“By the end of calendar year 2026, I expect that AI products will contribute at least 15% to 20% to revenue,” Whatfix co-founder and chief executive officer Khadim Batti told Mint in an interview.
The company’s core business is improving digital adoption of enterprise software through nudges, i.e., tooltips on different parts of a workflow. The intention is to make it easier and more comfortable for users to ‘adopt’ such software, thereby improving productivity and efficiency.
The company has increasingly been focusing on AI as it looks to become profitable within the next two years, the same timeline it has set for itself to make an initial public offering (IPO). “Currently, we’re at -27% Ebitda in 2025. In 2026, alongside growth, we will be at -5% or -6%,” Batti said.
Also read | Whatfix secures funding from Warburg Pincus, SoftBank at $900 million valuation; eyes IPO in two years
But Whatfix’s tryst with AI isn’t new. In fact, it goes as far back as 2019, when it acquired Airim, a startup that provides AI-powered personalisation for users and customers. “Combined with Airim’s AI-powered personalised engine, Whatfix users can now expect hyper-personalised in-app guidance at their point of need,” the company had said in a statement back then.
Whatfix last raised funds in September in a $125 million Series E round led by private equity firm Warburg Pincus, with SoftBank Vision Fund 2 also participating. The fresh funds bumped the company’s valuation to $900 million.
Backers of the company include Dragoneer, Peak XV Partners, Eight Roads and Cisco Investments.
Over the past nine months, several startups and companies have been pivoting their AI offerings to become agentic in nature. This technology has emerged as easily monetizable and promises to boost productivity, especially in the enterprise. As a result, investors, too, have been piling on, Mint previously reported.
Betting on agents
The only thing that’s missing with the agentic push across industries is adoption.
For context, adoption of agentic AI systems stood at 1% in 2024, according to research firm Gartner. That number is supposed to shoot up to 33% by 2028. The most common complaint from the enterprise has been return on investment.
Batti believes that its core product, digital adoption platform (DAP), which accounts for 93% of Whatfix’s revenue, will be the game-changer. “One of the reasons that organizations complain about return on investment on agentic AI systems is because people are not utilising it on a regular basis. Agents are sitting somewhere in the system, and users are trained on one or two use cases. But they don’t know that agents can do several other things.”
Also read | You’ve heard of LLMs. Prosus is building its own LCM for commercial smarts.
The remaining 7% is accounted for by other product verticals, analytics, and Mirror, an application simulation software builder. “Now we’ve started a journey where we want to embed AI into each product in our portfolio,” he added.
Whatfix is working on its own AI agent that can be integrated with existing enterprise software. Agentic AI companies usually develop agents for specific use cases or create a marketplace of agents suited to different tasks.
Instead, Whatfix has chosen to use a one-size-fits-all agent that can, in Batti’s words, “perform a task on any software. If our agent is not the default, and something else is, it’ll work with those.” The rollout for this is planned over the next two to three months.
AI focus
As part of Whatfix’s doubling down on AI, the company created a separate go-to-market division, whose sole job is to find different use cases and find a market for them. Currently, the team has pulled up two use cases, both of which have early adopters already. The first one is an assistant for task automation, while the other is for administrators to “simplify their job and make them more productive,” according to Batti.
From September this year, Whatfix plans to have one up and running in beta. By November, it plans to be monetized.
But as the company shifts to becoming AI-first, it’s but a given that compute costs will rise. But Batti disagrees, saying that with their implementations, there isn’t an expectation that people are constantly engaging their platform, but more realistically, between 3-5% of the time. “I haven’t seen a negative hit in terms of our gross margins.”
As part of its AI push, Whatfix created ScreenSense, a proactive AI-based technology that fits on top of its existing verticals. It crunches several clicks or parts of a workflow into fewer steps, therefore making users more productive.
Earlier, Whatfix’s DAP would simply guide users from one end of a workflow to another in a step-by-step manner to ensure that users became familiar with using an enterprise application. Simply put, it was reactive.
Also read | Cashfree Payments eyes profitability as it focuses on $250-300 million monthly in cross-border payments
For example, when a user begins creating a purchase order, ScreenSense is able to detect the intention behind it. Instead of having a user go through the entire workflow to create the order, the technology simply automates the entire process, reducing clicks and time taken.
“In the next 6-12 months, we want to continue to enhance ScreenSense so that it can not only work with our products but also can start working with AI agents or co-pilots, which are built by third parties,” said Batti. The hope is that with this, it will not only increase digital adoption of new technologies being introduced by Whatfix’s clients into their ecosystems but also boost user productivity.
No product left behind
Even as Whatfix reworks its own ecosystem to become more AI-heavy, the company says its core offerings won’t be phased out but will evolve alongside the new innovation it is planning.
On some level, the rationale makes sense. Companies don’t change their software stacks overnight. Solutions for customer relationship management and enterprise resource planning are often deployed on such a large scale that overhauling them overnight could be challenging, if not outright disastrous.
“We will definitely continue to increase our investment to become AI-first. But our existing product lines, I would say it will continue in your future. It’s not going to discontinue. We are continuing to invest there as well,” Batti said.
But even with those investments, Whatfix’s core products may be running on borrowed time. “Our core products will fundamentally shift. But it’ll take at least five to seven years. That’s the mid-term view.”
For example, with their analytics vertical, instead of generating dashboards for their clients’ users, Whatfix now uses natural language processing. Users can opt to simply input what they’re looking for, and Whatfix generates it for them.