The Agent Economy

Generational platform shifts don’t just spark new applications, they ignite entirely new economies:

• Public Cloud unleashed the SaaS economy
• The iPhone ushered in the app economy
• Social media enabled the creator economy

Recent AI innovations mark a similar tectonic shift (opens in new tab), and we are now entering the era of the agent economy. While we’ve written in the past about predicting the hype of agents (opens in new tab), the infrastructure required (opens in new tab) to deploy agents, and how the looming agentic web (opens in new tab) will revolutionize the way people browse, we are now turning to perhaps the most exciting part of the agent economy—the agentic AI applications themselves.

We’re excited to invest in AI agents for a few reasons, namely:

1. AI agents reinvent SaaS - incumbent SaaS typically pairs a point-and-click UX, siloed data moat, and seat-based model. AI Agents run counter to all of this DNA.
2. AI agents eat into labor - labor budgets are 35x larger than software budgets. AI Agents can tap into both budgets.
3. AI agents turn service into software - markets plagued by low-margin human services can improve efficiency with AI Agents.

For these reasons, we expect most AI and copilots to soon push into the world of autopilots, or agents. We see this happening and compelling opportunities for AI agents across end markets:

1. Horizontal (AI across industries) - ie. Legal AI sold horizontally to in-house teams at enterprises (Norm (opens in new tab))
2. Vertical (AI specific to one industry) - ie. Legal AI sold specifically to law firms (DeepJudge (opens in new tab))
3. Consumer (opens in new tab) (AI as a full-stack service for the public) - ie. Legal AI sold as a consumer-facing service (DoNotPay (opens in new tab))

Here’s a look at the agent economy’s SaaS and labor reinvention opportunity:

Examples of early winners

To power this generational platform shift, tech giants are estimated to spend more than $1T in capex in the coming years. That price tag suggests (and requires) an impending application ecosystem. There are some early signs of winners, but positively for investors and founders there’s also a lot of runway to go:

• Customer Service: companies like Sierra (opens in new tab) are taking a fresh approach, and will give Zendesk a hard time 
• Sales: companies like 11x (opens in new tab) are augmenting SDRs with better lead-gen
• Marketing: companies like Jasper (opens in new tab) rocketed off copywriting use cases, and now must expand into broader Marketing AI platforms
• Recruiting: companies like Mercor (opens in new tab) are finding unique ways to apply AI to talent assessment
• Healthcare: medical scribes like Abridge (opens in new tab) are performing well in highly regulated industries
• Legal: Harvey (opens in new tab) is taking the legal industry by storm and changing the way we think about billable hours
• Contact Center: Crescendo’s (opens in new tab) approach as an AI-first outsourced contact center is paying off
• General Purpose: companies like Brevian (opens in new tab) are giving companies the tools to configure secure agents for a wider variety of tasks

We are just entering the agent economy. More winners in these categories (and others) will emerge, and AI Agents will begin to impact every industry and nearly every role.

When the SaaS economy took off, we enjoyed well over a decade of new application winners. After the iPhone was released, it took over three years for mobile-first unicorns like Uber, Snapchat, and Rovio to emerge. The creator economy did not enter the public lexicon until eight years after Facebook acquired Instagram. The exact timing for decacorn AI apps is hard to predict, but with early signs of traction, robust foundational investments, and the speed of innovation cycles, we know killer apps are already being built today. 

If you’re building for this agentic future, please reach out.

‍

Thank you to Eric Flaningam for his help with research for this piece.