01 - Upload
Bring your export in. The system recognises common formats, maps the columns for you, and deduplicates against every existing campaign.
Every step from a raw list to reviewed, signal-led outbound.
Prospectlab takes one prospect export and walks it through the whole pipeline. Uploaded clean, enriched in the background, scored on signals, segmented against your real ICP, composed with Kat, assembled deterministically, and held in a review queue until you approve it.
START FREE TRIALOne list. Eight stages. You approve the output /
Upload once and walk away. Everything between import and review runs in the background.
How it works
02 - Enrich
Discover and verify every email; scrape public company surfaces for signal-worthy facts. A failed lookup is never billed.
03 - Signal
Score and rank 23+ signals per prospect, weighted by impact, freshness, rarity, and confidence.
04 - ICP
Match prospects against an ICP derived from your real customers, not a guess in a slide.
05 - Compose
Co-author the gold-standard email with Kat, your AI campaign strategist. You approve every paragraph before generation.
06 - Pattern
Decompose the approved email into reusable rotation pools (openings, body sentences, CTAs).
07 - Generate
Assemble a unique sequence per prospect, deterministically. No LLM in the send path.
08 - Review
Approve, edit, or discard at the email level. Push to Smartlead or Instantly when you're ready.
Feature 01/
Land any list clean, before a single email is written /
A guided upload flow turns any prospect export into a structured, deduplicated, campaign-ready segment. The cleaner the input, the better every step below performs.
Most outbound tools start with "send." Prospectlab starts with "structure."
The system recognises common export formats, maps the columns for you, and checks every prospect against your existing campaigns so the same person is never approached twice. When the import finishes, you get a receipt - exactly what came in, what was skipped, and why.
- Six source types: Apollo CSV, LinkedIn Sales Navigator, Clay, CRM exports (HubSpot, Salesforce, Pipedrive), generic CSV, and JSON.
- Automatic column mapping. The system proposes; you confirm or override.
- Cross-campaign deduplication on every upload.
- Import receipt: imported, skipped, and sent to background processing.
Feature 02/
Enrichment that fills the gaps and reads the company for you /
Once a prospect is uploaded, the system finds missing emails, verifies every address, and scrapes the company's public surfaces for signal-worthy facts.
An SDR spends about twenty minutes per prospect finding an address, checking it bounces, and reading the company's site. Prospectlab does that in the background, for every prospect, while you work on something else.
No copy-paste from LinkedIn. No "did this one bounce?" guessing. Every prospect moves through a transparent status flow you can watch, and only verified-deliverable prospects continue to composition.
- Email discovery - a three-provider waterfall proposes addresses for prospects without one.
- Email verification - every address classified valid, invalid, or catch-all; only valid prospects move forward.
- Company website scraping - heritage, case studies, hiring signals, recent news, product information, leadership signals.
- Transparent status pipeline: pending discovery, discovery, pending verification, verification, valid, enrichment, ready.
A failed lookup is never billed.
Feature 03/
It doesn't just enrich. It ranks the signal and matches it to the prospect /
23+ scored signals in production today, weighted by impact, freshness, rarity, and confidence. So the right fact opens the right email for the right prospect.
The if / then / else cascade
If - a recent funding round
The email opens with the funding round and what post-funding teams in this sector usually deal with next.
Else if - a strong case-study match
The email opens with the case study and the parallel problem it covered. Specific names, specific outcomes.
Else - a clean, honest generic
From your approved fallback pool, clearly flagged. Never padded, never guessed. You can see which path fired for each prospect in the review queue.
Three personalisation tiers, one cascade /
You always know which tier each prospect is in, before anything is generated. If a prospect previously worked at one of your existing customers, the opening references the shared connection automatically.
High signal
Proof-led emails anchored to specific facts. Funding, hiring, leadership change, case-study match, warm introduction.
Medium signal
Targeted on partial signals. Sector, technology stack, growth stage, geography.
Low signal
A clean, honest generic from your approved fallback pool. Clearly flagged, never padded, never guessed.
