Getting started

Introduction to Dezifi

Dezifi turns AI demos into governed, observable, reliable execution. Build agents, orchestrate them in multi-step workflows, and connect them to the enterprise systems your teams already use — all behind a single control plane.

What you'll learn
  • What an AI agent is and how Dezifi structures one
  • How agents, workflows, tools, and guardrails fit together
  • Which teams Dezifi is built for
  • How to ship your first production agent

What Dezifi gives you

Six capabilities ship together as one platform, so you don't glue together a stack from scratch.
  1. 1

    Agent Orchestration

    Design structured workflows with tools, conditions, approvals and multi-step execution on a visual canvas.
  2. 2

    Governance & Policies

    Define what agents can access, what they can do, and when a human must approve — at workspace, agent, or workflow scope.
  3. 3

    Runtime Guardrails

    Inspect every prompt, tool call and response. Block jailbreaks, PII leaks, and policy violations in real time.
  4. 4

    Observability

    Every run is fully traced. See LLM calls, tool spans, latency, cost, errors, and reviewer decisions for every execution.
  5. 5

    Evaluation

    Compare agent versions, score against golden datasets, run A/B tests, and detect quality drift before it hits production.
  6. 6

    Integrations

    Connect agents to CRMs, support systems, observability tools, source control and chat. The beta includes 50 integrations; add your own API as a custom tool.

Who Dezifi is for

Enterprises that need to move past AI proofs of concept. Common teams: operations automating handoffs across systems, engineering running PR review and on-call triage, support classifying tickets and drafting responses, sales ops updating Salesforce and qualifying leads.

How the platform fits together

An agent is your unit of intelligence — one LLM, one set of tools, one policy. A workflow orchestrates multiple agents with branching, loops, and human approvals. Both consume tools (integrations), retrieve from knowledge bases (RAG), are constrained by policies and guardrails, and emit traces to Monitor, Eval, and Analytics.

Frequently asked questions

Do I need to be technical to use Dezifi?
No. The visual agent builder and workflow designer let business and ops users assemble agents without code. Developers can extend via the SDK and add custom tools.
Can Dezifi run on-premise or in our cloud?
Today Dezifi runs as a hosted service, set up for you by the Dezifi team, with each organization isolated in its own database. Running inside your own cloud or on-premise is on the roadmap and is not available yet. Talk to us if that is a requirement.
Which LLMs does Dezifi support?
OpenAI, Anthropic, Google, AWS Bedrock, Azure OpenAI and a number of other hosted providers, plus self-hosted models through an OpenAI-compatible endpoint. You add your own provider key, can configure multiple providers per workspace, and pick the model per agent.