Generative London: How to Win AI’s App Layer, with Granola

While the conversation around generative AI often centers around foundational titans—OpenAI, DeepMind, Anthropic—the most intriguing innovations right now may be happening one layer up: the applications. Few products capture that momentum better than Granola, an intelligent notepad for meetings that uses AI to support human thinking, rather than replace it.

Lightspeed led the company’s $4.25 million seed round in May 2023, and has continued to support Granola in their $20 million series A and Series B financings. Co-Founders Chris Pedregal and Sam Stephenson recently welcomed Lightspeed to their Shoreditch office for a Generative London meetup.

What follows distills that conversation into a set of takeaways for anyone interested in building at the AI app layer, edited for length and clarity.

From One-Size-Fits-All to Purpose-Built Precision

Pedregal and Stephenson initially considered building their own language model, but the cost was prohibitive, and frontier models were improving by the day.

“From the very beginning, our view was that low-frequency, non-critical use cases are going to be eaten up by ChatGPT or Claude,” said Pedregal. “If you want to do something that really matters, bespoke tools optimized for that purpose are going to be way better.”

So they began looking for ways to apply GenAI to the things people struggle with each day. Their focus shifted to a universal pain point: back-to-back meetings and the avalanche of work that invariably accompanies them—summaries, follow-up emails, lists of action items, and other related tasks.

Out of that came Granola, which relies on third-party language models to automatically transcribe meeting notes, then helps you organize and personalize them in a way that makes sense to you.

“We think the tools you use matter,” Pedregal added. “And if that tool makes you 10% or 30% or 50% better at your job, that’s always going to have a lot of economic value.”

Balancing Functionality and Simplicity

During a six‑month beta with 150 users, the founders introduced features rapidly, discarded most, and kept only what felt effortless to people juggling nonstop calls.

“The hard thing was figuring out how to let users get the stuff they care about in a way that feels natural and effortless,” Stephenson explained. “We really can’t put too many buttons in front of you when you’re in that situation, you don’t have the brain space for it.”

The problems you choose not to solve are as important as the ones you do

Early on, real‑time transcription accuracy lagged. Rather than investing time and effort into building a proprietary transcriber, Granola doubled down on better note summarization and waited for transcription technology to mature. The same restraint applies today to multilingual support, retrieval augmented generation (RAG), and ultra-long meeting capture.

“We have to be quite selective about the challenges we bite off and the ones we choose to leave alone,” says Stephenson. “A lot of the game for us has been picking our battles, knowing what to innovate on, and where we needed to wait for the technology to get better.”

Toward a “Jetpack for the Mind”

While a tool that makes it easy to record and understand the content of meetings is certainly finding its audience, Granola’s founders have much bigger ambitions. If, as Steve Jobs famously said, the personal computer is a bicycle for the mind, Granola’s founders see AI as the jetpack.

“Our ambition is to build tools that help people think smarter, work better, and do great things,” he added. “What new heights can humanity reach?”

Yet that vision comes with guardrails: Stephenson and Pedregal are determined to hold to the minimalist ethos that has contributed to Granola’s early success.

“What people love about Granola is that it’s simple and gets out of your way,” said Pedregal. “If you add 50 buttons in there with new features, you kind of kill the golden goose. We’re figuring out how to find that balance—to move quickly, but also keep the soul of the product intact.”

To hear the full conversation with Chris Pedregal and Sam Stephenson, watch to the latest Generative Now episode here.

The content here should not be viewed as investment advice, nor does it constitute an offer to sell, or a solicitation of an offer to buy, any securities. The views expressed here are those of the individual Lightspeed Management Company, L.L.C. (“Lightspeed”) personnel and are not the views of Lightspeed or its affiliates.