World of DaaS Roundtable Recap: Selling to AI Agents

The latest World of DaaS roundtable brought together executives from across the data ecosystem to explore a new and rapidly evolving frontier: selling to AI agents. While the concept remains unfamiliar to many, the growing presence of AI agents in enterprise workflows makes it a timely and worthwhile area to examine.

1. Are We Ready to Sell to AI Agents?

The discussion opened with a fundamental question: are we ready technologically, organizationally, and commercially to sell to AI agents?

Most agreed that while agent adoption is early, the momentum is real. Some companies are already seeing inbound interest from agent-driven workflows, though often in experimental or proof-of-concept stages.

For now, the opportunity is mostly forward-looking. “There’s a lot of buzz, but it’s unclear where real demand will show up first,” one participant noted. Still, the group agreed it’s worth laying groundwork now, as agent adoption will likely accelerate.

2. The Integration Gap

Several participants highlighted the stark difference between selling data to developers building AI tools vs. directly supporting the end users or agents themselves.

"Startups integrating our data often don't know what their end users want. It becomes high-friction, low-return work," one founder explained. In contrast, direct relationships with enterprise users who deploy internal AI agents result in larger, more strategic contracts. These buyers care deeply about accuracy and reliability, and they’re willing to pay for it.

As one participant put it, “Enterprises don’t want 80% accuracy. They want 100%. Anything less is worthless.”

3. Microtransactions, Discovery, and Agent-Led Procurement

The conversation shifted toward what it will take to enable seamless AI-agent-led procurement. Will agents be allowed to transact autonomously? Will they buy data in small units, line by line, based on queries? Are sellers equipped to monetize those interactions?

Some participants expressed concern over a race to the bottom. “If agents are trained to find the cheapest data, it becomes a pricing war with no room for differentiation.” Others emphasized the importance of structuring pricing around value, not just access.

Another layer to this: discoverability. As agents start replacing traditional search, metadata, documentation, and semantic models will determine which datasets are surfaced and selected. “We need data that can describe itself,” one executive noted. “Free text descriptions aren't enough. You need context packages - sample queries, notebooks, and governance metadata to enable discovery.”

4. Rethinking the Data Stack for Agents

One recurring theme: the existing data infrastructure isn’t built for agentic workflows. Traditional APIs, licensing models, and compliance protocols assume human buyers. Selling to agents requires a new framework.

Several companies discussed building internal translation layers between buy-side and sell-side agents to manage permissions, pricing logic, and data matching. Others emphasized the need for shared context, especially in B2B use cases where agentic buying requires persistent knowledge of customer ICPs, personas, and preferences.

“Agents won’t just transact; they’ll need to understand who they’re buying for,” one participant explained. “And that context needs to persist across interactions.”

7. Lessons Learned

While most companies are still early in their agent strategy, some shared firsthand lessons:

• Don’t rely on intermediaries. Selling to agent startups rarely works unless you're also deeply embedded with the end customer.

• Charge for infrastructure, not API calls. “The price is in the data collection. Whether you make one call or a million, that’s not the variable cost.”

• Build with MCP (Model Context Protocol) in mind. It's emerging as the standard way for agents to discover and interact with external tools, but it still requires robust instruction and support.

8. Key Takeaways

• Agent demand is coming. Even if real-world adoption is still early.

• Enterprise buyers will demand accuracy from agents, not just affordability.

• Discoverability will depend on semantic context, not just keywords.

• Pricing models must evolve to support microtransactions and recurring agent queries.

• Shared context and translation layers will be critical to agent-to-agent transactions.

If you are a DaaS executive interested in participating in future roundtables, apply to join our World of DaaS community.