Signal Agents

Today, you tell SensorHub what to watch with keywords (or just let Sensorhub find the most relevant signals), then sift the shared feed yourself.

Signal Agents will make it easier to find the right signals based on your preferences. You describe what you're looking for in plain language, and an agent continuously hunts for matching signals on your behalf. Same underlying idea as keyword-based signals, but far more powerful and flexible.

What’s Included:

• Natural-language Signal Agents – give an agent custom instructions with a prompt ("find people frustrated with X migrating off competitor Y", "QA discussions about flaky tests in tool Z", "buyers evaluating AEO tooling"). The agent keeps searching for matching signals; no keyword tuning required. The prompt can be anything: a topic, a persona, a pain point, a buying intent, a competitor angle.

• Per-agent relevance scoring - each agent scores how relevant a signal is to its own instructions, not just to your core product. A signal that's only loosely related to the product overall can still rank highly for a specific agent. Specialists see what matters through their lens.

• Filter integration – signals surfaced by agents show up in a dedicated Found by agent filter section, so agent-discovered signals stay distinct and easy to review alongside your existing filters.

• Template gallery - a curated library of ready-made agent prompts by use case (e.g., AEO Signal Agents, Sales Signal Agents, and more) so you can spin up a high-quality agent in one click and customize from there.

Who Can Use This?

Anyone who wants hyper-personalized signals. Route different agents to different people or goals. E.g., a QA lead, a software architect, a sales rep, a founder tracking competitors, and let each one engage where they add the most value, without manually triaging one shared feed.