Convey
Platform to train, ship, and manage enterprise grade AI workers.
Convey is an enterprise AI workforce platform that lets non technical operators train, ship, and manage digital teammates that own operational work autonomously in the background, instead of co pilots that only make one person faster.
Founded in San Francisco in 2025 by Rohan Chopra, an early DoorDash employee who spent eight years helping scale it past ten thousand people, with Stanford friends Will Harvey and Diego Canales, the company raised a 38 million dollar Series A led by Andreessen Horowitz in June 2026, with Khosla Ventures and Pear VC continuing from the seed.
It has completed more than one million hours of automated work for customers including NBCUniversal, Samsara, Unity, TelevisaUnivision, Faire, and ChargePoint.
The core mechanic is teaching by demonstration. An operator describes a process or shares their screen, and the teammate observes, learns, and takes over execution, running reliably across hundreds of thousands of repetitions. Because teammates can operate a computer directly, clicking through applications the way a person would, Convey reaches legacy systems that lack modern APIs, not just tools with clean integrations. Provisioning is a partnership with IT.
Each teammate gets a distinct identity and role based permissions, so operators closest to the work can build in a governed way. Common jobs include invoice processing, financial reconciliation, campaign reporting, and advertising asset ingestion, with customers reporting outcomes like a forty percent EBITDA gain and tens of thousands of hours reclaimed.
Convey positions itself against enterprise AI's common failure mode of adoption without impact, tying teammates to measurable outcomes on high value operational work and using a white glove motion where its specialists help stand up the workforce and then hand back control.
It is SOC 2 Type II certified and HIPAA compliant, keeps customer data out of model training, and maintains zero retention agreements with providers. For an operations heavy enterprise that wants non engineers to automate recurring back office work end to end, including across legacy systems, Convey is a strong fit. Teams wanting a self serve developer framework or an open model choice will find it is a managed, outcome focused platform instead.
Vendor details
Canonical URL
https://convey.dev
Category
Agent builder
Subcategory
Enterprise AI worker platform
Funding status
Independent, founded in San Francisco in 2025 by Rohan Chopra, an early DoorDash employee, with Will Harvey and Diego Canales. Raised a 38 million dollar Series A led by Andreessen Horowitz in June 2026, with Khosla Ventures and Pear VC continuing from the seed round. Reports more than one million hours of automated work completed.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Connects to company systems either programmatically or by operating the computer directly, clicking through applications like a person, which lets it work with legacy tools that lack modern APIs. Each digital teammate is provisioned with its own identity and role based permissions in partnership with IT.
In practice
Your best operators lose two to three hours a day to invoice processing and reconciliation across portals. A Convey teammate learns the process by watching once, then owns it unattended across thousands of runs.
A critical workflow lives in a legacy system with no API, so nothing can automate it. Convey's teammate operates the application directly, clicking through it like a person, and takes the task off your plate.
You want operations staff, not engineers, to automate their own recurring work. An operator shares their screen, the teammate learns, and IT governs identity and permissions so it runs securely.
