Questions we get before every project.
How AI agents work, what they connect to, how they stay controlled, and what a deployment actually involves.
What exactly is an AI agent?
An AI agent is a system that takes a goal, reasons about the steps required, calls tools and APIs to perform them, and reports the result. Unlike a chatbot, it acts inside your systems rather than only answering questions.
How is this different from classic automation?
Classic automation follows fixed rules. Agents handle the parts that need judgment — reading unstructured documents, classifying requests, drafting responses — while deterministic steps still run as rules. Most Lumizy workflows combine both.
Which tools can Lumizy connect to?
Anything with an API or a supported connector: CRM, ticketing, ERP, storage, messaging, security tooling and internal services. If a system has no API, we look at file-based or database-level integration.
How do you keep AI execution under control?
Every agent gets scoped permissions, high-impact actions route to a human for approval, and inputs, tool calls and outputs are logged for review. Who can build, approve and run agents is defined by role.
Where is our data processed?
Agents only access the systems and records their workflow requires. Data-handling, retention and model-provider choices are agreed per engagement before anything is deployed.
How long does an implementation take?
A first scoped workflow typically moves from discovery to a supervised production run in a matter of weeks, depending on integration complexity and approval processes on your side.
How does pricing work?
Engagements are scoped per project: an initial workflow, then optional ongoing support and expansion. Pricing depends on scope, integrations and oversight requirements — the AI Audit is free and comes with an indicative scope.
What happens after deployment?
We measure execution volume, handling time and outcomes, tune the agent, and decide together which workflow is next. Agents can be paused or rolled back at any time.