Finding one client is easy. Finding 200, for what you sell, is not.
Today an expert in sales, marketing or data spends the week reading hundreds of pages and sources, sorting them, connecting the dots. LeadRadar does that work automatically, to your criteria. You paste your website or import your catalogue; we work out who to target, read the market, score every company on evidence and hand you the contact, the reason to call and the first draft. You review. You decide. You sell.
Researching one company is a task. Researching a hundred is a full-time job.
Any experienced salesperson can assess one company. At a hundred, the method breaks: tenders, job postings, news and annual reports for each, judged by experience, ranked by hand. About 25 minutes per company, 83 hours a month for two hundred, and the first signals are already stale when the list is finished. LeadRadar runs the same chain on every company, every day, and shows the reasoning behind each one.
Hundreds of sources in. A ranked list of companies, each with evidence and a reason to call, out.
Start from your website
Paste the address of your service page or import your catalogue. LeadRadar reads what you sell and proposes the ideal customer profile, the questions worth asking and the exclusions. You edit; nothing is code.
We read the market for you
Tenders, registries, newsrooms, annual reports, press and job boards, hundreds of pages a day. Every signal becomes a claim with its exact quote, source and date. If the text does not say it, we do not claim it.
We rank on evidence, in the open
Our own algorithm weighs fit, recency and relevance, subtracts explicit penalties, and gives every company a priority you can reconstruct line by line. It learns from each of your accept and reject decisions, as versions you approve.
You get the contact and the first draft
The right role at the company, the reason to call, and an email or call script written from the facts. You approve, edit or reject. One click creates the CRM record. Nothing leaves without you.
Tell us what you sell. We work out who needs it.
One address is enough to start. From your service page LeadRadar drafts the ideal customer, the signal questions and the disqualifiers. Change a weight, add a question in plain language, and the simulator shows how the market re-orders before you publish. A new product is a new template, not a new project.
ICP criteria
Market, industry, company size, geography, operational complexity. Mandatory criteria gate; preferred ones add fit.
Signal questions per service
"Is the company hiring RPA developers or process-excellence roles?" — high, medium or low importance, positive or negative polarity.
Negative rules
Penalties (a certified internal SOC: −15) or gates (a competitor bidding on the same tenders: excluded).
Scoring logic
Weights, half-lives and the 0.35 / 0.65 balance are per-service settings, saved as reviewed versions, never applied silently.
Signal Builder · SCUT NIS2
formula v3Eight templates ship with the MVP: NIS2, MDR, cloud, connectivity, IoT, analytics, IT consultancy and Intelligent Automation. Adding a service is configuration, not code.
The algorithm does the reading and the maths. You decide.
Weighted match on confirmed criteria. Unknowns earn nothing and cost nothing; the denominator is never shrunk to the data found.
Per rule: weight × quality × recency, with d = 2^(−age/half-life). Quality comes from a versioned rubric, not a free-text probability.
Only configured negative rules. Missing information never penalises; a failed mandatory condition is a gate, not a penalty.
Buying stage, momentum and a 45–90-day opportunity window from signal sequences, labelled as heuristic estimates until calibrated on outcomes.
Our own scoring algorithm processes every company the same way: the AI reads the sources and explains what it found, the versioned formula turns that into a priority, and every contribution stays visible on screen. Nothing is decided for you. You see the evidence, you approve, edit or reject, and each of your decisions proposes a better weight for the next version, which you approve as well. Swap the language model and the score does not move; that is what makes it yours to trust or contest.
Danubia Energy SA (fictional)
…Information Security Officer responsabil de conformitatea NIS2 și ISO 27001…
…două locații noi până în 2027.
Danubia Energy SA
We noticed from your NIS2 self-assessment announcement and the information-security officer opening that NIS2 readiness is on your agenda this quarter. Would a 30-minute clarification discussion on NIS2 scope be useful?
Every highlighted phrase links to its evidence card. Drafts quote only the supplier's public proof points, never state that a company is legally obliged, and are reviewed by a person before they reach the CRM.
The contact, the reason, the first sentence. Ready before your coffee.
For every prioritised company we hand you the role to talk to, the evidence that makes now the right moment, and a draft written from those facts, for email or for a call. Approve it and the CRM record is created once, never twice. Reject it with a reason and the algorithm learns. Sales sells; marketing runs the campaign; nobody reads two hundred websites again.
Drafts written from facts. Approved by you. Sent by you.
Intelligence is not the same thing as permission to communicate. LeadRadar may identify, enrich, personalise and recommend before it sends; a governance layer sits between the signal and the message, and every step below is enforced in the pipeline, not left to policy.
Why this is "friendly" rather than "compliant by declaration": B2B account prioritisation is not a listed high-risk use under the AI Act, scores produce no automated decision with legal effect on a person, and AI involvement is labelled and documented; on the GDPR side the personal data is minimised to name, role and professional address with a documented balancing test and an objection path that works from the first message. We describe these controls as we operate them, and verify hosting, sub-processors and retention with each customer in the pilot.