Sources & related URLs
Related / legacy domains
Research sources
Agentic Index coverage score
8.0 / 14 capabilities · 57%
| Integrations & Tool Calling | Full |
|---|---|
|
Teammates connect to critical company systems either programmatically or by operating the computer directly, including legacy tools and applications that lack modern APIs. Integrations are configured in partnership with IT. Named workloads span shipping portals, spreadsheets, financial systems, reporting tools and advertising asset systems. Teammates write as well as read, through invoice processing, financial reconciliation, order intake and report generation. Each teammate connects under its own identity with role based permissions. Sourceconvey.dev, Convey Series A announcement June 2026 and accompanying coverageread 2026-08-31 |
|
| Workflow Orchestration | Full |
|
Digital teammates own multistep operational workflows end to end across multiple tools and systems. Named workloads include invoice processing across shipping portals, financial reconciliation, recurring reporting, order intake and advertising asset ingestion. Teammates take ownership of responsibilities people previously handled manually, instead of assisting an individual, and they execute autonomously in the background. After teaching, execution is deterministic, following verified steps consistently across repetitions. Customers report tens of thousands of hours reclaimed annually. Teammates do not coordinate with or hand off work to each other. Sourceconvey.dev, Convey Series A announcement June 2026 and accompanying coverage; businesswire.com/news/home/20260617486214/en/convey-raises-$38-million-series-a-led-by-andreessen-horowitz-to-automate-enterprise-operations-with-ai-teammatesread 2026-08-31 |
|
| Knowledge Grounding & RAG | Partial |
|
Teammates learn tasks and gather context as an operator describes a process or shares their screen, and the teammate observes, learns and takes over execution. At run time, teammates read live data from connected systems and application interfaces across invoices, financial records, orders and reports. There is no knowledge base, document index, ingestion pipeline, retrieval configuration or curated corpus for customers to maintain. Sourceconvey.dev, Convey Series A announcement June 2026 and accompanying coverageread 2026-08-31 |
|
| Human Oversight & Guardrails | Partial |
|
Teammates operate autonomously but reach out for help when they need the operator. In one example, a teammate asks by email whether to pause a campaign or roll an overspend into the next month, and its daily view flags items for review. Each teammate has per app permissions configured with IT, set to read and write, read only, or a single module. Organization wide roles govern who can manage it. The teammate decides when to escalate, and there is no configured approval step before a class of action. Sourceconvey.dev and convey.dev/enterpriseread 2026-09-29 |
|
| Security, Identity & Governance | Full |
|
The platform is SOC 2 Type II certified and HIPAA compliant, with regular penetration testing and reports in its Trust Center at trust.convey.dev. Users sign in with SAML SSO and MFA through the customer's identity provider. Organization wide roles and granular permissions control access, and user access management assigns people to teammates by role. Each agent has an isolated identity with dedicated credentials, explicit access controls and per app read, write or scoped permissions. Customer data is not used for training, and zero data retention agreements are held with model providers. BAAs are available for qualifying enterprises. Sourceconvey.dev/enterprise and convey.devread 2026-09-29 |
|
| Observability & Auditability | Full |
|
Audit Logs record every change, action and integration. They appear as an activity feed of a teammate's individual actions, such as opening a named document, building a media plan and exiting the document. A view of each teammate's work in progress has step counters, a daily recap of completed tasks, items flagged for review and time saved. Every teammate acts under its own identity, so actions are attributable. Sourceconvey.dev/enterprise and convey.devread 2026-09-29 |
|
| Memory & State Persistence | Partial |
|
Every agent has its own private data store. Teammates learn a task once by observation and then own it across repeated runs. In one example, a teammate recalls vendor issues over thirty days and proposes new responsibilities. Convey does not say what a teammate keeps, how long it keeps it, or how anyone reviews or deletes it. Sourceconvey.dev/enterprise and convey.devread 2026-09-29 |
|
| Deployment & Data Residency | Partial |
|
Tenant specific execution environments run automations in the customer's own private sandboxed environment, and every agent has its own private data store and dedicated credentials. This single tenant isolation runs on Convey's own hosting. Customers cannot deploy into their own cloud or on premises, and there is no region list or way to choose where data resides. Sourceconvey.dev/enterpriseread 2026-09-29 |
|
| Prebuilt Agents, Templates & Packs | Not documented |
|
Operators with no technical background create teammates by teaching a process through description or screen sharing. Each teammate is built for the customer's own process instead of being selected from a library. Convey specialists help with onboarding in a white glove motion before handing back control, and an end to end teammate is onboarded in around three hours. There is no template catalog, prebuilt agent library, shareable teammate format, component library or starter pack. Sourceconvey.dev, Convey Series A announcement June 2026 and accompanying coverageread 2026-08-31 |
