AI agents are now over 30% of CapRelay user traffic

Agentic traffic to CapRelay is now over 30% of all usage. Nearly all of CapRelay's traffic was human at the start of the year.

What changed? Professional investors started giving Claude Code, Codex, and other agents direct access to CapRelay's data and tools. These agents then use CapRelay to drive research processes.

Why? Agents are incredible at crunching data and using tools, but they need buy-side grade data and tools to deliver analyses that professional investors can use. This is where CapRelay comes in.

The problem: Agents don't have the right tools or data

Would you be interested in running analyses like this?

go through each software company above 2b of market cap and summarize the specific AI bear case for it

You probably already know AI tools are going to struggle with this task. The AI can't list all 160+ software companies with $2B+ of market cap, can't access the latest data consistently, and often relies on web sources that professional investors would never use.

Claude Code, running the latest Opus model, knows its limitations when asked to do this:

The solution: CapRelay + Claude Code or Codex

This is why many of CapRelay's customers work with us specifically to give their agents the right tools and data.

Using CapRelay's data and tools, the agent delivers a full list of 160+ software companies, with a specific AI bear case grounded in buy-side thinking.

A small excerpt from one generated run:

• ADBE - Adobe's stock business decline is a preview of a broader AI substitution risk... This is a concrete, real-time example of AI eroding an Adobe revenue stream - and stock was approximately a $450M book of business...
• SHOP - Agentic commerce is a channel where Shopify has no guaranteed privileged position, and significant adversaries could route around it. Specific concerns include: Amazon has its own "Buy with Prime" capability, deep consumer checkout trust, and stronger direct relationships with the leading AI platforms. If Amazon strikes preferential deals with OpenAI or Google for agentic checkout placement, Shopify merchants could be at a disadvantage...
• MDB - Postgres is winning the default database decision for the majority of new application builds, particularly in the AI-native startup community in Silicon Valley - precisely where the next generation of high-growth workloads is being built. MongoDB's own management concedes it "neglected" the Bay Area and had to launch a "reclaim the Bay" campaign to rebuild mindshare...

You can also view a longer excerpt of generated examples.

And this goes on for all 160+ software companies.

How other investors drive research velocity with agents

This is one of many use cases for agents + CapRelay. Some other use cases and prompts that agents can execute with CapRelay's data: - Large-scale iteration or idea generation: "which commodity industries have added a lot of supply in the past 1-2 years?" - Industry trends or read-throughs: "what are recent trends in non-resi earnings that might impact AECOM?" - Data analysis: "create a summary table laying out takeover defenses for all software companies below 1b of market cap."

This allows the investment team to allocate more time to the creative research tasks that actually drive alpha, which AI can't do.

How CapRelay supports agents

What makes this possible is that CapRelay is engineered for both agents and humans. - CapRelay is designed to be navigable via a simple skill file, so agents can imitate the way human analysts work. - CapRelay's content is instantly available and machine-readable. - And most importantly, CapRelay provides buy-side quality research, so agents can think like professional investors.

If you're interested in driving research velocity with tools like Claude Code or CapRelay, sign up for our free trial, or contact us at support@caprelay.com. We'd love to chat.