In January 2017, AppDynamics was days away from its IPO. The roadshow was done, the price range was set, and every employee held equity. Then Cisco offered roughly US$3.7 billion and Jyoti Bansal took it.
By any measure it was a triumph. He has since said it left him unfulfilled, because he had not finished what he started.
IIT Delhi to a 2008 startup
Bansal was born in India in 1977 and studied computer science at IIT Delhi. Before building software for the world’s largest enterprises, he worked selling irrigation machinery — a detail worth keeping, because the instinct to understand what a customer actually does with a product shows up throughout his career.
He founded AppDynamics in 2008, as the financial crisis made starting anything look reckless. The product monitored complex applications so companies could catch failures before customers noticed.
Enterprise software of that era was sold through long sales cycles and lengthy pilots. Bansal’s team shipped AppDynamics Lite — a free version you could download, install and get useful data from in minutes. Let engineers try it before any executive was involved. That distribution decision did as much as the technology.
He led the company as CEO until 2015, by which point it had over 900 employees.
Starting again, deliberately
Rather than retire into angel investing, Bansal built a startup studio, BIG Labs, which produced Harness in 2017 and Traceable in 2020. He also co-founded Unusual Ventures, a firm now managing around US$1 billion.
Harness automates everything that happens after code is written: testing, deployment, security and compliance. Traceable secured APIs. For years he ran both as CEO, because the two products sold to different buyers inside the same customer.
In 2025 he merged them under the Harness name. In December that year Harness raised a Series E at a US$5.5 billion post-money valuation. He has been explicit that this time he wants a stock ticker, not an acquisition.
The bet he is making now
Bansal’s current argument is straightforward: if AI writes an increasing share of code, writing code stops being the bottleneck. Everything downstream — testing it, securing it, deploying it, verifying it works and rolling it back when it does not — becomes the constraint.
Harness is positioned entirely on that thesis, and in 2026 announced AI agents aimed at responding to vulnerabilities at machine speed.
What repeat founders know
Let the user try it without a meeting. The free, self-serve version was a strategic weapon in a market built on gatekeeping.
Experience can mislead you. Bansal has said his AppDynamics background tempted him to install mature sales machinery at Harness too early. Knowing how the last company looked at scale does not tell you what this one needs now.
A good outcome is not the same as the right one. $3.7 billion was an excellent result and still left him with unfinished business. Only he could know that.
Most founders spend a career trying to get one exit like his. He got it, concluded it was the wrong ending, and went back to the workbench.
