He Scaled a Company to a Billion Users. Then He Started Over | Deepak Gupta, GrackerAI

What we talked about.

In this episode of the Sunny Ray Show, host Sunny Ray talks with Deepak Gupta, who helped scale LoginRadius past a billion end users and then started over with GrackerAI, a generative engine optimization platform for B2B companies, with a focus on cyber security vendors. Gupta explains why he left a decade of identity and security work after seeing search shift toward AI answers in 2024, and how a half-correct but well-structured ChatGPT comparison convinced him that AI search is the future. He covers what scaling a security product taught him about trust and infrastructure, why generative engine optimization is a superset of SEO rather than a replacement, and what breaks for companies that ignore AI answer engines. He also describes the blind spots of B2B SaaS teams, including optimizing for too few engines and too few prompts, and predicts that Google will move fully to AI mode and that agents will soon handle software buying.

After scaling LoginRadius to a billion users, Deepak Gupta bet on AI search and started GrackerAI to help brands get found in it.

The questions, and the answers.

What is GrackerAI and what problem are you solving?

We are a generative engine optimization platform for B2B companies, specifically serving cyber security. In a nutshell, we help companies get visible in AI. When someone asks ChatGPT, Claude or Perplexity for the best email tool or antivirus tool, we help companies rank well there.

What made you walk away from LoginRadius and start over?

LoginRadius began around user engagement, privacy and security. After ten years, AI arrived in 2022 and 2023, and in 2024 I saw search shifting as people looked for synthesized answers, not just Google links. I saw a bigger opportunity, and if agents are the future of search, the work has to start now.

What was the moment you realized search itself was about to change?

When I researched identity solutions, Google gave me pages of links. AI gave me an answer that was only half correct, but it presented comparisons, pros and cons, use cases, and asked what kind of company I was. I did not want to spend hours on research, so that structure became my turning point.

What is the biggest misconception about generative engine optimization?

Many people say it is fake or just SEO. Generative engine optimization is a superset of SEO. You still need good content and an optimized website, but a middle layer appears when AI engines ingest and answer from that content. Many companies miss that answer engine layer on top of legacy SEO.

What breaks first for a company that ignores AI answer engines?

Potential customers cannot find them. When people search for the problem the company solves, it will not show up, so it will not generate good leads or customers. AI mode is becoming core to Google, and agents will soon search for answers too, so ignoring this hurts future customer acquisition.

Where are most B2B SaaS companies still blind to how AI search ranks them?

They optimize for too few engines, often only ChatGPT or Perplexity, and skip Claude or Gemini. Many security companies ignore Microsoft Copilot, which is a top source for them. They also test only a few hundred prompts, when they need a wider horizon to see and fill their gaps.

How do you know what prompts real buyers might ask?

We do not know exactly what buyers will ask, so we work with third party partners and our algorithms to generate hundreds and thousands of prompts. We fire them at AI engines, and the more data we collect, the better we can see which gaps a brand is missing and how to fill them.

What do you believe about the future of search that most founders disagree with?

Google will shift from its search pages to full AI mode sooner than people expect. Soon agents will do the shopping for us, including in B2B, comparing software, getting pricing, negotiating, and taking a purchase nearly to the contract stage. Agentic commerce will be a key part of search.