Platforms We Deploy

    OpenAI AgentsAgents your whole team can run in ChatGPT

    Pyra designs, deploys, and manages OpenAI agents for GTM, sales, marketing, and operations teams. According to OpenAI's documentation, ChatGPT workspace agents can be built and tested in ChatGPT, connected to apps and tools, shared across a workspace, used in Slack, run on a schedule, or triggered through an API.We turn that into scoped, reviewed workflows your team actually adopts — built to your stack, with approval gates before anything acts.

    Independent implementation partner: ChatGPT and the Agents SDK are products of OpenAI. Pyra is not affiliated with, sponsored by, or endorsed by OpenAI — we design, deploy, and manage agent workflows on the platform for our clients. Platform facts verified August 18, 2026.
    Human gates
    Approval before actions
    Test first
    Preview before publishing
    Access control
    Role-based permissions
    Managed rollout
    Admin-controlled enablement
    Who this is for

    Teams already living in ChatGPT and Slack

    GTM & RevOps Leaders

    Want repeatable research and prep work packaged as agents the whole team can run — with access controls your admins set and results humans review.

    Sales & Marketing Teams

    Need meeting prep, account research, and drafted content available where the team works — in ChatGPT and Slack — left as drafts to review before anything ships.

    Operations Leaders

    Want recurring workflows to run on a schedule or fire from an API trigger, with role-based access controls and human checkpoints where it matters.

    What we deploy it for

    Four workflows we build on OpenAI's platform

    Every workflow below is grounded in documented platform capabilities and configured with human review and approval gates. Pyra sells the implementation and management service.

    1) GTM: Account research & prep
    A workspace agent that owns read-and-prepare work across your GTM stack and leaves results for review.
    • Research target accounts using connected apps and tools
    • Assemble briefs from your files and connected sources
    • Run on a schedule or on demand from ChatGPT or Slack
    • Human review before any downstream action
    2) Sales: Meeting prep to drafted outreach
    Mirror the sales meeting-prep workflow from OpenAI's own cookbook: prep packaged as an agent the team shares.
    • Pull account context from your connected CRM and files
    • Prepare meeting briefs before every call, automatically
    • Draft follow-ups and outreach in your team's voice
    • Leave drafts to approve — nothing sends without sign-off
    3) Marketing: Content & campaign prep
    A shared agent with your brand files and skills attached, producing drafts for team review.
    • Work from uploaded brand files, skills, and connected apps
    • Assemble draft briefs, copy, and campaign material
    • Share with the team or workspace under access controls
    • Approval gate before publishing or scheduling
    4) Custom: Agents SDK builds in your product
    For workflows beyond ChatGPT, we build with OpenAI's Agents SDK inside your own application.
    • Custom agents in Python or TypeScript on your infrastructure
    • Long-horizon tasks in controlled sandbox environments
    • Wired to your data and tools with scoped access
    • Human checkpoints designed into the workflow
    Why OpenAI

    What the platform actually does

    The following facts are drawn from OpenAI's documentation, verified August 18, 2026. Where a claim is made by the vendor, we attribute it as such.

    Workspace agents, built in ChatGPT

    Per OpenAI's documentation, Business and Enterprise workspaces can build agents for repeatable tasks directly in ChatGPT — with a choice of model and reasoning effort, a preview mode to test before publishing, and sharing to teammates or the whole workspace.

    Apps, tools, MCP, and files

    OpenAI's documentation describes attaching tools, apps, custom MCP servers, skills, and files to an agent in the builder — and running the agent in ChatGPT, in Slack, on a schedule, or through an API trigger.

    Admin controls by default

    Per OpenAI, workspace agents are off by default at launch for Enterprise workspaces and admins enable them for eligible workspaces. Role-based access controls determine which agents each user can see and run — per OpenAI, the same controls apply in ChatGPT and in the Codex Workspace Agents plugin.

    Agents SDK for custom builds

    OpenAI updated the Agents SDK in April 2026 to help developers build agents that can inspect files, run commands, edit code, and work on long-horizon tasks within controlled sandbox environments. Separately, OpenAI announced in June 2026 that the AgentKit Agent Builder and Evals products are being wound down as of November 30, 2026.

    How Pyra deploys it

    Scope, integrate, then manage

    1. Scope & pilot

    We pick one workflow with a clear result, choose the right surface — a workspace agent in ChatGPT or a custom Agents SDK build — and start with read-and-prepare work.

    2. Integrate & configure

    We connect your apps, tools, MCP servers, and files, work with your admins on enablement and role-based access, and set approval gates before any action that sends or changes something.

    3. Manage & iterate

    We monitor, tune, and expand — adding schedules, Slack, or API triggers only once results hold up in review — and track OpenAI's platform changes so your workflows don't get stranded.

    FAQ

    Common questions about our OpenAI deployments

    Is Pyra affiliated with OpenAI?

    No. Pyra is an independent implementation partner. We are not affiliated with, sponsored by, or endorsed by OpenAI. We design, deploy, and manage agent workflows on OpenAI's platform for our clients. ChatGPT and the Agents SDK are products of OpenAI.

    Which OpenAI agent surface do you build on?

    Two, depending on the workflow. For team-facing work we build ChatGPT workspace agents — per OpenAI's documentation, these can be built and tested in ChatGPT, shared with a team or workspace, used in Slack, run on a schedule, or triggered through an API. For custom application work we build with the Agents SDK. Note that OpenAI announced in June 2026 that the AgentKit Agent Builder and Evals products are being wound down as of November 30, 2026 — we do not start new work on them, and part of our management service is tracking platform changes like this.

    Who controls what an agent can see and do?

    Per OpenAI's documentation, workspace agents are off by default at launch for ChatGPT Enterprise workspaces, admins enable them for eligible workspaces, and role-based access controls determine which agents each user can see and run. On top of those platform controls, Pyra designs the workflow itself with human review: agents prepare work and leave drafts, and approval gates sit before any action that sends or changes something.

    What's the fastest way to start?

    Start Free to scope one repeatable workflow with a clear result. From there we build and test the agent, connect your apps and tools, set sharing and access controls with your admins, and pilot before expanding — with monitoring and human review throughout.

    Next step

    Put an OpenAI agent to work on one real workflow

    Start Free to scope a single read-and-prepare workflow with a clear result. Then Pyra builds and tests the agent, connects your tools, sets access controls and approval gates, and manages the deployment as it expands. Independent implementation partner — not affiliated with OpenAI.