Q.QIBA AIAI model, product, industry, and policy news
Sites
SearchSaved0

Explainer / Published 2026-10-29

OpenAI agents move into ServiceNow, and ServiceNow workflows into ChatGPT

OpenAI and ServiceNow expanded their January partnership: more models and agents in ServiceNow products, and governed ServiceNow actions inside ChatGPT and Codex. The release names no price.

Q.QIBA AI deskSource-backed article · 5 min read
Event date Prepared Published 2026-10-29
A glass desk holds a workflow board with colored connectors running through several permission gates, beside a single plug and a closed folder of policy notes.
Conceptual illustration of governed workflow actions; it does not depict ServiceNow, ChatGPT, or any real interface.

THE SHORT VERSION

  • OpenAI and ServiceNow expanded their partnership to put more OpenAI models and agents into ServiceNow products and connect ServiceNow workflows to ChatGPT and Codex.
  • A new integration lets an employee investigate incidents and update ServiceNow records inside ChatGPT, with ServiceNow's Action Fabric applying existing business rules and permissions.
  • The ServiceNow MCP plugin is available in the OpenAI marketplace now, with more capabilities planned through 2027; neither company published pricing for the integration.

The expansion runs in both directions

OpenAI and ServiceNow are expanding the partnership they announced in January. The release summary published by Markets Insider says the deal brings more OpenAI models and agents into ServiceNow products and connects ServiceNow workflows with ChatGPT and Codex. The companies describe the goal as making it easier for businesses to build applications, resolve IT requests, and address security issues with AI.

The direction matters for each company. OpenAI gets its models and agents into the systems where large enterprises already run IT service management, security operations, and internal workflows. ServiceNow gets its workflows into the tools where employees already ask questions, and a route into Codex for development work. The January agreement is the base the companies are building on, and the release describes the expansion as bringing OpenAI into more of the work ServiceNow does for its customers and inside its own business. The release summary is a company statement, not an independent test of the integration.

What an IT team can do inside ChatGPT

A new ServiceNow integration brings business context and workflow actions into ChatGPT. According to the release, an IT team member can find similar incidents, assess the risk of a change, and update a ServiceNow record without leaving the conversation. Customers connect ChatGPT to their own ServiceNow instance and choose which capabilities to make available.

The permission layer is ServiceNow's Action Fabric, which applies existing business rules and permissions so the actions a person can take reflect the organization's configuration and access. In a preview example, an on-call engineer reviews linked monitoring alerts, recent changes, and affected services, then assigns the right team and works toward a fix inside ChatGPT. ServiceNow also plans to give all its employees access to ChatGPT Work and Codex, beyond the engineering organization. The permission model is the part that determines whether an agent can change a record or only read it. The example is a preview described by the companies rather than a tested workflow. The integration does not replace the ServiceNow interface; it exposes a subset of records and actions inside the chat, chosen by the customer.

The rollout has dates, and they are staggered

The ServiceNow MCP plugin is available in the OpenAI marketplace today, with enhanced capabilities expected by the end of 2026. Voice capabilities in the Super Intelligence Control Tower are expected in the first quarter of 2027. The OpenAI-powered security capabilities are expected to enter early access in January 2027, with general availability planned for the second quarter of 2027. Codex in the AI Workflow Factory is expected in 2027.

The sequencing means the pieces arrive at different times. The plugin and the governed actions ship first, while voice, security, and deeper Codex integration are dated later. For an enterprise planning a rollout, the current capability is narrower than the full description of the partnership. The gap between the plugin available today and the dated features is where most enterprise rollouts will sit for the next year. The MCP plugin is the piece an administrator can install today; the rest is roadmap. The security capabilities have the longest lead time, moving from early access in January 2027 to general availability in the second quarter. The plugin is available now; the security and voice features are not. ServiceNow's own employees are the first internal deployment the release describes.

