Bourne
Also known as: usebourne.com, Bourne Studio, Bourne Solutions, Bourne by HockeyStack, Bourne AI labor platform
Bourne is the AI labor platform from the team behind HockeyStack. Bourne Studio connects a company's systems, maps how its work flows into a shared ontology, and has a coding agent build agents, automations and apps from plain language, which Bourne hosts in its cloud or the customer's AWS, Azure or GCP account. Every change is versioned and tested against real business cases, people approve the actions they choose, and Bourne Solutions works alongside companies on their first deployments.
Bourne comes from HockeyStack, Inc., the company behind the HockeyStack attribution platform, and launched on September 22, 2026 under HockeyStack co founder and CEO Bugra Gunduz. It sells two things: Bourne Studio, where builders create, host and eventually sell AI agents, apps and automations, and Bourne Solutions, which works with companies to find work AI can take on and deploy it around their existing systems. The company describes itself as an autonomous enterprise business and says it is building toward a marketplace for AI labor once builders have stocked it.
Studio works in four moves: connect the company's systems, map how work flows, restructure that work into a form machines can run, then automate it. Connectors reach systems with APIs, databases and file feeds, documents are read as data, and computer use covers legacy applications and portals with no API.
An ontology holds the business's objects, such as customers, orders, parts, assets and people, along with process maps and a memory of past decisions, so that when a person resolves an exception the next similar case can run without them. On top sit agents that plan and act over hours or days, automations that fire on a trigger, and apps built around each person's job, reachable from Slack, Microsoft Teams, email and voice. A coding agent builds all of it from plain language descriptions, with version control and rollback, and the platform picks a model for each step.
Every change is tested against real business cases before it ships, a failing eval sends the coding agent back to fix and retest, and production runs are tracked so failures and drift are flagged. Customers set which actions need a person's approval, and uncertain cases reach a person with the evidence attached.
Governance covers per agent permissions and an audit trail, security covers SAML single sign on, role based access, TLS 1.2 in transit and AES 256 at rest, and HockeyStack's trust center lists SOC 2 Type II. Bourne hosts and scales the work in its own cloud or the customer's AWS, Azure or GCP account. Sign up and a first prototype, built from data extracts the customer sends, are free, and Bourne lists no paid tiers.
Bourne fits operations heavy companies such as manufacturers, distributors and services firms whose work spans ERP, CRM and legacy portals and who want a partner to build the first automations with them.
Its examples include an industrial manufacturer quoting in five minutes, a steel company selling excess steel worldwide, a software company resolving 68% of support tickets and a benefits administrator running payroll and benefits work, and it cites more than 300 enterprises, logos such as Microsoft, Rakuten, Mastercard, BNP Paribas, Proofpoint and Tipalti, and 230,000 hours of AI labor completed a year. It fits less well for developers who want an API or SDK, buyers who must pick their own models, or teams wanting published prices before they start.
Vendor details
Canonical URL
https://usebourne.com
Category
Agent builder
Subcategory
HockeyStack's platform for building, hosting and governing AI agents, automations and apps that run a company's operational work, with a coding agent, an ontology and evals
Funding status
A product of HockeyStack, Inc., which has raised $45 million from Bessemer Venture Partners, Y Combinator, General Catalyst, Madrona and Kleiner Perkins.
Company status
first party product
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Connectors for systems with APIs, databases and file feeds, including ERP, CRM and machine data; document reading; computer use for legacy applications and portals without APIs; Slack, Microsoft Teams, email and voice as channels; hosting in Bourne's cloud or the customer's AWS, Azure or GCP account.
In practice
An industrial manufacturer connects its ERP and pricing sheets. A Bourne agent reads each incoming request for quote, pulls part data from the ontology, prices the job and returns a quote in about five minutes, while quotes above a set value wait for a sales engineer's approval.
A software company routes support tickets from email and Slack to a Bourne agent that resolves about 68% of them on its own and sends the rest to an agent with the evidence it gathered, and each human resolution goes into memory for the next similar ticket.
