Isha Juneja left Microsoft, sold a pet-tech startup, then built ContraVault AI

The first paying customer for ContraVault AI was won inside a cooling towers company. Isha Juneja and her co-founders had built a rough chatbot that could read a tender document and answer questions about it, and they took it to Paharpur Cooling Towers for a live demo. The bid team there spent its days buried in hundred-page government and infrastructure tenders, hunting for the clause that could sink a project. The demo worked. Paharpur signed. That single yes told Juneja she had found a real problem hiding in plain sight, the kind that billion-rupee construction firms live with quietly because nobody had bothered to fix it.

Early life

Isha Juneja grew up in India and came into entrepreneurship the long way around, through engineering school and a corporate product career rather than a family business. She has spoken publicly about marrying into a family connected to the infrastructure trade, and it was that world, the one of tenders, contractors and bid deadlines, that later handed her a company. Much of her personal early life is not a matter of public record, and this profile sticks to what she has said on the record in interviews and on her own channels.

Education

Juneja studied engineering at the Netaji Subhas Institute of Technology in Delhi, one of the capital’s better-known technical colleges. She followed it with an MBA at XLRI Jamshedpur, finishing the two-year programme in 2022. The combination matters to how she works. The engineering grounding lets her sit in product conversations with technical co-founders without losing the thread; the business school training shows up in how ContraVault talks about return on investment to procurement heads rather than selling them on the novelty of AI.

Career before ContraVault AI

Before she was a founder, Juneja built her career capital inside large organisations. She worked at NatWest and then at Microsoft, where she was a product manager on the team behind Excel at the company’s India development centre. Shipping inside a product as widely used as Excel is a particular kind of training: you learn to reason about millions of users, edge cases and backwards compatibility.

Then she jumped. Juneja was part of the founding team at UltraPawz, a venture building Fitbit-style smart trackers for dogs and cats, the sort of wearable that watches a pet’s activity and location. The team raised a little over $250,000 in pre-seed money and built real hardware, which is hard. UltraPawz was eventually acqui-hired by Supertails.com, one of India’s larger pet-care companies. It was not a blockbuster exit, but it was an exit, and it taught her what it feels like to build something from nothing, run out of road, and find a home for it anyway. That experience is the bridge between the Microsoft years and ContraVault.

Why she started ContraVault AI

The spark was specific. Juneja watched her father-in-law work through tender documents that ran to a thousand pages, the dense, repetitive, legally loaded paperwork that infrastructure companies must digest before deciding whether to bid on a project. A single missed clause, a penalty term or an unworkable delivery schedule can turn a winning bid into a money-losing one. Big contractors throw teams of people at the problem and still take days to reach a go or no-go call.

Generative AI had just become good enough to read and reason over long documents, and Juneja saw the match. In 2024 she co-founded ContraVault AI in Bengaluru with Sayan Sen, who became chief executive, and Tanmay Juneja, the chief technology officer. She took the chief operating officer role. The pitch was plain: point the model at a tender, and have it surface the risks, summarise the obligations and help a bid team decide fast. This was not a chatbot looking for a use case. It was a use case that had been waiting years for the technology to arrive.

The early struggle

Selling software into construction and heavy infrastructure is not glamorous, and it is not quick. These are firms with long memories and longer procurement cycles, where a new vendor has to earn trust before anyone lets a model near a live bid. ContraVault’s early answer was to make the product earn that trust in a demo, the way it did at Paharpur, rather than in a slide deck.

The harder technical problem was reliability. An AI that confidently invents a clause is worse than useless to a bid team, because a wrong answer carries the same penalty as a missed one. So the founders broke the work into narrow, specialised models instead of asking one general model to do everything: separate engines for the go/no-go decision, for risk detection, for reading and negotiating clauses, and for drafting the proposal itself. Splitting the job that way kept each step checkable and cut down the hallucinations that would have killed the product in this market.

Funding & revenue

ContraVault’s first institutional cheque came in May 2025, when Titan Capital led a seed round of about ₹5.1 crore, roughly $600,000. The round pulled in a notable list of operator-angels: Rajiv Ahuja of Navayuga, Haresh Chawla of True North, Jaswinder Ahuja of Cadence, Dilip Kumar Khandelwal formerly of Deutsche, Tracxn’s Abhishek Goyal, and backers connected to Paharpur. “This fundraise is a critical milestone for us as we step into a greenfield opportunity,” Juneja said at the time.

