Kindo raises 30 million to better oversee AI agents
Kindo announces a $30 million fundraising round for its enterprise AI agent management platform. The challenge is simple: controlling their access, costs, and outcomes, as adoption progresses faster than governance.
American companies are already giving names, roles, and tasks to their artificial intelligence agents, but oversight is not keeping pace. Kindo, a company based in Venice Beach, California, has raised $30 million for its AI agent management platform, while a separate survey shows that 63% of decision-makers surveyed name their agents and 60% assign them a title, compared to only 41% who assign a human supervisor. This issue matters to technical, financial, and compliance departments, as an agent acting within a company's systems can also leave costly traces—or cause damage.
Quand les agents ont un nom, mais pas toujours un superviseur
The tension between rapid adoption and insufficient control is strongly highlighted by both sources. According to the Digital Workforce Readiness Survey by Vida Global Inc., 72% of respondents say quality and reliability of results are their main concern, but only 12% actively monitor accuracy or error rates. At the same time, over 80% report testing or deploying digital workers in their internal workflows or customer-facing roles.
This rise is especially seen in customer service, where 72% of companies use or evaluate digital workers. The study, conducted in late October 2026 among 215 U.S. decision-makers in mid-sized and large companies, also gives an idea of the gap between promises and practices: 85% report a positive return on investment, but only 3% use outcome-based pricing. The rest are largely billed via traditional monthly subscriptions, which does not always help link the bill to actual utility.
La promesse de Kindo : voir, contrôler, facturer
Kindo positions itself precisely on this blind spot. The company aims to provide a common operational layer for IT, security, finance, and compliance, with a control center that aggregates activity, access, costs, and outcomes. Its approach covers both agents created on its platform and those developed elsewhere, via gateways that govern calls to approved data, tools, and models.
The core of the product relies on several components. A registry identifies who owns the agent and what permissions it inherits, while a cost allocation system links token usage and failed executions to individuals, departments, or projects. For sensitive deployments, Kindo adds an agent harness, an execution environment that allows for increased monitoring, assistance, and intervention, including taking back control if something goes wrong.
Le financement sert aussi à industrialiser la discipline
The fundraising was co-led by Crosspoint Capital and Forgepoint Capital. The round brings Kindo's total capital raised to $57.6 million, and two fund executives, Andre Fuetsch and Ernie Bio, are set to join the board. The company says it recorded approximately 800% year-over-year revenue growth in 2026, without further detailing this comparison base.
The new funds will be used to strengthen the sales team and accelerate development of the AI Control Center and Deep Hat modules. These include four tools: Sift, which filters certain events after learning; Tune, which adapts a smaller model to a given task; Shrink, which converts repetitive steps into code to reduce model cost; and Attack, intended for authorized offensive security testing. The idea is less glamorous than an agent that talks to itself, but much more useful when it comes to paying the bill.
Des budgets qui montent, une gouvernance qui court derrière
Kindo says it is already in production in large groups and mid-sized companies in financial services, professional services, healthcare, and air transport. The company also cites Aireon, whose security director estimated in May 2026 that using Kindo for threat hunting and level 1 and 2 security operations generated over $2 million in annual savings. This is a client-reported result, not a guarantee for all deployments.
Vida Global Inc.'s figures confirm that a market is structuring around this governance issue. Nearly 84% of surveyed organizations plan to increase their AI agent-related spending within two years, and 70% believe digital workers could absorb up to a quarter of human workloads within three years. In this context, the battle is no longer just about creating agents, but about controlling them, their access rights, and their real cost. This is often less visible than a spectacular demonstration. But it is what determines whether a deployment lasts.
Key points
- 63% name their AI agents, according to Vida Global Inc.
- Only 41% assign them a human supervisor.
- Kindo raised $30 million in 2026.
- 84% plan to increase their AI spending within two years.
- 215 U.S. decision-makers responded to the survey.
- Aireon reports over $2 million in annual savings.
By the numbers
- $30 million — fundraising announced by Kindo in 2026.
- $57.6 million — total capital raised by Kindo after the round.
- 800% — year-over-year revenue growth, according to Kindo.
- 215 decision-makers — sample size of the U.S. Vida Global Inc. survey.
- 84% — share of organizations planning to increase AI spending.