A benefits administrator describes its payroll reconciliation in plain language. The coding agent builds the automation, tests it against last quarter's real cases before release, and Bourne flags drift when a carrier changes its file format.
Sources & related URLs
Related / legacy domains
Agentic Index coverage score
9.5 / 14 capabilities · 68%
| Integrations & Tool Calling | Partial |
|---|---|
|
Connectors reach any system with an API, a database or a file feed, which in Bourne's examples means ERP, CRM and machine data, and computer use handles legacy applications and portals that offer no API. The agents act on what they find, issuing quotes, selling surplus stock, closing support tickets and running payroll steps. Bourne names no specific products it connects to and offers no catalogue of connectors or published way for a customer to add a tool of its own. Sourceusebourne.comread 2026-11-19 |
|
| Workflow Orchestration | Full |
|
Work is first mapped as a process and restructured into steps a machine can run, then carried out by three kinds of operation: agents that plan and act over hours or days, automations that run whenever their trigger fires, and apps built around each person's job. A quote request, for example, moves from intake through part lookup and pricing to delivery, with uncertain steps handed to a person and the run resuming afterward, and the platform chooses a model for each step along the way. Sourceusebourne.comread 2026-11-19 |
|
| Knowledge Grounding & RAG | Full |
|
An ontology sits at the center of the platform: a shared model of the business's objects, such as customers, orders, parts, assets and people, joined to process maps that show how work moves between them. Connected systems and documents feed it, with document reading turning PDFs and forms into structured data. Agents, automations and apps all work from this one model rather than assembling context per run, so a pricing agent and a support agent see the same customer record. Sourceusebourne.comread 2026-11-19 |
|
| Human Oversight & Guardrails | Full |
|
Customers decide which actions need a person's approval, set per agent in Bourne's governance settings, and anything the agent is unsure about goes to a person with the supporting evidence attached rather than proceeding on a guess. The person's answer both releases that case and feeds the decision memory. Per agent permissions limit what each agent can touch, and every change an agent or builder makes is versioned so it can be rolled back. Sourceusebourne.comread 2026-11-19 |
|
| Security, Identity & Governance | Full |
|
Sign in runs through SAML single sign on, role based access sets what each person can do, and per agent permissions set what each agent can reach. Data travels over TLS 1.2 or higher and rests under AES 256 encryption. Bourne's security link opens HockeyStack's trust center, which lists SOC 2 Type II and SOC 2 Type I reports with access on request, alongside infrastructure and organizational controls such as restricted encryption key and production database access. Sourcesecurity.hockeystack.comread 2026-11-19 |
|
| Observability & Auditability | Full |
|
Every production run is tracked, and failures and drift are flagged as they happen, so an automation that starts going wrong surfaces before its errors pile up. An audit trail records what agents did under governance, version control keeps the history of every change to an agent, automation or app, and escalations carry their evidence with them. Together these let an operator reconstruct why a case was handled the way it was, not only that it ran. Sourceusebourne.comread 2026-11-19 |
|
| Memory & State Persistence | Partial |
|
Bourne keeps a memory of past decisions, history and how each exception was resolved, and every decision a person makes is written into it so that the next similar case can run on its own. That makes the agents learn from approvals over time. How long those memories last, whether they are kept per agent, team or customer, and how one entry would be reviewed or removed are not set out, which leaves the store closer to an accumulating record than a governed memory. Sourceusebourne.comread 2026-11-19 |
|
| Deployment & Data Residency | Full |
|
Bourne hosts and scales agents, automations and apps in its own cloud or in the customer's, and the customer's option covers its own AWS, Azure or GCP account. Deploying inside the customer's account keeps data and running agents within infrastructure the customer already governs and in whatever region it already uses. Bourne does not list the regions of its own hosted cloud, and an on premises option outside those three clouds is not offered. Sourceusebourne.comread 2026-11-19 |
|
| Prebuilt Agents / Templates / Packs | Partial |
|
