We're designing the next iteration of our Agent Studio, what's the one feature you'd want most?
by•
Building toward the next major release. The team has internal priorities, but I'd rather hear from people actually shipping agents.
The shortlist we're considering:
- Visual flow editor for multi-agent orchestration
- Real-time cost forecasting before agent runs
- Per-agent observability dashboard
- Native A/B testing across LLM models in production
- Approval gate templates by industry (BFSI, healthcare)
Which one would actually unblock your team? Or is there something not on this list that would matter more?
Feel free to share your opinion below
31 views
Replies
I'd probably prioritize per-agent observability first. Once agents move into production, debugging failures, understanding why an agent made a decision, tracking latency, token usage, and tool calls become much more valuable than adding new capabilities.
One feature I'd also love to see is conversation replay with full execution traces—being able to step through an agent's reasoning, tool invocations, and state changes would make debugging and optimization much easier, especially for multi-agent workflows.
Curious—are you targeting technical teams building custom agents, or are you aiming for less technical users as well?
@kartikbatchu2003Â Thanks Kartik, this is exactly the kind of feedback I was hoping to get. The more agents we've put into production, the more I've realized observability becomes the product. New capabilities are exciting, but when something breaks at 2 AM, nobody cares about the feature roadmap, they care about understanding what happened and how to fix it fast.
I also love the idea of conversation replay with full execution traces. We've had a few debugging sessions where seeing the final output wasn't enough, we needed to know which tool was called, what state changed, and why the agent chose that path in the first place. That's especially true once multiple agents are collaborating.
To your question, we're trying to serve both audiences, but not with the same experience. Less technical users need a simple "build and deploy" path, while technical teams usually want visibility, controls, and debugging tools like the ones you mentioned. Balancing those without overwhelming either side is probably the hardest design challenge we're working through right now.
@nolan_vu That's a thoughtful approach. I like the idea of separate experiences rather than trying to force one UI for everyone.
One thing I've noticed while building ChatHop is that once people start relying on AI agents daily, the biggest challenge isn't generating responses—it's finding and revisiting the right conversation later. Search, navigation, and context retrieval become surprisingly important parts of the workflow.
If you end up building conversation replay, it could pair really well with execution traces. Being able to jump to a specific tool call, decision point, or state transition instead of scrubbing through an entire session would make debugging much faster.
Looking forward to seeing where Agent Studio goes!
@kartikbatchu2003Â I love the idea of making replay more "jump-to-state" than "scrub-through-everything." If someone is debugging, they usually care about the exact tool call, state transition, or decision point that caused the outcome, not the entire conversation history. That's a really useful framing for how we think about execution traces going forward.
Really appreciate the support and the upvote. Hoping ChatHop keeps growing and solving real problems for users too. Already following along and looking forward to seeing where you take it next. If you ever get a chance, I'd love your thoughts on AI Hive as well. Any review, follow, or upvote means a lot to our team: AI Hive on Product Hunt.
@nolan_vu Thanks, Nolan! I really appreciate that.
I just followed AI Hive and gave it an upvote. Wishing you and the team all the best with Agent Studio—it looks like you're building in the right direction.
If you get a chance, I'd love for you to check out ChatHop as well. It's a browser extension I built to help people navigate and search long AI conversations across ChatGPT, Claude, and Gemini. If you find it useful, an upvote or any feedback would mean a lot.
Always happy to exchange ideas as both products evolve.