Feature 04/
Derive your real ICP from the customers you already have /
Upload your existing customer list. Prospectlab finds what those customers actually have in common and turns it into a profile you can prospect against.
Most ICPs are assumptions. This one is evidence.
The system clusters your customers on shared attributes, scores each cluster, and flags the outliers that don't fit. Then hands you a plain-language summary you can use to qualify new prospects. Once you have an ICP, every new upload runs through it and gets pre-segmented campaigns on the other side.
- Upload your customer list (CSV).
- Clustering on shared attributes: sector, size, growth stage, geography, hiring patterns, technology stack.
- Cluster scoring and outlier detection.
- A plain-language ICP summary you can act on.
- ICPs feed straight into segmentation - pre-segmented campaigns from new uploads.
Usage is measured in customers analysed, nothing to decode.
Feature 05/
Meet Kat - your AI campaign strategist, not a prompt box /
Kat walks you from "I have a list" to "I have an approved campaign" through a guided session, working from a structured knowledge base of what makes cold email get replies.
Kat doesn't generate copy from nothing. She composes from a foundation.
Campaign hypothesis, target / problem / offer fit, sequence design, voice rules, anatomy rules, CTA logic, signal usage. She works in stages, and you always know which one you're in. She gathers the hypothesis first - who is this for, what specific problem do they have, what first step is worth their time - and she refuses to compose copy until that hypothesis is solid.
- Knowledge base, not prompts. She composes against tested rules, not vibes.
- Golden Library. Start from proven, recognised frameworks (message structures and multi-step cadences) and adapt one to your voice, or bring your own. No blank canvas.
- Stage-aware: intake, composition, review, approval, each with its own toolset.
- Composes against explicit voice and anatomy rules: subject length, opening shape, body structure, CTA logic.
- Over 80 specialised tools - signal lookup, voice profiles, sequence rules, CTA libraries, draft validation, version history - and you see what she's doing.
- Human review at every step. Kat advises and composes; you author and own the copy.
If Kat's critique conflicts with your edit, your edit wins.
Feature 06/
One gold-standard email, scaled without turning into a template /
Pattern Studio decomposes the email you co-authored into reusable rotation pools, then assembles a unique sequence for every prospect in the segment.
Templates feel like templates. Prompts produce slop. Pattern Studio is the middle path.
You control the gold standard at index 0, and the system handles scale without losing your signature. Four hundred prospects produce four hundred unique sequences, no two the same, and every variant is a controlled version of the email you approved. Never a hallucinated alternative.
- Decomposition - the gold-standard email becomes rotation pools (openings, body sentences, CTAs, sign-offs).
- Rotation engine - selects from the pools deterministically per prospect; 400 prospects = 400 unique sequences.
- Degradation paths - when a prospect lacks a high-tier signal, the engine falls back through a defined cascade, never to a blank or a guess.
- Conditional logic - each pool slot can carry conditions (use this opening only if signal X; this CTA only if signal Y).
- Pattern library - build once, reuse across campaigns and segments.
Feature 07/
Generated deterministically. Checked four ways. Approved by you /
Once the pattern is approved, generation is rule-based assembly from your rotation pools. No LLM in the path, no hallucination risk, no per-email cost. Reliability comes from the architecture, not from a better prompt.
Most "AI email" tools generate at send time, with no QA, and hope. Prospectlab separates the concerns: composition is intelligent, generation is deterministic, QA is automatic, review is yours.
Four hundred prospects produce four hundred sequences in seconds. Every one passes a four-layer quality gate before it reaches you. Then every email lands in a review queue where you approve, edit, or discard at the email level. When you push to your sending tool, the campaign stays paused until you explicitly start it.
- Deterministic assembly - same prospect, same pools, the identical email every run.
- Review workflow - approve, edit, or discard per email; nothing leaves until approved.
- Push to your sending tool - Smartlead or Instantly. The campaign stays paused until you start it.
