Viktor

Viktor is an AI coworker that lives inside Slack and Microsoft Teams, where a whole team can message it like a colleague to get real work done, not a chatbot that answers questions but an agent that executes them. Built by Zeta Labs, whose founders Fryd Wiatrowski and Peter Albert are former Meta engineers behind the earlier Jace email assistant, Viktor launched publicly in early 2026 and raised a 75 million dollar Series A led by Accel in May, with angels including Slack cofounders Stewart Butterfield and Cal Henderson. The company reported reaching a 15 million dollar annualized revenue run rate within three months and more than ten thousand workspaces using it.

Each workspace gets its own cloud computer where Viktor writes and runs code, operates a real browser to fill forms and navigate applications, and connects to more than three thousand tools like HubSpot, GitHub, Stripe, Google Drive, and Meta Ads. Team members ask it to pull a cross tool report, reconcile invoices, build an internal app, audit ad spend, or triage a support queue, and it does the work in its sandbox and posts the result back into the conversation. A heartbeat system watches how the team operates and proposes automations it can then run on a schedule, so recurring work like a Monday revenue briefing runs on autopilot once approved. It builds institutional memory over time, learning a company's brand, terminology, and preferences.

Viktor keeps humans in control by design: it proposes and waits, asking before anything it cannot undo, pushing back when a mistake looks likely, and holding every sensitive action for approval, with admin controls over who can use sensitive integrations. It is SOC 2 Type 1 certified with Type 2 and ISO 27001 in progress, and its credit based pricing starts free and scales from a Team plan to enterprise. For a Slack or Teams centric team that wants an agent to actually finish operational work across many tools rather than assist one person, Viktor is a strong fit; organizations needing self hosting, deep audit tooling, or open model choice will find those are still maturing.