Eric Glyman on using AI to radically boost internal productivity

FirstMark is an early-stage venture capital firm (backer of Pinterest, Shopify, Airbnb, Dataiku, Synthesia, and more), dedicated to helping their founders stay on the bleeding edge of AI, machine learning, and data technology. And at Ramp, we operate under the same principles—which is why it was a no-brainer for our CEO and co-founder Eric Glyman to join their recent Data Driven event to discuss how startup leaders can use AI, not only in their products, but also internal operations.

Check out the full conversation on YouTube, Spotify, or Apple Podcasts. Read on for a quick overview of Eric's top recommendations.

1. Create an applied AI team focused on driving innovation

Tools like ChatGPT are great starting points for AI, but as more companies are realizing, AI’s true power lies in its ability to train on internal datasets and access processes. To efficiently identify these opportunities, you need employees with expert knowledge of AI models and their potential. Creating an applied AI team to explore use cases across the organization could help you get faster value from the technology.

At Ramp, our team of AI specialists asks one simple question: where can AI be applied at Ramp to make things better? To date, they’ve embedded with dozens of teams, from product to sales, marketing, and customer support. “I don't think there's a function at Ramp that doesn't use AI in some way,” says Eric, “The team looks for places where we can use data signals to apply automation to manual tasks.”

Take Ramp’s underwriting process as an example. Previously, businesses applying for Ramp’s corporate card had to wait two days to hear back from our underwriting team. After a two-week sprint during which the AI and underwriting teams sat side by side to apply AI to the process, the time dropped to as few as four hours per application.

The partnership didn’t just save everyone time, it allowed the underwriting team to scale their expertise and focus on less repetitive, higher-value work.

2. Use AI to accelerate—not replace—people

Popular AI discourse likes to focus on when and how AI will automate entire jobs—but in reality, AI often generates better results when business leaders use it to empower their team rather than replace it.

“Sometimes people tend to jump to full automation when what’s really valuable in the short-run is augmenting what people are already doing,” says Eric.

One area where we’ve seen this to be true is sales development. With a little help from AI, our sales development reps (SDR) are able to book up to 4x more meetings than the industry average.

“Our reps regularly work on a couple sets of activities,” Eric explains, “If you focus on streamlining parts of these tasks instead of immediately buying an AI SDR, you can make employees radically more productive. It leads you to start thinking about what data they need about different companies, signals that can be detected, and how to run and measure messages to improve outcomes.”

3. Explore AI agents to scale knowledge transfer

News headlines talk about AI agents being the next frontier for the technology—but the idea of a digital agent isn’t a new concept.

Eric’s first company, Paribus, founded alongside Ramp co-founders Karim Atiyeh and Gene Lee in 2014, was essentially a limited digital agent that helped consumers recoup price adjustments for their purchases.

Now, a decade later, Ramp is testing agentic AI to analyze hundreds of thousands of customer calls and generate custom summaries for teams. Team members can ask our AI agent—affectionately named Toby—what our customers have said about any topic and get a detailed digest based on real customer conversations.

Toby at work in Ramp's Slack instance

“What’s interesting is that there’s no job that [Toby] is displacing. But actually, it is unlocking new use cases.” Eric observes. “Think about the nature of the data you have and how you can query it in interesting ways.”

The arrival of GPT-4o, which allows AI models to seamlessly digest images, video, and audio in addition to text, means it’s time to explore AI agents for advanced customer support as well. One day soon, users will be able to ask a Ramp AI agent how to do anything, and it'll show them in real time directly within their account.

Watch the entire conversation

During the fireside chat, Eric also discussed how Ramp deploys AI to help customers reduce the cost of their software, audit expenses automatically, and auto-code transactions.

“We believe that AI is a mechanism to save people time, save money, and learn how to improve their process,” Eric says. “Ramp is more than software for managing spend. We’re really a productivity and data-driven company in the guise of a friendly neighborhood [corporate] credit card.”

To see the full conversation—plus exciting demos from Nomic and Reality Defender—check out the replay here.