We started with 3,000 signal-active companies and narrowed them to 39 Tier 1 leads in under 20 minutes using Claude, Clearcue MCP, and a 4-step qualification workflow. Here is the exact process with prompts you can copy.
We ran a live workflow during our April 2026 webinar that took 3,000 raw signal-active companies and narrowed them to 39 Tier 1 leads ready for outreach, all in under 20 minutes using Claude with Clearcue's MCP. The secret was not better outreach copy or a bigger database. It was qualifying based on real-time buying signals rather than static firmographic filters.
This article walks through the exact 4-step process we use, with the prompts you can copy and customize for your own business. The workflow combines Claude's analytical capabilities with Clearcue's real-time signal data to replace what would otherwise take a full day of manual research.
Most B2B teams qualify leads by filtering a database like Apollo, ZoomInfo, or LinkedIn Sales Navigator: industry, company size, job title, location. This approach tells you who fits your ICP on paper. It says nothing about whether those companies are actively looking for a solution right now.
The result is predictable. You build a list of 1,000 companies that match your criteria and blast cold outreach to all of them through Salesloft, Outreach, or Lemlist. Reply rates sit at 1-3%. Your sales team burns hours on people who have no intention of buying.
Signal-based qualification flips this. Instead of starting with "who matches our profile," you start with "who is showing buying behavior right now" and then filter for ICP fit. The companies that survive both filters are the ones worth your time.
| Approach | Starting Point | Output | Typical Reply Rate |
|---|---|---|---|
| Database filtering | Contact list filtered by firmographics | Large list, unknown intent | 1-3% |
| Signal-based qualification | Companies showing real-time buying signals | Small list, high intent | 30-50% |
The difference in efficiency is dramatic. Signal-qualified outreach reaches people who are already thinking about your category. They engaged with competitor content, downloaded a lead magnet, or interacted with your brand this week. When your message arrives, it lands in context rather than out of nowhere.
This workflow is built for B2B teams who already have signals flowing and need a repeatable process to turn that data into outreach-ready leads:
If you are sending cold outreach to thousands of contacts and getting 1-3% reply rates, this workflow will change how you prospect.
Here is the complete workflow we demonstrated live. Each step builds on the previous one, progressively narrowing your list from thousands of raw companies to a handful of warm, qualified leads with identified decision makers.
What you need before starting:
The entire workflow runs inside a single Claude conversation. Each prompt takes 2-5 minutes to process depending on the volume of data.
The first step retrieves every company that has triggered your tracked signals. This is your raw universe of potentially interested companies.
The prompt:
Can you use Clearcue and give me a list of all the companies that have interactions with any of My Brand signals including Martin and Ralitsa, and any of my Competitor signals.
How to customize: Replace the signal names with whatever you track in Clearcue. Common configurations include brand engagement (your team's LinkedIn activity), competitor engagement (interactions with competitor content), lead magnet downloads, and topic-based signals (people discussing relevant pain points).
What Claude returns: A spreadsheet with every company showing signal activity. In our case, this pulled approximately 3,000 companies. Each row includes the company name, employee count, location, last interaction date, and number of engaged people.
Why this matters: These 3,000 companies are not random. Every single one showed some form of buying-adjacent behavior. Someone at that company engaged with your content, a competitor's content, or both. This starting pool is already warmer than any database export.
The challenge is that 3,000 companies is too many to reach out to meaningfully. Most of them will not match your ICP. That is what the next three steps solve.
Tiering separates your raw signal list into priority groups based on how well each company matches your ideal customer profile. Think of it as applying your ICP filter to the signal-active universe.
The prompt:
Go through every company and assign Tiers using these rules.
Tiering logic:
Tier 1 - B2B company based in the US or the UK, with 10 to 300 employees operating in Finance, HR, or Recruitment.
Tier 2 - B2B company based in the US, the UK, or Europe, with more than 10 employees, operating in Finance, HR, or Recruitment.
Tier 3 - B2B company, based in the US or the UK, any SaaS or AI company.
Output: Build the final spreadsheet with separate sheets per tier.
How to customize: Define your own tiers based on your business. Tier 1 should be your dream ICP where everything aligns perfectly. Tier 2 relaxes one or two criteria. Tier 3 is the broader net. Be specific about geography, company size, industry, and any other firmographic criteria that matter for your sales process.
