Research
Generative & Agentic AI
Generative AI creates content and answers. Agentic AI goes further and takes action. This area covers the models, assistants, agents and retrieval systems that enterprises are putting into production.
Topics in Generative & Agentic AI
Foundation Models & LLMs
The large models behind generative AI and how to choose among them.
Overview
Large language models and other foundation models are now available through cloud APIs, embedded in business software and as open-weight models that organizations can run themselves. Copilots and assistants bring them into daily work. Agents connect them to tools so they can complete multi-step tasks.
Value depends on grounding these systems in accurate, permitted information and on controlling what they are allowed to do. Retrieval-augmented generation (RAG) and well-governed knowledge sources are often more important than the choice of model.
What leaders need to get right
- Model strategy. Match models to tasks on capability, cost, latency and data terms.
- Grounding. Connect AI to curated, access-controlled knowledge.
- Action controls. Limit what agents can do and require approval for high-impact steps.
- Measurement. Track quality, adoption and cost from the first deployment.
Questions leaders should ask
- Which models are approved for which kinds of data?
- How do we ground AI answers in approved sources?
- What actions can our AI agents take without a human?
- How do we measure answer quality over time?
Other research areas
Related research reports
Reports, ebooks and guides connected to this topic.
Latest insights on Generative & Agentic AI
New articles in this area are on the way. In the meantime, explore the Research library.