Clearing chat history helps AI agents find better solutions
Key points
- Microsoft Research found long-running agents stall in narrow solution regions, with eight runs ending between 68.3 and 100.0.
- Fork-and-Flush splits agents into parallel branches and wipes chat context while keeping workspace files intact.
- On 13 tasks with equal compute, the method scored 0.78 normalized versus 0.47 for single runs, a reported 66 percent gain.
- Simply clearing chat context without branching improved results by 22 percent, according to the paper.
Why it matters for marketers: Teams running long agent workflows may get better results by periodically resetting conversation context while preserving saved files, rather than adding more compute.
The DecoderAgents & automation