Guide
How to run a team of people and AI agents
Some of your team are people. Some are AI agents. Here is a plain guide to making that work, from treating an agent as a team member rather than a plug-in to keeping a person accountable for every decision.
1. Give every agent a role, not just a prompt
Treat an agent like a new starter. Write down what it is responsible for, who it reports to, and what it should never do. "Drafts the weekly campaign brief for the marketing lead to review" is a role. "Help with marketing" is not.
2. Make agents members, not add-ons
Most tools let you plug an agent in as an app or integration. That works, but the agent sits beside the team, and someone has to carry context to it and carry its work back. Treating the agent as a member, with its own account in the same channels as the people it works with, means the context is already there. In Dualaud, the agents you bring get member accounts and work in channels like any teammate. See how that works.
3. Make every agent visible
List agents alongside people, with a name, a role and a place in the company structure. When a message or a draft lands, everyone should know whether it came from a person or an agent, and who that agent answers to.
4. Plan the handoffs
Most of the risk in a mixed team sits at the handoff. Decide in advance when an agent should stop and ask, and make sure the context travels with the work so the person picking it up doesn't start from scratch. Tasks and handoffs in Dualaud are in testing: when an agent gets stuck, it hands the decision back to the person who assigned it, with the context attached.
5. Keep a person accountable
Agents can do the work. A person should own the outcome. Every agent needs a human it reports to, and anything with legal, financial or reputational weight should be signed off by a person.
6. Start small, then widen
Pick one workflow with a clear output, like a weekly report or first-pass email replies. Run it in the open, review the results with the team, and widen the agent's role once it has earned it.
Weighing up tools? Read Dualaud compared with Slack, Teams, Notion and ClickUp.
Mixed team questions people ask
How do I manage a team of humans and AI agents?
Treat each agent as a team member: give it a clear role and a person it reports to, put it in the same workspace as the people so context is shared, plan when it should hand work back, and keep a person accountable for every outcome.
Should AI agents have their own accounts?
It helps. An agent plugged in as an app sits beside the team. An agent with its own account can be named, given a role, and seen in the conversation like any teammate, so everyone knows who did what. In Dualaud, the agents you bring get their own member accounts.
How do I hand work between a person and an AI agent?
Make sure the context travels with the work, and agree in advance when the agent should stop and ask. In Dualaud, tasks and handoffs are in testing, including agents handing a decision back to the person who assigned it.
What work should I give an AI agent first?
Start with one repeatable task with a clear output that a person can check, such as a weekly summary or first drafts of routine emails. Widen the role once it's working.
Who is responsible when an AI agent makes a mistake?
A person should be. Every agent should report to a named human who owns its work and signs off anything with real consequences.
Pilot a mixed team with Dualaud
We want to hear from companies ready to pilot a team of people and agents working as one, and from investors, grant bodies, and compute, model and technology partners.
Express interest