Qualify leads against your ICP
A raw list of leads is mostly noise: out of fifty posts that mention your keyword, maybe five are worth your time. With this skill, your AI agent runs every lead through Buska's score_lead tool, then sorts the pile into hot, nurture and ignore, each with a written reason. You read the top, skip the rest.
get_signalsscore_lead1Before you start: a lead list and the MCP connection
A Buska account (any plan, trial included) with the MCP server connected to your agent via OAuth (setup: buska.io/docs/mcp-server/setup). You also need leads to sort: your Buska signals, or any list you paste into the conversation.
2Gather the pile
Your leads can come from anywhere: a list you paste into the chat, a CSV export from your CRM, or Buska itself. If you want Buska's own findings, the agent calls get_signals with a minimum score and a date range to pull the buying signals the engine already detected for your configured keywords, complete with AI score and reasoning.
3Score every lead
For each lead in the pile, the agent calls score_lead. Buska rates buying intent and fit with your ICP on a 0-100 scale and explains why. This is the step that turns a flat list into something sortable: a founder complaining about your competitor's pricing scores very differently from a student mentioning the category in passing.
4Sort into three buckets
The agent splits the results into hot (worth a reply today), nurture (right profile, weak timing) and ignore. Each hot lead keeps its score, its post URL and a one-line reason, so you can sanity-check the call in seconds instead of re-reading the whole post.
5The three-bucket verdict
A short ranked list: the hot leads first with author, platform, link and reasoning, a compact nurture table below, and a simple count of what was ignored. From there you draft replies with Reply Studio, move leads through your kanban, or sync a column to Google Sheets. Nothing is sent or posted without you.
The prompt to paste into your agent
Copy it into Claude, ChatGPT or any MCP client connected to Buska.
Qualify these leads against my ICP: B2B SaaS founders and heads of growth at companies under 50 people.
If I pasted leads below, use those. Otherwise call get_signals with minScore 5, since the last 7 days, limit 30.
For every lead, call score_lead to rate buying intent and ICP fit.
Sort them into three buckets: HOT (worth a reply today), NURTURE (right profile, wrong timing), IGNORE.
For each HOT lead give me the author, platform, post URL, score, and a one-line reason.
Show NURTURE as a short table and just give me a count for IGNORE.
[paste your leads here, or delete this line to use my Buska signals]Related documentation
Frequently asked questions
Will the agent reply to leads for me?
No. Buska never posts to social networks on your behalf, and this skill does not change that. Reply Studio drafts replies inside the app and you send them yourself. On the Scale plan, the AI SDR agent can send email outreach from your own inbox, but only after you approve it.
Do my leads have to come from Buska?
No. score_lead works on anything you feed it: pasted posts, a CRM export, links you collected by hand. get_signals is simply the fastest source, because Buska has already filtered and scored those signals against your configured keywords across 30+ platforms.
Can I trust the scores?
Treat them as triage, not truth. Every score comes with written reasoning, so you can spot-check the hot bucket in a minute instead of reading fifty posts. Quality also depends on how clearly your ICP and signals are set up in Buska: vague inputs produce vague scores.