an ocean.

Then

Harvard University

An ALM - a master's in liberal arts. The engineer picks up a humanist's degree.

Career

Oat Systems → State Street → Fidelity

A decade-plus learning exactly where money gets stuck inside big institutions.

2018

Founds TERA

An AI-powered finance platform, headquartered in Acton, Massachusetts.

2020

Seed Funding

TERA raises early capital to build out the platform.

2021

The Essays

Publishes a run of writing on AI-first fintech, low-code, and corporate banking.

2024

Scale

Reports roughly $7M revenue, a ~51-person team, and 1,200+ customers.

The Product

One screen to run the money.

01 // CARDS

Corporate & Prepaid Cards

Spending controls built in, so the limit lives in the card instead of in a memo nobody reads.

02 // PAY

UPI Wallet

Business payments that move at the speed customers now expect by default.

03 // SPEND

Expense Management

Categorization, approvals and policy enforcement, automated rather than chased.

04 // AP

Accounts Payable

Invoice automation and vendor payments, with AI watching for the duplicate nobody caught.

05 // BUY

Procurement

Vendor and purchase workflows folded into the same command center.

06 // AI

Finpilot

The in-product co-pilot for spend insights - meant to replace the spreadsheet, not assist it.

"Small businesses generate 44% of U.S. economic activity, yet over 91% feel underserved by their banks."

- Madhu Tera, "Extreme Makeover: Business Banking Edition"

The Credentials

The unlikely transcript.

Bachelor of Technology

Andhra University

Master of Science

Kansas State University

ALM, Liberal Arts

Harvard University

The Margins

Three things worth knowing.

01

His finance platform sits in Acton, a quiet New England town - not San Francisco, not New York. The product travels further than the zip code suggests.

02

He helped build the back-end systems at the very institutions TERA now aims to outpace. The reformer used to be the architect.

03

To explain the future of corporate cards, he cited a novel from 1887. The engineer reads further back than the engineering.

The Thesis

"AI-first" is easy to say and hard to mean.

Every software company in 2026 claims to be AI-first. Madhu made the claim in 2021, and in his telling it carries a specific, almost stubborn meaning. In "AI-First Approach to Building Fintech Systems," published in the Geek Culture tech collection, he draws a line most product teams blur: there is a difference between a system that has AI features and a system whose foundation is AI. The first sprinkles intelligence on top - a fraud flag here, a chatbot there. The second assumes from the first schema that machines, not people, will do most of the noticing.

You can see why that distinction would matter to an architect. Bolting AI onto a legacy stack is exactly the kind of seam Madhu spent his career untangling at State Street and Fidelity. Patch jobs accumulate debt. A foundation does not. TERA's promise - automated invoice matching, smart expense categorization, duplicate-invoice detection, fraud detection - only holds if the intelligence lives in the plumbing rather than the paint.

There is a second idea running underneath all of it: low-code. Across several essays, Madhu argues that the people who actually run finance - controllers, AP clerks, founders doing their own books - should be able to bend their software without filing a ticket with engineering. Configurable workflows. Dashboards a non-engineer can build. The unspoken target is the spreadsheet, the duct tape of every finance department on earth. TERA's tagline does not whisper this. It says it out loud: replace the spreadsheet with something that thinks.

The spreadsheet is not a tool. It is a confession that the real software never showed up. TERA is the bet that it finally has.

It is worth pausing on the audience. TERA's customer list - Spinny, Medicover, Wheelocity, Scoops India - leans toward fast-growing companies in markets where digital payments moved faster than the back office could keep up. The platform supports UPI wallets and multi-currency spend, the kind of features that signal a company thinking globally from a small office in Massachusetts.

That global-from-Acton posture is the quiet contradiction at the center of the story. A founder trained at Harvard and seasoned at Boston's biggest financial institutions, building payment infrastructure for businesses scattered across continents, from a town better known for colonial history than cloud software. There is no glass tower in the picture. There is a thesis, a product, and 1,200-plus companies that decided the thesis was worth wiring their money through.

What he is chasing is not a feature roadmap. It is a reordering of who gets good banking. The large enterprise has always had a treasury team, a forecasting model, a fraud desk. The small and mid-size company had a spreadsheet and a prayer. Madhu's wager is that AI flattens that gap - that the intelligence once reserved for the Fortune 500 can be packaged, automated and handed to a 40-person manufacturer in the form of a screen and a co-pilot named Finpilot.

Whether the wager pays off is a story still being written. But the shape of it is unmistakable. The architect who learned where the money gets stuck decided the most interesting thing he could do was unstick it for everyone who never had an architect of their own.