The big platforms find companies. We find the reason, for your product, in your market.
6sense, Bombora, Demandbase, Apollo, Clay and Common Room prove the category. They do not read SEAP, MTender or Termene.ro, and they rarely map a signal to the exact service in your catalogue with a formula you can inspect. Their claims are theirs; ours are hypotheses with a test attached.
| Dimension | Category pattern | LeadRadar |
|---|---|---|
| Where intent comes from | De-anonymised web visits, publisher co-ops, keyword research | Public, citable events: tenders, hiring, announcements, incidents, reports. Weaker on anonymous research, stronger on verifiable evidence. |
| Level of identity | Person-level contacts at the core | Company-level by design; buying roles recommended as roles. Persons only with a documented legal basis. |
| Scoring | Black-box predictive scores and buying stages | Deterministic, versioned formula with every contribution visible; stages labelled as estimates until calibrated. |
| Product mapping | Generic scoring; mapping left to the customer | Service-specific templates: a signal maps to an exact catalogue item with its own ICP, rules and playbook. |
| Internal context | CRM only | CRM plus accounting and invoices: existing customer and purchased services change the action. |
| Geography | US/UK-centric data, Europe via partners | Romanian and Moldovan registries, procurement portals, press and job boards as first-class sources; Romanian and Russian extraction. |
| Execution | Sequences, dialers and ads that send at scale | Intelligence and decision layer with a human gate. Drafts, never sending; complementary to sequencing tools. |
| Price point | Enterprise contracts, annual commitments | 500–1,500 EUR per month test range and a four-week pilot, accessible to mid-size IT service providers. |
Where they are better and we say so: breadth and freshness of contact data, anonymous-research intent at global scale, years of outcome data, mature integrations, agentic execution.
Built against a real brief, judged on real criteria.
Orange Systems asked for a platform where a sales team defines its own questions, monitors public sources, scores companies and prioritises on evidence. The blog maps every requirement and every judging criterion to the screen that answers it.
How LeadRadar answers the Orange Systems brief
Configurable ICP and signal questions, data ingestion tiers, explainable scoring, the optional outreach module, React + PostgreSQL + HubSpot, and the 25 / 20 / 20 / 15 / 10 / 10 criteria with evidence for each.
Read the breakdownFour weeks. One market. One service. One hundred accounts. Then decide.
Put it next to the alternative. One analyst, one company at a time, about 83 hours a month to cover two hundred accounts, with the first ones already out of date by the end. LeadRadar covers the same two hundred every day, for a monthly fee in the range of 500 to 1,500 EUR, and hands your team back those hours for calls. We agree the numbers before we start: what counts as a good recommendation, who judges it, and the precision we must reach. If we miss, you keep the evidence and walk away.
One analyst, one company at a time
Every company, every day, for one monthly fee
Pilot · 4 weeks
- Week 1 scope, catalogue, identities, baseline
- Week 2 first batch researched and reviewed
- Week 3 repeated use, approved CRM actions
- Week 4 held-out evaluation and decision
Starter
- One portfolio, up to 100 accounts
- Scoring, predictions, dashboard
- Explicit research credits
Growth
- Several services, alerts, CRM
- Rules, simulator, review workflow
- Accounts and users by contract
Enterprise
- Approved sources, API, integrations
- Audit and extended governance
- Private deployment priced separately
Time figures are the team's working hypothesis (25 → 8 minutes per company) and are measured in the pilot; the 2.50 EUR per account is the illustrative delivery cost from the whitepaper, not a price. Test range 500–1,500 EUR per month per organisation, central scenario 1,000 EUR; tiers not yet allocated. Results-as-a-Service packs ("NIS2 Opportunity Pack": 50 accounts with evidence, scores and recommendations) are available as a verified deliverable, without a guarantee of meetings or revenue.
Knowing a company is ready is not the same as permission to write to it.
So LeadRadar identifies, enriches, personalises and recommends, and then stops. A governance layer sits between the signal and the message, a person approves every action, and the data stays in Europe.
Privacy by design
Company-level facts by default. Where a professional contact is used, the record carries its source, lawful basis (a documented Legitimate Interest Assessment), retention and suppression status. Objections are honoured globally, before any draft is generated.
Advisory, not automated decisions
B2B account prioritisation is not a listed high-risk use. Scores stay advisory; human review is mandatory in the architecture, not just in policy. AI involvement is labelled in the interface and documented (model, data, limits, oversight).
Data stays in the EU
Target environment: Hetzner, Falkenstein (Germany), with a self-hosted, EU-resident language model option so no CRM or accounting data leaves the EU. Region, sub-processors and transfers are confirmed contractually and technically in the pilot.
Security programme
Tenant isolation with row-level security, purpose-based access, audit trail per decision, token rotation and revocation, incident response with 72-hour authority notification. Certification is under way, not claimed.