The governance layer comes with eleven members

ServiceNow also launched a Control Tower Alliance with eleven founding members: Anthropic, Amazon Web Services, Cisco, HiddenLayer, Kong, NVIDIA, Okta, OpenAI, Stripe, WitnessAI, and Zenity. Forkast's account describes the alliance as addressing security, governance, value measurement, and token metering inside the ServiceNow Super Intelligence Control Tower, a plane for discovering, securing, governing, observing, and measuring enterprise AI deployments.

The alliance is the governance half of the announcement. It includes competitors of OpenAI and names companies whose products cover identity, security, payments and policy. Whether an alliance of this composition can set shared standards, rather than produce a list of integrations, is a question the announcement does not answer. The alliance scope covers security, governance, value measurement and token metering, which are the controls a buyer would use to check spending. Payments and identity members suggest the alliance is meant to cover billing and access, not only model safety. For ServiceNow, the partnership is also a way to sell its workflow layer into conversations that start in ChatGPT.

The price is not in the release

Neither release lists pricing for the integration. Forkast reported that there are no pricing figures, tiers, or per-seat costs in ServiceNow's releases for the ChatGPT integration, while the alliance's members describe building usage measurement and cost controls. Scott Woody of Metronome, a Stripe product, is quoted saying pricing is increasingly tied to usage and outcomes; Dan Graves of WitnessAI is quoted describing AI policy control, usage analytics and cost controls.

The asymmetry is the point a procurement team should notice: the tools for measuring token use, spend and value are being built into the platform, while the rate card for the integration is not published. Forkast also noted that date metadata for the ServiceNow release is inconsistent across systems, with the page stamped October 28 but crawler metadata reading October 1 and October 22 in different places. The release leaves open how a mid-market customer would budget for a pilot before signing a contract. The date metadata inconsistency is a small but concrete example of how release records can differ across systems.

What is established, and what a buyer should ask

Established: the two companies expanded their partnership, the MCP plugin is available now, the governed action path runs through ServiceNow's existing rules and permissions, and the security, voice and Codex features have dated plans through 2027. Not established: what any of it costs, how usage will be metered in a customer contract, or how the alliance's standards will be enforced.

For a buyer, three questions follow from the release. Which ServiceNow records and actions can the ChatGPT integration reach under our configuration? How will ServiceNow and OpenAI bill for the usage, and at what volume does the contract change? And when the security capabilities reach general availability in the second quarter of 2027, which controls will be on by default? The announcement answers the first question in general terms and leaves the other two open. A buyer can answer the first question from the documentation; the other two require a negotiation. The integration is aimed at the IT service desk first, where incidents and change requests already have records and owners. The preview label means the capabilities may change before general availability. Related coverage includes OpenAI Dots agents, Google Cloud Gemini agent work, and Products. The site also covered Notion 3.8's agent context.

Sources & reporting limits

  1. OpenAI, ServiceNow expand partnership across enterprise workflows
  2. OpenAI Agents Get Governed Workflow Access via ServiceNow Action Fabric — With No Price Named
Keep the thread.

Follow Products to find related stories on your next visit.

Your next read

More Products →
Conceptual illustration of a dark room of branching glass corridors lit from one side, with thin green lines running along the floor toward a single open door.

Launch5 min read

Ex-GitHub CEO Thomas Dohmke launches Entire for AI agents

Thomas Dohmke, GitHub's former chief executive, launched Entire, a code platform for teams whose AI agents write faster than reviewers can keep up. Pricing is not public.

Products
Conceptual government workspace with an AI document review, an administrator setting a spending cap, and an audit log beside a protected device.

Explainer5 min read

Claude for Government is generally available, with capped usage billing

Anthropic moved its FedRAMP High-authorized government environment from public beta to general availability. Agencies receive administrative controls and prepaid spending caps, while some Microsoft 365 and command-line tools remain early access.

Products
THE WEEKLY EDIT

The week, made clear.

Stories worth understanding, brought together in one weekly edition.

Share this page

Instagram, TikTok & other apps

Copy the page link, then paste it where your chosen app supports links. You choose whether to send or publish.