Our results:
| Tier | Companies | Percentage of Total |
|---|---|---|
| Tier 1 | 39 | 1.3% |
| Tier 2 | 109 | 3.6% |
| Tier 3 | ~850 | ~28% |
| No match | ~2,000 | ~67% |
From 3,000 signal-active companies, only 39 landed in Tier 1. This is normal and expected. Most companies showing signals will not be your dream ICP. A competitor engagement from a 5-person agency in Southeast Asia is still a signal, but if you sell to mid-market finance companies in the US and UK, it is not worth your outreach time.
The tiering step eliminates roughly 95% of your raw list before you spend any time on deeper qualification. This is where the real time savings happen.
Optional: Tag companies in Clearcue
After tiering, you can push the tier assignments back into Clearcue for visual review:
Add a tag (Tier1, Tier2) in Clearcue for Tier 1 and Tier 2 companies.
Tagging lets you go into the Clearcue platform and review signal details, people, and interaction history for each company. The spreadsheet gives you the overview. The platform gives you the depth.
Tiering tells you which companies fit your ICP. Scoring tells you which of those companies are actively showing purchase intent right now. A company can be a perfect Tier 1 fit but have stale signals from three months ago. You want the ones that are hot today.
The prompt:
Score each company 0-100 based on how hot a lead they are right now. 100 = hot, 0 = not interested.
What makes a hot lead: Recent interactions (last 1-2 weeks) with Martin/Ralitsa/Clearcue content, PLUS engagement with lead magnets or competitors. The best leads show multiple signal types (direct engagement + lead magnet + competitor interaction) rather than the same signal repeated many times.
Recency matters most: Fresh signals (≤14 days) score highest. Older signal combinations: Martin/Ralitsa/Clearcue interactions can still matter IF paired with recent competitor or lead magnet activity. Old signals alone = low score.
Weak signals to penalize: Only top voice interactions with nothing else → cap at 15. Only many interactions with a single competitor and nothing else → cap at 20. Only stale (30+ day) signals → cap at 10.
Output per company: Company: [Name] Score: [0-100] Reasoning: [1-2 sentences human-readable text, summarised reasoning]
How to customize: Define what "hot" means for your specific business. What signal combinations indicate genuine buying intent versus noise? Be explicit about caps and penalties for weak signals. This is what separates a useful score from a meaningless number.
Key scoring principles from our experience:
Recency beats volume. Someone who engaged once yesterday is more valuable than someone who engaged 20 times two months ago. Buying windows close quickly. Fresh signals indicate current evaluation, not historical curiosity.
Signal diversity beats signal repetition. A company showing three different signal types (brand engagement + competitor interaction + lead magnet download) is warmer than one showing 50 interactions with a single competitor. Diversity suggests active evaluation across the category. Repetition might just mean someone follows that competitor personally.
Penalize ambiguous patterns. If someone only interacts with top voice content and nothing else, they might just appreciate the content without any buying intent. If someone engages heavily with one competitor only, they might already be a customer or have a personal relationship. Set explicit caps so these patterns do not inflate scores.
Typical score distribution:
| Score Range | Label | Action |
|---|---|---|
| 70-100 | Hot | Outreach immediately |
| 50-69 | Warm | Outreach within the week |
| 30-49 | Interested | Monitor, nurture with content |
| 0-29 | Low intent | Do not outreach, revisit later |
We typically outreach only to companies scoring 50+ combined with Tier 1 or Tier 2 classification. Between tiering and scoring, the original 3,000 companies narrow to a handful of genuinely warm leads.
The final step flips from company-level to person-level qualification. Now that you know which companies are both a good ICP fit and actively showing buying signals, you need to find the right person to contact.
The prompt:
For the hot and warm companies in Tier 1 and Tier 2, tell me the best person to reach out to who might have the best signals. Ideally, they should be in the Sales or GTM department; for smaller companies, it could be the founder/CEO.
How to customize: Replace the department and role criteria with whoever buys your product. For dev tools, look for engineering leads. For HR software, target People Ops. For marketing tools, find the Head of Growth or VP Marketing. Match it to your actual buyer persona.
What Claude returns: For each qualified company, the best contact person based on two criteria: they hold the right role AND they personally show signals. This is the warmest possible intro. You are reaching someone who is a decision maker at a company showing buying intent, and who has personally engaged with relevant content.
Optional: Push to outreach
Add a tag OutreachCampaign in Clearcue for these people.
With tags in Clearcue, you can push qualified leads directly to HeyReach, Lemlist, or your outreach tool of choice. If your outreach tool has an MCP connection too, Claude can handle the entire flow end-to-end: from signal detection to personalized outreach sequence, in one conversation.