01 - Structural QA
Tier integrity, token coverage, guardrail compliance. High-signal prospects got high-signal emails; every personalisation slot filled; no banned phrases or broken seams.
02 - Proofreading
Grammar, readability, Americanisms in British copy, spam-trigger flags, sentence-level smoothness.
03 - Fix engine
Traces a defect back to the pool or rule that caused it, proposes a fix, and applies it only after your approval.
04 - Tier reconciliation
Compares the projected tier (what the data should support) against the actual generated tier; flags any discrepancy.
Built for teams running considered outbound, not blasts /
Prospectlab is built for teams that want precise, signal-led campaigns rather than another high-volume sending tool.
Agencies
Run segmented, deterministic campaigns across multiple client workspaces. Every email is checked before it's sent, and your clients can approve the copy before anything goes out.
Founders
Move from a raw list to structured, signal-led outreach without hiring an SDR team. Kat guides each step.
B2B teams
Stand up a repeatable outbound system: one ICP, one pattern, reused across segments, at a cost you can predict.
The manual workflow, and the one Prospectlab replaces it with /
Each row pairs the step your team runs today with the step the pipeline above replaces it with.
| Traditional outbound prepThe manual workflow most teams run today. | ProspectlabA single pipeline, structured at every step. |
|---|---|
| Spreadsheet imports and manual column cleanup | Guided upload with automatic column mapping |
| Dedupe by hand across campaigns, or not at all | Cross-campaign deduplication on every import |
| Copy-paste research from LinkedIn, ~20 min per prospect | Background enrichment: discovery, verification, signal scraping |
| Every prospect gets the same template | Signal-matched personalisation tiers per prospect |
| ICP is a guess in a slide | ICP derived from your own customer list |
| One clever prompt, hoping it holds at volume | Gold-standard email decomposed into deterministic rotation pools |
| An LLM writes each email fresh at send time | Deterministic assembly. No LLM in the send path |
| Spot-check a few, send the rest on faith | Four-layer QA on every email before you see it |
| Unclear what was imported or skipped | Import receipt with per-prospect status |
| Review scattered across tabs and tools | One review queue. Approve, edit, or discard per email |
Works with the tools your data already comes from /
Bring your list from anywhere; push approved campaigns to the sender you already trust. Prospectlab makes your list worth emailing - it isn't where you buy the list, and it isn't the tool that sends.
Import sources
- Apollo
- LinkedIn Sales Navigator
- Clay
- HubSpot
- Salesforce
- Pipedrive
- CSV
- JSON
Send to
- Smartlead
- Instantly
Your data moves through a controlled, inspectable workflow /
Prospectlab is built around structured processing, visibility, and human review. Customer and prospect data moves through clear upload, enrichment, and review states. Each workspace is isolated, access is protected with multi-factor authentication, and campaigns hand off through your own sending keys, so your data and your sending reputation stay with you.
Questions, answered plainly /
Which files and sources can I upload?
Apollo CSV, LinkedIn Sales Navigator, Clay, CRM exports (HubSpot, Salesforce, Pipedrive), generic CSV, and JSON. The system recognises common formats and maps the columns for you.
Does Prospectlab send my emails?
No. Approved campaigns hand off to Smartlead or Instantly using your own keys. Your sending reputation, warm-up, and inbox placement stay in the infrastructure you already trust.
Can agencies manage multiple clients?
Yes. Multi-client workspaces let agencies run segmented, deterministic campaigns across separate client accounts, with copy clients can approve before anything sends.
Where are your customer results?
We don't have published reply-rate numbers yet, and we won't invent any. Instead of a metric to take on faith, we give you a mechanism you can inspect - and you can test all of it on your own list at $49 a month, no annual lock-in. When we publish numbers, they will be real.
See how Prospectlab fits your outbound, on your own list.
Starter from $49/month
Kat, Pattern Studio, enrichment access, deterministic assembly, and Smartlead or Instantly hand-off on one per-seat price. Enrichment is metered from $0.10 per prospect, only when used.
Upload a list, walk it through the pipeline, and review the output before anything sends.