AI Agents: A digital team that works the way your company does
Describe an agent's job in plain words, connect agents into a workflow, and keep a person in the loop at every step that matters.
What is AI Agents?
AI Agents in kanês are a digital team that carries out work for your company: built-in specialists for operations, sales and strategy, plus agents your company defines. They follow workflows you draw, hand work along from step to step, and stop for a person wherever a review is needed.
How it works
- 1
Define the agent
Its responsibilities, how it should work, what to avoid, when to ask and what done looks like. What it produces (a written deliverable, tasks or a proposal draft) fixes which tools it may use.
- 2
Draw the workflow
A playbook connects up to twelve steps: agent work, tasks, a proposal draft, a teammate's stage and review steps. Edges say what must be done first.
- 3
Run it
Each run uses a frozen copy of the published workflow, so editing a draft never changes work in progress.
- 4
Review and approve
A review step stops the run until the person who started it checks the work. A teammate's step becomes a task on their list.
What it does
- Built-in specialists: Operations, Sales and Strategy
- Company agents you define, each limited to the tools its output needs
- Workflows (playbooks) of up to 12 steps, at most 6 of them using AI, with no cycles
- Review steps and teammate steps that pause the run for a person
- Every run recorded step by step on its own page
What it does not do
- An agent cannot give itself a tool; tools are fixed by what it produces
- Agents do not send email or make payments
- Agents act with the permissions of the person who started the run, never more
Questions about AI Agents
What are AI Agents in kanês?
A digital team inside the workspace: built-in Operations, Sales and Strategy specialists plus agents your company defines. They follow workflows you draw and stop for a person wherever review is needed.
Can kanês AI Agents act without approval?
They do internal work such as writing a deliverable, adding tasks or drafting a proposal. A review step pauses the run for a person, nothing is sent outside the workspace, and agents never have more access than the person who started the run.
How is a kanês agent different from a chatbot?
A chatbot answers. A kanês agent works on your records through checked tools, inside a workflow with steps, limits and approvals, and leaves a record of every step.
Start with the list
you already have.
Import a spreadsheet. It gets researched, scored and ordered before you open it.