ServiceNow launches the Control Tower Alliance with Anthropic, OpenAI, Okta, AWS, NVIDIA and others to govern enterprise AI

ServiceNow (NYSE: NOW) has announced the Control Tower Alliance, a new ecosystem of companies spanning AI, cloud infrastructure, security, governance, value measurement and token metering. Its goal is to help customers adopt enterprise AI securely and at scale. The press release is dated October 28, 2026. It is a corporate announcement, so the benefits it describes are claims by the company and its partners, not verified results.

The founding members are Anthropic, Amazon Web Services (AWS), Cisco, HiddenLayer, Kong, NVIDIA, Okta, OpenAI, Stripe, WitnessAI and Zenity. The alliance centers on what ServiceNow calls Super Intelligence Control Tower. According to the company, it is a centralized control plane for discovering, securing, governing, observing and measuring AI across the enterprise. The alliance sits alongside the existing ServiceNow Store and Workflow Data Network Partner Passport programs.

Amit Zavery, president, chief product officer and chief operating officer of ServiceNow, defended the approach this way: with the rise of AI agents, no single solution can see everything or protect everything. That is why, in his view, AI governance requires an ecosystem of partners innovating together on an open platform.

The alliance offers partners three ways to collaborate:

• Build together: partners and ServiceNow develop differentiated solutions that run on the Control Tower and are easy to deploy. For partners, this means building on the platform companies already use to govern AI, rather than offering a standalone tool the customer has to integrate on its own.

• Deliver with governance: pre-integrated, curated solutions go through ServiceNow qualification and certification. In addition, there are shared standards for agent identity and security guardrails, so that AI is accountable to the business.

• Sell and lead together: select partners integrate their solutions with the Control Tower to create joint offerings, which ServiceNow sells and supports globally. This way, partners do not have to make their way to enterprise buyers on their own.

For customers, the pitch is a curated set of AI solutions that have passed qualification, certification and integration testing before appearing in the catalog. The promise is faster adoption, with governance active from the start.

Comments from the founding members offer clues about the technical content:

• Cisco: will connect Cisco AI Defense with ServiceNow to provide visibility into AI assets and security findings, and ultimately drive remediation.

• HiddenLayer: contributes specialized AI security, focused on emerging agentic threats.

• Kong: will extend governance to real-time enforcement over the APIs, models and tools that agents depend on.

• NVIDIA: according to Pat Lee, NVIDIA OpenShell enforces runtime boundaries, while ServiceNow provides the workflows and business context. The goal is for agents to operate with clear permissions, visibility and control.

• Okta: its executive Ely Kahn noted that AI adoption stalls when there is no central way to cut off a rogue agent's access. Okta combines its "kill switch" capability with ServiceNow's containment decision layer to block new token requests from agents acting outside their intended scope. Looking ahead, the shared vision is to drive open standards that enable a cross-platform kill switch, revoking an agent's access across all necessary systems without disrupting the rest of the ecosystem.

• Metronome, a Stripe product: will provide reliable billing, real-time usage visibility and spending controls, at a time when software pricing is increasingly tied to usage and outcomes.

• WitnessAI: integrates its LLMs, designed to understand behavioral intent, with the Control Tower. They promise policy control, usage analytics and cost control, with behavior-based prompt routing.

• Zenity: presents itself as a pioneer in AI agent security and aims to extend that protection across the entire agent lifecycle within ServiceNow workflows.

The release does not detail what Anthropic, OpenAI or AWS will specifically contribute. The text received contains no comments from these companies, nor availability timelines, pricing or commercial terms. It also does not explain how solutions will be qualified or which agent identity standards will be used.

ServiceNow's own legal disclaimer lists risks worth keeping in mind. Among them are delays and difficulties in developing, integrating, certifying and deploying the alliance's offerings, regulatory changes in AI, collaboration challenges among members and third-party providers, and uncertainty over whether demand will justify the investment.

Overall, the announcement reflects a clear trend: governance, identity and cost control for AI agents are becoming a layer of enterprise software in their own right. It is notable that direct competitors, such as Anthropic and OpenAI, are taking part in the same alliance. For now, however, it is a commitment to collaborate rather than a product with measured results.