Five questions to Nakul Aggarwal on BrowserStack's latest acquisition
BrowserStack, an Indian SaaS company that is among the leaders in providing app and browser testing products, recently acquired Bird Eats Bug, a Berlin-based bug detecting and reporting platform
(L to R) Nakul Aggarwal and Ritesh Arora, Co-Founders at BrowserStack.
Nakul Aggarwal, co-founder and CTO of BrowserStack, spoke about the acquisition and how the company is evolving from infrastructure to platform provider, in a recent interview with Forbes India. Edited excerpts:
Q. Tell us about the rationale for this acquisition
We’re transforming from a test infrastructure provider to a cohesive, ubiquitous testing platform company, where the testing landscape is quite fragmented. Currently, the process from QA request to production is disjointed, so we're working on integrating it into a seamless end-to-end platform. This transformation involves expanding our portfolio and addressing various challenges.
Bug filing, though seemingly minor, has been a significant friction point. The new bug tool we’re adopting aligns with our philosophy of a developer-first, frictionless experience, making it much easier to file bugs without getting bogged down in details. This tool complements our product philosophy perfectly.
Related stories
The bug filing process, in general, is quite a painful journey today. We’re working on enhancing our test platform, where we have a manual testing product, called Live. Testing happens in different environments, like Chrome or mobile devices, and filing bugs can be a painful process.
The new tool simplifies this significantly, especially for Live users who rely heavily on manual testing. It not only benefits Live users but also those testing in their own browsers. The biggest advantage is for Live manual testing users, streamlining their workflow and improving the overall testing experience.
Q. You are well-known for browser testing. Are you expanding into adjacent opportunities?
Our focus has shifted from test infrastructure to a broader testing platform. Back in 2021, mobile testing was a significant part of our business, with a lot of emphasis on iPhone and Android testing alongside Chrome and Safari. Over the last two to three years, we’ve expanded our offerings considerably.
We launched four products in accessibility testing, acquired Percy for visual testing, and built a no-code testing platform. We also developed an observability tool for testing. Our goal is to broaden our scope beyond traditional testing infrastructure.
Q. How are you taking advantage of generative AI?
From my perspective, Copilot technology is about making testing faster and easier. Our products now incorporate AI and Copilot tech extensively. For instance, it can generate test cases from JIRA stories, automatically create code from these test cases, and even heal test scripts if site changes break them.
These features are already available on the BrowserStack platform, and we're continually enhancing them with Copilot and LLM technology. This has been a significant boost for developers worldwide. We're also tackling the challenge of testing AI technology, which is complex due to its non-deterministic nature. Unlike traditional applications with predictable outcomes, AI can produce varying results, and finding a solution for this is a key focus for us.
At this moment, we use existing LLM models such as GPT, Llama and Gemini. We are evaluating if we should build our own LLM model as well, because we have a lot of user data over the years and we can train our own model which can add a lot more value to our customers.
Q. To what extent is AI automating your work?
I don’t think we've made a significant impact on our internal productivity with AI-led automation yet. For example, with Copilot, while we're using it for development teams across BrowserStack, we're now exploring tools to help our support teams. We’re assessing whether Copilot-like tools can help resolve L1 tickets and assist with documentation.
For engineers, Copilot has received mixed reviews—some teams find it useful, while others do not. We’re still evaluating its effectiveness and expect to make a decision in a couple of months. We need to determine where it works best and apply it accordingly, but we haven’t yet found a major breakthrough that makes a noticeable difference in productivity.