Bourne's approach is to build for each company rather than hand over finished agents. When Bourne Solutions solves a problem other builders will meet, it places the reusable piece on the platform so each deployment costs less than the last, and its first domain solution covers an industrial manufacturer's customer lifecycle from inquiry onward. A marketplace where companies would pick ready made AI labor from other builders is planned for later and is not open. Sourceusebourne.com/newsread 2026-11-19 |
|
| Triggers & Channel Coverage | Partial |
|
People reach Bourne's agents through Slack, Microsoft Teams, email and voice, and automations start when a trigger fires, such as a new order, ticket or file arriving in a connected system. Long running agents keep working over hours or days once started. The kinds of trigger on offer, whether schedules, webhooks or record events, are not spelled out, so it is unclear which events a team can wire to an automation without the coding agent building that link. Sourceusebourne.comread 2026-11-19 |
|
| Model Flexibility & Routing | Not documented |
|
Bourne selects the best model for each step itself, and it names neither the models nor the companies behind them. Customers get no picker, no bring your own key option and no routing rules of their own, so the choice of model stays with Bourne whether the work runs in its cloud or the customer's. The coding agent that builds automations works the same way, with its model left unstated. Sourceusebourne.comread 2026-11-19 |
|
| APIs / SDKs / MCP Extensibility | Not documented |
|
Building happens inside Bourne Studio through its coding agent, with version control, rollback and hosting included. Bourne offers no developer documentation, public API, SDK or MCP server through which outside code could create agents, start runs or read results, and the site links to no docs. Teams that want to extend what an automation does describe the change to the coding agent rather than writing against an interface of their own. Sourceusebourne.comread 2026-11-19 |
|
| Testing, Debugging & Optimization | Full |
|
Every change is tested against real business cases before it ships, so an edited quoting agent must handle last month's actual requests correctly first. When an eval fails, the coding agent fixes the automation and retests until it passes, which makes testing part of the build loop rather than a manual step. After release, an improvement loop and production monitoring watch live runs for failures and drift. Sourceusebourne.comread 2026-11-19 |
|
| Browser / Computer-use | Partial |
|
Computer use is part of Bourne's data layer, aimed at legacy applications and supplier or customer portals that have no API, so an agent can log in and work through screens a person would otherwise click through. How it runs, whether in a hosted browser, a virtual desktop or the customer's own machines, and what limits or approvals apply to it are not described, which keeps it a stated capability rather than a mechanism a buyer can inspect. Sourceusebourne.comread 2026-11-19 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Pricing
Not published; sign up and a first prototype are free
Not stated
What is public
Bourne offers a free sign up with no credit card and a free first prototype, which it builds from data extracts the customer sends. It lists no paid plans or prices.
Billing mechanics
Bourne gives no billing unit for Studio or for Bourne Solutions engagements.
Cost watchouts
Deploying in the customer's own AWS, Azure or GCP account puts the hosting bill on the customer's cloud account.
Variable cost rationale
Bourne gives no billing unit, so it is unknown whether cost follows agents, runs, hours of AI labor or a services engagement.
Additional watchouts
Running agents in the customer's own cloud adds that cloud's hosting costs to whatever Bourne charges.
Sales call required
Mixed (some tiers require a call)
Free / trial
Free sign up with no credit card, and a free prototype built from data extracts the customer sends
Commercial notes
Bourne Solutions works alongside companies to find and deploy the work AI can take on, and Bourne plans a marketplace where builders sell AI labor once Studio holds enough products.
Missing data
Bourne gives no price for Studio, hosting or Bourne Solutions work.
Related vendors
- Lindy — Lindy is an AI teammate that lives in Slack, answers from your tools…
- Activepieces — Open source, MIT licensed automation platform where every one of 700…
- Agenta — Agenta is an open source workspace where teams build AI agents by…
- AgentX — No code platform to build, evaluate, and deploy multi agent…
- AI Library — No code platform for building and running AI agents, with a…
- Aigensei — Enterprise agentic workflow platform that runs business processes…