What types of agents does DataRobot offer for enterprise use?

DataRobot offers four agent types for enterprise use: foundational agents, business agents, co-developed agents for SAP, and purpose-built agents delivered as a managed service. Each type targets a different starting point — from teams that want to build their own agents to organizations that want DataRobot to build and run them. The right choice depends on how much control, customization, and delivery responsibility you want to keep in-house.

The four agent types at a glance

Foundational agents

Foundational agents are positioned as the easiest starting point: agents you can stand up quickly and infuse into your organization's specific processes and systems. The emphasis is on low setup friction rather than deep customization from day one.

This type fits teams that already know which internal processes they want to augment but don't want to build an agent stack from scratch. You get a working agent and adapt it to your environment, rather than designing the architecture yourself.

Business agents

Business agents bundle integrations, data models, and UIs so you can build and integrate your own agents into business tools — SAP is the named example. Here the agent isn't a standalone assistant; it's a component that plugs into systems your teams already use.

Choose this type when the value of the agent depends on reaching into existing enterprise software and data, and when your team wants to own the build while relying on provided integrations and interfaces.

Co-developed for SAP

DataRobot describes itself as the exclusive agentic AI partner fully certified to run inside the SAP ecosystem. This is a distinct category from general business agents: certification and co-development mean the agents are designed to operate within SAP's environment rather than alongside it.

If your enterprise runs core processes on SAP and requires certified in-ecosystem operation, this is the relevant option. If SAP is only one of many tools, business agents may cover the same ground with less constraint.

Purpose-built agents

Purpose-built agents are service-led: DataRobot builds and manages them for you. The stated goal is to help you quickly create a full agentic workforce without carrying the build and operations burden internally.

This suits organizations that want outcomes on a short timeline and are willing to delegate design, delivery, and management. The trade-off is less direct control over how each agent is constructed compared with the build-it-yourself paths above.

How to choose

• Want to build and customize: start with foundational agents, then move to business agents when you need integration into tools like SAP.
• Run on SAP and need certified in-ecosystem operation: the co-developed SAP offering is the specific fit.
• Want DataRobot to build and manage agents for you: purpose-built agents are the service-led route.

These types sit on a shared Agent Workforce Platform, which DataRobot describes as covering the build, operate, and govern phases of the agent lifecycle — including deployment to edge, cloud, or on-prem, real-time quality monitoring, and lifecycle tracking with enforceable controls. That means the choice of agent type doesn't lock you out of the surrounding operational and governance capabilities.