The reply rates from signal-qualified outreach are not magic. They follow directly from the qualification logic:
You reach companies already in-market. Every company in your final list showed real buying signals within the past two weeks. They are actively thinking about your category.
You reach the right person. Instead of guessing which contact to email, you identify someone who personally engaged with relevant content and holds a decision-making role.
You have context for personalization. The signal data tells you exactly what the prospect engaged with, when, and on which platform. Your outreach can reference specific behavior rather than generic pain points.
You invest effort where it counts. Instead of spreading thin across 3,000 companies, you focus deeply on 39 Tier 1 leads. More research per lead means better conversations.
| Metric | Cold Database Outreach | Signal-Qualified Outreach |
|---|---|---|
| Starting list | 3,000 contacts | 3,000 signal-active companies |
| After qualification | 3,000 (no filtering) | 39 Tier 1, 109 Tier 2 |
| Reply rate | 1-3% | 30-50% |
| Meetings booked | 5-15 from 3,000 | 15-25 from 148 |
| Time to qualify | 0 (skip straight to outreach) | 20 minutes |
| Rep time per lead | Low (templated) | Higher (personalized) |
The math favors qualification. Twenty minutes of Claude-powered filtering saves hours of wasted outreach and generates more meetings from fewer contacts.
The step-by-step version gives you control and review at each stage. Once you have validated the process and tuned your tiering and scoring rules, you can combine everything into a single prompt and schedule it.
Combined daily prompt:
With Clearcue, give me all the new companies from the past 24 hours. Identify only the top leads for us: SaaS companies with 10-50 people, ideally in Europe or the US. Companies that interact with a competitor or lead magnet are a good signal; if they have a signal with our brand, it's even better. Add a tag NewLeadToReview in Clearcue.
In Claude, go to Schedule Task, create a new task, paste the prompt, and set it to run daily at 9 AM. Your computer needs to be on and Claude open for the task to execute.
The daily version is lighter than the full qualification workflow. Think of it as your morning radar that surfaces new companies worth watching. For the full tiering, scoring, and decision maker identification, run the complete 4-step workflow weekly or bi-weekly.
Running one massive prompt instead of step-by-step. When the context becomes too large, Claude needs to summarize and information gets lost. Run each step separately so you can review outputs at each stage. Combine into a single prompt only after you have validated each step individually.
Relying on a single signal type. One signal is noise 70-80% of the time. Someone might have a personal connection to a competitor, or they might be selling to them. Signal stacking, combining multiple different signal types, is what separates genuine buying intent from random activity.
Skipping company-level qualification. Always qualify the company before the person. A signal from an intern at the right company is still valuable. It tells you the company is in-market. Then you go find the decision maker. Researching individuals before confirming company fit wastes time on people at companies that will never buy.
Ignoring signal recency. A prospect who engaged yesterday is exponentially more valuable than one who engaged two months ago. Buying windows close fast. Weight recent signals heavily in your scoring and deprioritize anything older than 30 days.
Setting scoring thresholds too low. If you outreach to every company scoring above 20, you are back to volume-based prospecting. Hold the line at 50+ for outreach. Companies scoring 30-49 go into a nurture track, not an outreach sequence.
| Tool | Role | Cost |
|---|---|---|
| Clearcue | Signal detection and stacking across LinkedIn, X, Reddit, news, jobs, podcasts, events | From €79/month, unlimited users |
| Claude | Analysis engine that processes signals, assigns tiers, scores leads, identifies decision makers | €100/month (Max subscription) |
| HeyReach (optional) | LinkedIn outreach automation from qualified lead lists | Varies by plan |
| Lemlist (optional) | Email outreach if combining LinkedIn with email sequences | Varies by plan |
Clearcue provides the raw signal intelligence. Claude provides the analytical processing power. The outreach tool of your choice handles delivery. Unlike enrichment-first tools like Clay or PhantomBuster that focus on data appending, this workflow starts with buying behavior and only enriches the leads that pass qualification. Connect them all via MCP and Claude becomes the orchestrator that moves data between tools automatically.
The entire process requires two setups:
Once connected, copy the prompts from each step above, customize the tiering rules and scoring criteria for your business, and run the workflow. The first run will show you exactly how your signal data translates into qualified leads.
For the full prompt library with advanced workflows including engagement dashboards, meeting preparation briefs, conference planning, and champion identification, visit our prompt library.
Start using Clearcue today and never miss a buying signal again.