- Model: Basic AI systems that interpret signals, generate responses, and make predictions
- Tools: The integration layer that connects AI to enterprise systems, such as APIs, protocols, and connectors
- Context: Before making decisions, information agents need to understand the complete business picture, including customer history, product catalogs, and supply chain networks.
- Governance: Policies, controls and processes that ensure data quality, security and compliance
This framework helps assess where reliability gaps emerge. When an enterprise agent fails, which quadrant is the problem? Is the model intended to be misunderstood? Are tools unavailable or broken? Is the context incomplete or inconsistent? Or is there no mechanism to verify what the agent did?
Why is this a data problem, not a model problem?
The temptation is to think that the model will improve. Still, the model’s potential is growing rapidly. Evaluation is the price Dropped about 900 times In three years, Delusion rates are decliningand AI’s ability to perform longer tasks Doubles every six months.
Tooling is also getting faster. Integration frameworks such as the Model Context Protocol (MCP) make it dramatically easier to connect agents to enterprise systems and APIs.
If the models are powerful and the tools are mature, then what is the point of adoption?
To borrow from James Carville, “It’s data, stupid.” The root cause of most abusive agents is incorrect, inconsistent, or incomplete data.
Businesses have accumulated data debt for decades. Acquisitions, custom systems, departmental tools, and shadows have left data scattered across silos that rarely agree. The support system is not identical to what is in the marketing system. Supplier data is replicated in Finance, Purchasing and Logistics. Places have multiple representations depending on the source.
Drop some agents into this environment, and they’ll perform admirably at first, because each one is given a ready-made set of systems to call upon. Add more agents and the rift widens, as each carves out their own slice of the right.
This dynamic has ended before. When business intelligence became self-service, everyone started building dashboards. Productivity increased, reports failed to match. Now imagine this trend not in static dashboards, but in AI agents that can take action. With agents, data inconsistency creates real business results, not just arguments between departments.