EntrepreneursOfLA icon
Original Startups

Kindo Raises $30M to Rein In AI Agents: Lessons for LA Startups

LA's Kindo raised $30M to help companies control what AI agents access and spend. What founders building or using agents should take from the round.

Kindo Raises $30M to Rein In AI Agents: Lessons for LA Startups

Los Angeles-based Kindo has raised $30 million to help companies keep track of something many of them are deploying faster than they can manage: AI agents. The new round was co-led by Crosspoint Capital and Forgepoint Capital and brings the company's total funding to $57.6 million, according to Kindo's Nov. 12 announcement.

The raise is a signal worth paying attention to, even for LA founders who will never buy enterprise software. It shows where investors think the next problem in AI lies. The question is no longer only whether an agent can do the work. It is who is watching what the agent does, what it can reach and what it costs.

The deal

TFX Capital and CircleRock Capital joined the round, along with existing investors including Drive Capital, Riot Ventures, RRE Ventures, Eniac Ventures, Marlinspike Partners, New Era Ventures and Sunset Venture. Andre Fuetsch of Crosspoint and Ernie Bio of Forgepoint will join Kindo's board.

Kindo says its revenue grew about 800% year over year in 2026. That figure comes from the company and has not been independently verified. The company plans to use the money to expand its go-to-market team and keep building its product.

What Kindo sells

Kindo describes itself as an AI agent management platform. In plain terms, it gives a company's IT, security, finance and compliance teams one place to see the AI agents running across the business, set rules for them and step in when something goes wrong.

According to the company, the product starts by discovering "shadow AI," the agents and tools employees are using without official approval. It then tracks what each agent accesses, what it does and what it costs. Kindo says its gateways audit every call an agent makes to connected data, tools and models, and that it applies a company's existing permissions to both people and agents. For critical agents, it offers around-the-clock monitoring and the ability for a person to take over.

Cost is a big part of the pitch. Kindo says its control center records every token used and every agent run, including failed ones, so spending can be rolled up by person, agent, department or project. Finance teams can set budgets, alerts and hard spending caps. A set of add-ons the company calls Deep Hat aims to cut costs by filtering out routine events, using smaller models for simple tasks and turning repeatable steps into regular code.

The company says it is used by large and mid-size companies in financial services, professional services, healthcare and aviation, and that it can run on a customer's own servers, in its cloud or in an air-gapped environment.

Why this matters for LA founders

Founder and CEO Ron Williams put the core problem this way in the announcement: it is easy to deploy powerful agents, but the tooling to centrally control how they are built, what data they can reach and what they cost is missing. That gap, he argued, is why the teams responsible for AI are slowing down adoption.

That is a useful lens for any startup, whether you are building agents or just using them. A few takeaways:

If you sell AI agents to businesses, expect tougher questions. Buyers increasingly want to know what data your agent can touch, how its actions are logged and how a customer can shut it off. Having clear answers, and audit logs to back them up, can shorten a sales cycle. Being able to show predictable costs may matter just as much.

If your team uses AI agents internally, write down who approved what. Even a small company can end up with agents connected to email, calendars, a CRM and a payment system. Keep a simple list of every AI tool with access to company data, what permissions it has and who owns it.

Treat AI spend like any other budget line. Usage-based pricing can surprise you. Set spending alerts with each provider, review usage monthly and check whether a smaller, cheaper model would do the job for routine tasks.

Give every agent the least access it needs. An agent that only drafts replies does not need permission to send payments. Narrow permissions limit the damage if something goes wrong.

Plan for a human override. Decide in advance who can pause an agent and how. For anything that touches customers or money, make sure a person can step in quickly.

The bigger picture

Kindo's backers are leaning into the idea that AI governance is becoming its own category, much like cybersecurity tools did as companies moved to the cloud. Crosspoint's Fuetsch said in the release that companies are putting agents into real business processes but most still lack a reliable way to see what they are doing.

For LA's startup scene, the round is another sign that the money is not only going to companies that build AI, but also to those that help others manage it safely and affordably. Founders who build those controls into their own products, or their own operations, will be better positioned as customers and regulators start asking harder questions.

← More Startups