Technology
Convey's $38 Million Round Shows The AI Coworker Pitch Is Moving From Demo To Workflow Ownership
Convey has raised $38 million to build AI coworkers for repetitive office workflows, according to Business Insider. The round shows enterprise AI shifting from assistants that draft text to agents that are expected to own repeatable business outcomes.
By Patrick T ·

Convey has raised $38 million to build AI coworkers for repetitive office work, Business Insider reports, with Andreessen Horowitz leading the round. The young startup is entering one of the most crowded parts of the market, but its framing is revealing: the next enterprise AI sale is not about a better chatbot. It is about software that can be assigned recurring work and judged on whether the work gets done.
That language matters. Enterprise buyers have spent the last two years experimenting with copilots, search boxes, and meeting summaries. Now they are asking whether AI can become a unit of labor for sales operations, recruiting, finance, support, compliance, procurement, and other back-office workflows where the work is structured enough to automate but messy enough to require judgment.
The Teammate Metaphor Has A Budget Line
The AI coworker pitch works because it maps to how managers already think. A team has tasks, owners, handoffs, deadlines, metrics, and escalation paths. If an AI product can fit into that operating model, it can be evaluated less like a tool and more like a role. That makes procurement easier to justify when the workflow is painful and measurable.

The risk is that the metaphor gets ahead of the controls. A real coworker can be trained, corrected, promoted, fired, or asked to explain a decision. An AI worker needs the software equivalent: permissions, logs, test cases, exception handling, audit trails, and a clear boundary around when it should stop and ask a human.
Why Office Work Is The Wedge
Office workflows are attractive because they are usually digital, repetitive, and spread across systems. A human coordinator may copy information from a spreadsheet to a CRM, draft a message, check a policy, update a ticket, and notify a manager. None of those steps is dramatic, but together they consume enormous operating time.

This is why the market is fragmenting into vertical and role-specific agents. General models are powerful, but businesses buy packaged outcomes. A startup that understands the exact workflow, data permissions, and approval path can often beat a more general assistant that knows more but owns less.
Agents Need Accountability
Topics: Convey, AI agents, enterprise AI, workflow automation