← back
What the Best Agents Share — Mardu Swanepoel, Flinn AI
Takeaway
Strong agents combine constrained modes, visible execution, personalization, and undo controls to support collaboration.
Summary
- Focus modes constrain tools and prompts while setting user expectations; Cursor's planning and debugging modes exemplify this pattern.
- Transparent execution exposes progress, context, and tool activity so users can intervene early and trust results.
- Personalization through Harvey playbooks, memory, skills, and connectors improves understanding of how a user wants work performed.
- Reversibility bounds the cost of mistakes through line-level acceptance, file rollback, and conversation checkpoints.
agentsagent-uxhuman-in-the-loop
Original description
Harvey, Cursor, Manus, and Claude operate in completely different domains but share four patterns: focus modes that constrain the action space to improve output quality, transparent execution that surfaces tool calls and reasoning to build user trust, personalization that optimizes for speed to understanding rather than just speed to output, and reversibility that bounds the downside of mistakes so users take on higher value tasks. Mardu Swanepoel from Flinn AI breaks down how each company puts these into practice. Cursor lets you roll back changes at the line, file, or conversation level and run multiple model outputs in parallel from the same input. Harvey builds playbooks from a firm's legal methods so the agent works the way the firm would. Claude surfaces a live task list alongside every tool call's inputs and outputs so users can intervene before the agent goes further in the wrong direction. Speaker info: / mardu-swanepoel-000