Having sold AppDynamics to Cisco for a record $3.7 billion just ahead of its IPO in 2017, serial entrepreneur Jyoti Bansal has been back at building solutions for enterprises.
His latest venture Harness.io was valued at $5.5 billion in a Series E round of funding led by Goldman Sachs announced earlier in December 2025. As enterprises struggle with accelerated code production due to AI deployment and deployment, Harness’s tools help them with all the work that happens after coding, including testing, security, deployment, optimisation, compliance and verification, among other things.
“The need for everything from testing, deployment, verification, and security is much higher with nearly three to five times more code coming out due to AI. We are building an AI fabric for security, deployment, testing and compliance,” Bansal tells Forbes India.
From his learnings in building AppDynamics, Bansal has created Harness based on a ‘startup-in-a-startup' model which includes 16 different product modules which work independently within the company with a CEO and fair degree of autonomy. This helps the company come up with new product modules faster. Earlier this year, Harness merged with API security platform Traceable to form a single platform in the DevSecOps space. Interestingly, both Harness and Traceable were founded by Jyoti Bansal as part of his venture BIG Labs.
Among the 16 product modules, four come from Traceable’s portfolio to leverage the platform effect.
“We give them (product module CEOs) autonomy to build the best product for their task, and to compete in the market with competitors that are trying to build along those lines, to be responsible for revenue for that, be responsible for customer adoption success, so that the teams can move fast,” adds Bansal. The company plans to expand its repertoire to 25-30 modules over the next few years.
Harness recently announced expansion in India operations at its Research and Development Centre in Bengaluru. The company grew to 480 employees in 2025, with plans to expand to 1000 employees in the coming years.
In an interview with Forbes India, Bansal talks about Harness’ plans of expansion into Asia and India as new markets, and upcoming product lines. Edited excerpts:
Q. What problem does Harness help enterprises with? Are there companies in India that use the platform?
What Harness focuses on is everything that happens after coding. About the first 30 percent of software engineering is writing code and then there’s 70 percent of the work that happens after coding, which includes testing, security, deployment, optimisation, compliance, verification, etc to make sure that software works. Even a small glitch can bring everything down. So that’s where Harness’ focus is.
Now that there’s more code coming from AI, the need for everything from testing and deployment and verification and security is higher. But we are also leveraging a lot of AI to drive all of that.
We work with nearly 1000 enterprise companies and in the US, nearly 8 out of 10 banks are our customers. In India we work with the likes of HotStar and InMobi. We plan on focusing on expanding in the Asia market after focusing on the US as a primary market and our expansion in Europe last year.
Q. Why did you decide to merge Harness and Traceable—both companies founded by you—this year? What benefits does it bring?
I started Harness to automate the DevOps tasks, and Traceable was a separate company to secure all the code and started with the APIs. We realised over the last few years that people are thinking of both—DevOps and application security—in an integrated manner. So that’s why we merged the two companies earlier this year, and that has worked out very well for us. We are seeing a lot of growth in both parts of the business.
Q. What are the different products the combined entity has?
They both come together in solving one problem, which is streamlining code written by the companies before it gets into production. This includes seven different kinds of testing, different security tasks, deployment tasks, cost control etc. So we have 16 product modules and each of these ‘startups’ focuses on carrying out one task.
The reason we could run it as a startup is because there is a product leader whose job is like that of a CEO of a startup and the primary purpose of this setup is autonomy. With greater autonomy, the teams can move faster due to greater flexibility. It also creates internal accountability.
It is a model I have been perfecting for some time, and it is working well for us. Our customers can pick and choose what capabilities they want, and we don’t force them to buy all capabilities from us. There is no bundling of these 16 modules and our customers can pick and choose—some start with two or five, and we earn our business by proving that each module is the best in class. That said, we get a lot of platform effect.
Q. How do you evaluate which module is working for you and which isn’t? What are the capabilities you start with for the ‘startups’ and how do they scale?
We constantly evaluate these modules every quarter, similar to a business review at a startup. Sometimes things don’t work out in the start, but we repurpose the technology or the people in the team and it also allows us to run experiments for new problems we want to solve.
We typically start with five to seven people on a problem statement, and they can leverage our platform. Once they start getting more revenues and a product market fit, we fund the teams. Once the team reaches a revenue of 1 million, we fund them to a Series A stage, and so on. Internally we have ‘startups’ at all these levels—ranging from 50 million plus revenue to those in the seed stage.
But we also have internal timelines—like if you don’t get there in six months you have to change course or pivot and try a different approach to solving the problem.
Q. And how are you leveraging AI for efficiency and outcomes?
Like most companies, we are figuring out where we can use AI. On the software engineering side, we definitely use AI for coding in a significant way. We use our own tool chain for testing, security and deployment, and we also use AI for other tasks such as customer support, marketing, and sales.
The greatest thing about AI is the notion of a knowledge graph capturing institutional knowledge for an organisation. Generic AI is not very useful without understanding the institutional knowledge of a business. With AI agents we are using the context specific to a particular business, and a lot of our secret sauce and IP is built around that.
Q. Will you look at a public market listing for Harness in the coming years?
We can build a much bigger platform for the next 10 years and continue to grow the company. So that is the number one goal. How do we do it now—as a private company, a public company or as part of some other company?
I think we are best served to be an independent company and keep doing it for a long time. Going public is a natural part of a company's progression, especially if you are an independent company, and I do think we will go IPO.
Harness is bigger in revenue with $250 million ARR target in 2025, than my previous company AppDynamics was ($150 million ARR) when we were going IPO, but the IPO markets are different now and there is a higher bar to it. IPO is not an exit strategy, it is the beginning of the next chapter.
First Published: Dec 24, 2025, 19:57
Subscribe Now