Common questions
about engagements.
What an engagement includes, who owns the agent, how privacy works, what systems connect, how long it takes to launch, and how DeployCo differs from giving your team an AI chat subscription.
What does an engagement actually include?
Each engagement is the design, build, and ongoing operation of one or more agents purpose-built for your business. We define the agent's job in a working session with you, configure its goals and the tools it can use, encode the rules it must operate under, and operate it for you. Agents run on DeployCo managed infrastructure, or are deployed into your environment as part of an engagement, to your specification. You see consequential actions in an approval queue. We handle monitoring, drift, and improvements.
What if the agent produces something wrong?
Consequential actions are approval-gated by default. The agent proposes the action, whether a reply, a reorder, a refund, or a payment, and you approve, adjust, or reject before anything executes. An agent earns more autonomy only through verified results, within limits you set, and a kill switch returns it to fully supervised operation at once. Every decision is recorded, and a wrong proposal is a non-event, because nothing consequential executes until you say so.
Who owns the agent you build for us?
You do. The configuration, rules, reference material, and any institutional knowledge encoded into the agent are yours. The runtime is what we operate. If you ever leave, we provide the agent's configuration and policy documents so a future team can rebuild it.
How does data privacy work?
Your data is segregated by organization and access-controlled at the database level, enforced on every query. Approvals, audit history, and any business data the agent touches are visible only to your team. We do not train models on your data, and your data is never shared across customers. Specific privacy and data-handling questions are answered in the discovery call.
What systems can an agent connect to?
An agent connects to the systems where your work already happens: email, CRM, support tools, and internal APIs. If your stack has an API, we connect to it; if it does not, we build a webhook bridge. We build the specific integrations your engagement needs as part of the build.
How long until we see something working?
Most engagements have a working agent proposing actions for your review within one to three weeks of kickoff. The first week is design and configuration. The second week is integration with your systems and policy encoding. The third week is private testing and refinement before you start approving real actions.
Can we cancel?
Engagements are monthly. You can cancel any month with 30 days' notice. There is no long-term commitment, no enterprise contract minimum, and no termination fee. The engagement continues because the agent earns its place, not because a contract requires it.
How is this different from giving our team a ChatGPT subscription?
An AI chat tool helps an individual employee one prompt at a time: they ask, it answers, they act. An agent does the work itself, on its trigger or schedule: it watches for the trigger, does the task against the rules of your business, and takes the action in the system where the work lands, with audit logging and an approval gate on consequential steps. Different category. We do not replace ChatGPT for your team; we replace the operational work your team would otherwise do by hand.
A question we haven't answered?
Start a conversation