A year on, the company raised again. In June 2026 ContraVault closed a pre-Series A of $3.1 million, about ₹29.3 crore, led by Chiratae Ventures with the Titan Capital Winners Fund returning. That takes total disclosed funding to roughly $3.7 million. The startup said it would use the money to push into the United States, where it has already set up a subsidiary, and to keep building product and team.

The traction underneath the funding is the more interesting number. By mid-2026 ContraVault was serving around 40 customers in India, a roster that includes Tata’s Voltas, Shapoorji Pallonji, the Adani Group and Sterling & Wilson. The company says its customer base doubled in a two-month stretch. On the product’s own metrics, clients cut the time spent working a single tender by 60 to 70 percent and place 30 to 35 percent more bids as a result, because the cost of evaluating each opportunity drops.

Isha Juneja’s net worth & ownership

There is no credible public figure for Isha Juneja’s net worth, and any number you see quoted for an early-stage private founder should be treated with suspicion. What can be said is grounded. She is a co-founder and the chief operating officer of a company that has raised about $3.7 million across two rounds and is still privately held, which means her stake is an equity position in a young, growing business rather than realised wealth. She also had a prior exit when UltraPawz was acquired by Supertails, though the terms of that deal were not disclosed. The honest summary: her value is tied up in ContraVault, and it will be worth whatever the company eventually proves to be worth.

Growth strategy

Three choices stand out. The first is the decision to sell into a deeply unsexy, deeply painful workflow. Tender analysis is not a market people pitch on stage, and that is exactly why it was open. The second is the specialised-model architecture, which trades the demo-friendly magic of one all-knowing assistant for the boring reliability that an Adani or a Shapoorji Pallonji actually needs before they trust it. The third is going to the United States early, while the India business is still young, on the bet that bid-and-proposal pain is universal across construction, energy, defence and aerospace, and that the company that owns the workflow in one country can carry it to another.

What Isha Juneja’s story teaches Indian founders

The lesson is about where good problems hide. Juneja did not go looking for a market; she noticed someone she knew losing a weekend to a document, and recognised that the technology to fix it had only just arrived. The second lesson is that an unglamorous, hard-to-sell market can be a moat rather than a warning sign, because the same friction that scares founders away keeps competitors out once you are in. And the third is that a modest first exit is worth more than its headline. UltraPawz did not make anyone rich, but it gave Juneja the scar tissue and the credibility to raise for ContraVault and to sell to companies that do not hand their trust to first-timers easily.

Frequently asked questions about Isha Juneja and ContraVault AI

Who is Isha Juneja?

Isha Juneja is an Indian entrepreneur and the co-founder and chief operating officer of ContraVault AI, a Bengaluru startup that uses AI to analyse tenders and proposals for infrastructure companies. Before ContraVault she was a product manager at Microsoft and co-founded the pet-tech startup UltraPawz, which was acquired by Supertails.

What is Isha Juneja’s net worth?

There is no reliable public figure for her net worth. As a co-founder of a private, early-stage company that has raised about $3.7 million, her wealth is held as equity in ContraVault rather than as a reported or realised sum, so any specific number quoted online should be treated as guesswork.

What does ContraVault AI do?

ContraVault AI builds software that reads long, complex tender and RFP documents and helps bid teams decide whether to go after a contract. Its models flag risks, summarise obligations, support clause negotiation and help draft proposals, mainly for firms in construction, energy, power, defence and aerospace. The company says it cuts tender-processing time by 60 to 70 percent.

How much funding has ContraVault AI raised?

ContraVault has raised roughly $3.7 million in total: a seed round of about ₹5.1 crore ($600,000) led by Titan Capital in May 2025, and a $3.1 million (about ₹29.3 crore) pre-Series A led by Chiratae Ventures with the Titan Capital Winners Fund in June 2026.

What can founders learn from ContraVault AI?

That painful, unglamorous workflows can make the best businesses, and that timing matters: ContraVault works because generative AI became good enough to read thousand-page documents just as Juneja recognised the need. Reliability over flash, and a willingness to sell into a slow, trust-heavy market, did the rest.