Fuel Cycle Research Harness is designed to be implanted directly into a research organisation’s environment with workflows designed around the existing insight function, studies that the company has already run and the data they own.

Fuel Cycle said that Research Harness would examine what AI can be reliably asked to do in the organisation and what’s safe to bring in front of leadership, and will keep AI agents within the organisation connected and operating from within the same context.

In practice, Fuel Cycle said that this meant that new studies build on previous work, with context retained between handoffs and findings able to be traced back to their source. 

The framework is opening early access for select design partners, who will get direct input into how the product develops.

Daryush Laqab, chief product and AI officer at Fuel Cycle, said: “Every vendor in this category has agents now. That was never the hard part.

“The hard part is what happens between the agents – making sure a survey-design agent and a data-quality agent and an analysis agent are operating off the same brief, the same audience rules and the same brand guardrails, and that a research director can see exactly what each one did and why.”