Key Takeaways
- • Zenzap, a work chat platform with embedded AI agents, launched publicly with thousands of businesses already running on it, according to Bessemer Venture Partners, which has backed the company since inception.
- • Agents live in group chats as teammates and connect to more than 400 business apps, pulling data into threads, acting on it, routing approvals, and writing results back.
- • Any employee can create an agent in under 90 seconds without code or training, positioning the product against IT-led automation.
- • Two weeks after rollout to existing customers, 60% of those who enabled agents use them daily, agent activity grew roughly 20x, and one clinic recovered $1,583 in missed payments on its first day.
- • Security claims include isolated compute per conversation, no stored credentials, human approval for actions such as sending invoices, SOC 2 Type II audit completion, GDPR compliance, HIPAA support under a BAA, and EU or US data residency with BYOK.
Summary
Bessemer Venture Partners, which has continued to back Zenzap, describes the company as building chat for work designed to move employees off personal messaging apps such as WhatsApp. Co-founders Guy Weiss, Jonatan Bien and Noam Neumann set out three years ago to build a mobile-native product that is as intuitive as WhatsApp but governed like Slack and connected to the applications where work takes place. Bessemer characterized the bet as contrarian, since much of the technology sector regarded messaging as a solved market. Zenzap launched publicly today with thousands of businesses already on the platform.
The company's pitch centers on the shift from personal AI assistants to agents that operate across people, teams, tools and approvals. On Zenzap, agents sit in group chats as teammates, connected to more than 400 business apps. Bessemer's article gives the example of a clinic in which an agent picks up a patient cancellation, posts it to the reception channel, and a receptionist claims the slot. Zenzap distinguishes business agents, which run on roles with their own skills and tool access, from consumer super-agents.
Zenzap's design premise is that employees, rather than IT departments, should build agents. Any employee can create one in under 90 seconds with no code or training. Bessemer argues that IT cannot see the long tail of workflow inefficiencies, and that a large share of frontline staff, including restaurant managers, construction crews and hotel employees, never adopted Slack and are more likely to use agents they built themselves.
The company frames the security model as the answer to shadow IT. Each conversation runs in an isolated computer, agents never hold credentials, requests pass through a firewall the company does not control, and actions such as sending an invoice require a human to tap Approve. All activity is logged. Zenzap reports completing a SOC 2 Type II audit, GDPR compliance, HIPAA support under a BAA, and EU or US data residency with bring-your-own-key encryption.
Early usage figures were disclosed in the article. Two weeks after agents rolled out to existing customers, 60% of those who enabled them use them daily, having configured them independently. Agent activity grew roughly 20x over that period, and one clinic recovered $1,583 in missed payments on its first day. These figures are company-reported and have not been independently verified.
The company's pitch centers on the shift from personal AI assistants to agents that operate across people, teams, tools and approvals. On Zenzap, agents sit in group chats as teammates, connected to more than 400 business apps. Bessemer's article gives the example of a clinic in which an agent picks up a patient cancellation, posts it to the reception channel, and a receptionist claims the slot. Zenzap distinguishes business agents, which run on roles with their own skills and tool access, from consumer super-agents.
Zenzap's design premise is that employees, rather than IT departments, should build agents. Any employee can create one in under 90 seconds with no code or training. Bessemer argues that IT cannot see the long tail of workflow inefficiencies, and that a large share of frontline staff, including restaurant managers, construction crews and hotel employees, never adopted Slack and are more likely to use agents they built themselves.
The company frames the security model as the answer to shadow IT. Each conversation runs in an isolated computer, agents never hold credentials, requests pass through a firewall the company does not control, and actions such as sending an invoice require a human to tap Approve. All activity is logged. Zenzap reports completing a SOC 2 Type II audit, GDPR compliance, HIPAA support under a BAA, and EU or US data residency with bring-your-own-key encryption.
Early usage figures were disclosed in the article. Two weeks after agents rolled out to existing customers, 60% of those who enabled them use them daily, having configured them independently. Agent activity grew roughly 20x over that period, and one clinic recovered $1,583 in missed payments on its first day. These figures are company-reported and have not been independently verified.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.