AI agent security startup Reco has raised $55 million, with AT&T Ventures joining as both an investor and a customer. The new money brings Reco’s total funding to $140 million. The company helps security teams see which AI agents are running at work, what they can access, and what they are allowed to do.
Inside the Reco Funding Round
Besides AT&T Ventures, the round includes Forestay and Quadrille Capital. Reco says it will use the cash to grow sales, partnerships, sales channels, and customer support. The pitch is simple: large companies are moving AI agents from small trials into daily operations, and they need a way to keep them in check.
| Detail | What we know |
|---|---|
| New funding | $55 million |
| Total raised | $140 million |
| Investors named | AT&T Ventures, Forestay, Quadrille Capital |
| Apps it connects with | More than 280 |
| Use of funds | Sales, partnerships, channels, customer support |
Why AT&T Is Both Investor and Customer
AT&T already uses Reco inside its own business. The telecom giant says the software gives it better visibility into agent risk, access management, and third-party integrations across its enterprise apps. AT&T’s security chief, Rich Baich, said it also helps the company stay ready for audits.
Why AI Agent Security Is Becoming Urgent
AI agents now live inside chat tools, browsers, copilots, and automated workflows. They often connect through OAuth grants, API keys, and service accounts, which makes them easy to add and hard to track. Reco calls its focus “agentic security,” meaning knowing what agents exist, what identities they use, and what actions they can trigger.
Reco points to a Gartner forecast that the average global Fortune 500 company will run more than 150,000 agents by 2028. That compares with fewer than 15 in 2025, according to the company. It also cites Grand View Research figures putting the agentic AI security market at $1.3 billion in 2025, growing to $17.8 billion by 2033.
How the Reco Graph Works
At the core of Reco is a map it calls the Reco Graph. It links agents to the identities, apps, permissions, data, and workflows they touch. A single agent might look harmless on its own but still hold connections that lead to customer records, source code, or finance systems.
With that map, security teams can:
- Spot agents with more access than they need
- Revoke stale permissions and outdated access
- Turn off risky integrations
- Apply policy controls without breaking legitimate workflows
Reco connects with platforms including OpenAI, Anthropic, Microsoft Copilot, Salesforce, ServiceNow, and Workday. CEO and co-founder Ofer Klein says agents are gaining access to business data faster than many companies can map it.
A Crowded Field for AI Agent Security
Reco is not alone in chasing this problem. Investors have backed several startups this month that promise to watch over workplace AI, including Pluto Security and Harmonic Security. Model makers are adding their own tools too, such as Claude’s new agent tracing support for monitoring platforms.
The common thread is that governance and oversight are becoming a standard line item in enterprise AI budgets. Companies no longer want to choose between blocking AI and accepting blind spots.
What This Means for You
If you work in IT or security, this funding round is a sign that agent sprawl is now a board-level concern. Even smaller teams can borrow the same approach:
- List every AI tool and agent that connects to your work apps, including browser add-ons and copilots.
- Review which accounts and permissions each one uses, and cut anything it does not need.
- Remove old OAuth grants and API keys from tools nobody uses anymore.
- Set a simple approval step before new agents can connect to company data.
If you are just starting, our guide to AI threat detection covers the basics, and you can browse more cybersecurity AI tools for ideas.
Sources
Details in this article come from funding coverage by SecurityBrief, including statements from Reco and AT&T.
Conclusion
Reco’s $55 million round shows how quickly AI agent security has turned into its own market. As companies hand agents real access to data and systems, knowing what each one can touch is becoming as important as the agents themselves. AT&T’s dual role as customer and investor adds weight to the bet. Expect more money to flow into tools that keep workplace AI visible and under control. For more funding news, check our AI news page.