|
| Triggers & Channel Coverage | Full |
|
Each teammate has a calendar of upcoming scheduled work, such as checking campaign pacing in thirty minutes and a NetSuite prepaid reconciliation in an hour. Teammates are built for unsupervised execution and work in the background, and scheduling and coordination are among their common uses. Each teammate has its own email identity, through which it writes to people and receives replies. Teams and Outlook Calendar are among the connected apps. There is no webhook or event trigger. Sourceconvey.dev and convey.dev/enterpriseread 2026-09-29 |
|
| Model Flexibility & Routing | Not documented |
|
The platform manages models itself. Customers get no model selection, model picker, bring your own key arrangement, endpoint configuration or routing. Zero data retention agreements are in place with model providers, and customer data stays out of model training. Convey sells the product as a managed, outcome focused platform, not one with open model choice, and execution is deterministic, following verified steps. Sourceconvey.dev, Convey Series A announcement June 2026 and accompanying coverageread 2026-08-31 |
|
| APIs, SDKs & MCP Extensibility | Not documented |
|
Convey connects to company systems programmatically or by operating the computer directly. It also works with IT managed identity, permissions and governance, so the customer's identity provider governs Convey. There is no public application programming interface, software development kit, webhook endpoint or Model Context Protocol server, and no route by which an external system could invoke a digital teammate or manage the platform. There is no developer documentation site. Convey sells the platform as managed and outcome focused, not as a self serve developer framework. Sourceconvey.dev, Convey Series A announcement June 2026 and accompanying coverageread 2026-08-31 |
|
| Testing, Debugging & Optimization | Not documented |
|
Execution is deterministic. Operators teach tasks through visual interfaces, and teammates then follow verified steps every time. Convey contrasts this with probabilistic models that may produce incorrect financial data. Customers get reporting on completed work and cite outcome measures such as hours reclaimed. There is no test suite, scoring, accuracy measurement, simulation or regression comparison between versions of a taught procedure. Nothing checks a teammate's output before it is committed. Sourceconvey.dev, Convey Series A announcement June 2026 and accompanying analysis; aicerts.ai/news/conveys-38m-bet-on-enterprise-ai-teammatesread 2026-08-31 |
|
| Browser & Computer Use | Full |
|
Digital teammates operate a computer directly, clicking through applications the way a person would, to reach legacy tools and applications that lack modern APIs. They also connect programmatically where an interface exists. An operator teaches a teammate by describing a process or sharing their screen, and the teammate observes, learns and takes over execution. One customer automated hundreds of hours of invoice processing across shipping portals. Sourceconvey.dev, Convey Series A announcement June 2026 and accompanying coverage; businesswire.com/news/home/20260617486214/en/convey-raises-$38-million-series-a-led-by-andreessen-horowitz-to-automate-enterprise-operations-with-ai-teammatesread 2026-08-31 |
|
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 public; enterprise sales, plan based and scaling with value delivered
plan scaled to value delivered
What is public
No public rate. Convey publishes no list pricing; plans and terms are set through sales.
Billing mechanics
Plan based enterprise pricing that scales with the value delivered, meaning the volume and impact of automated work. Onboarding is a white glove engagement where Convey specialists help stand up teammates before handing back control.
Cost watchouts
Because pricing scales with value delivered, expanding teammates across more processes or departments raises the plan. Confirm the onboarding cost and what implementation support the plan includes.
Variable cost rationale
Plan based pricing that scales with the value or volume of automated work; predictable once scoped, but expanding teammates across more processes raises the plan.
Additional watchouts
Value scaled pricing is hard to benchmark without a public rate; size the plan against the operator hours each teammate reclaims and confirm how adding teammates or processes changes the plan.
Sales call required
Yes, required for paid access
Free / trial
Demos and white glove onboarding through sales; no public free tier
Key ambiguities
No entry rate or billing unit is published; billing is described only as plan based and value scaled, negotiated under sales.
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…
Alternatives to Convey
The closest documented capability profiles to Convey among agent builders tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Aigensei8.0 / 14Adds documented Prebuilt Agents, Templates & Packs and APIs, SDKs & MCP Extensibility
- Flozic10.5 / 14Adds documented Prebuilt Agents, Templates & Packs and Model Flexibility & Routing, among others
- Toolhouse8.5 / 14Adds documented Prebuilt Agents, Templates & Packs and APIs, SDKs & MCP Extensibility, among others
- Maisa9.0 / 14Adds documented Prebuilt Agents, Templates & Packs and Model Flexibility & Routing, among others
- Minded12.0 / 14Adds documented Prebuilt Agents, Templates & Packs and Model Flexibility & Routing, among others
- Play9.0 / 14Adds documented Prebuilt Agents, Templates & Packs and Model Flexibility & Routing
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded