Prospecting tools that cite the research behind the pitch
Most prospecting tools hand you a list. This one reads the account you already chose, writes down what it found with the line each fact came from, and leaves the boxes it could not fill empty.
Four products call themselves prospecting tools
They solve different problems and a tool that is excellent at one is usually mediocre at the others. This one is the fourth, and it needs the first to exist.
A list
Contact data: who works there, what their address is, what the company does. You are buying coverage and accuracy. Saleswoman is not this and does not include it.
Signals
Something changed: hiring, funding, a technology added or dropped, a leadership move. You are buying timeliness. Point them at us as a source and we will cite them.
Execution
Sequences, timing, reply detection, reporting. You are buying reliability. We do this, and the sending side is on the email outreach platform page.
The research and the writing
Something reads the account and produces the actual message. You are buying judgement, which is the hardest of the four to evaluate in a demo. This is the one.
The empty box is the feature
Every research tool has a field it could not fill. What separates them is what happens next. Most fill it with something plausible, because a slot has to contain something and a plausible sentence is what a language model is for.
Here it stays empty and the drafting step is told so. A fact that is not available as evidence is a different thing from a fact that is false, and the difference is the whole product.
You can test this in about a minute: point the demo at a real account with almost no public presence and read what comes back. A careful tool produces a short, slightly boring email. The other kind produces a confident paragraph about their digital transformation journey.
The brief it assembled for a regional logistics firm, before it wrote a word.
Three fields found, one left empty. The empty one is not an error and is not filled in.
What it found, and where
3 of 4 cited
-
Two dispatch roles open in Columbus.
CITED
Source: the careers page you connected, read at draft time.
-
The renewal on their routing tool lands in July.
CITED
Source: CRM note dated 14 March.
-
A second facility opened in Reno in March.
CITED
Source: their press page.
-
Headcount growth over the last year.
UNVERIFIED
Nothing found in the sources connected. Left empty rather than estimated, so nothing in the draft can lean on it.
What it reads, and what it will not touch
The research is only worth something if you know exactly where it came from, so the sources are a list you control rather than a black box labelled "the web".
Sources it reads
- Account pages you connect: the site, the careers page, the press page, a documentation site.
- CRM records and the notes your team left, with their dates.
- The history of previous contact with this person and this account.
- A signal feed you already pay for, if you connect it. We will cite it by name.
- Anything you paste in yourself, which is all the demo has.
Things it does not do
- Sell you contact data. There is no database here, no email finder and no enrichment to resell.
- Scrape a company it was never given access to.
- Guess seniority, budget or intent and then write as though it knew.
- Treat an inference as a fact. An inference is allowed in a draft, labelled. It is never allowed to pass unlabelled.
- Record or transcribe calls. That is a different category.
If you are still choosing a vendor, the questions worth sending before a demo are in b2b prospecting software, including the three tests that take an afternoon and settle it. If the vendor on the table is a contact database rather than a writing tool, what a quote is built from is on zoominfo pricing, and the contract terms behind a Seamless.AI order form are on seamless ai pricing. A Cognism quote, five seats in each package and one shared credit pool, is turned into a cost per contact on cognism pricing. For a vendor that does print its prices, rocketreach pricing works through the cost per exported contact (the best alternative to RocketReach depends on which job you bought it for), and lusha pricing shows what a five credit phone reveal does to it. uplead pricing shows what a contact really costs on a single user plan you do not fill, and leadiq pricing covers the opposite model, one credit pool shared by five users where a phone number costs ten emails, and lead411 pricing shows the cheapest printed cost per contact with a direct dial, about 5 cents an export on a one person plan. For a team that wants unlimited seats on one bill, snov io pricing shows the three meters behind it and why the phone number, not the plan, grows the bill. If the tool you are replacing is Clearbit, which no longer sells new standalone plans, clearbit pricing shows what the same data costs inside HubSpot. If the tool on the table bundles the database with sequences, a dialer and AI agents, amplemarket pricing shows what the $600 Startup plan costs per user and per contact. If the tool is a workflow builder over many data providers rather than one database, clay pricing splits its card price into data credits and actions and works out the cost per researched row. If the platform on the table ranks accounts by intent and buying stage rather than selling contacts, 6sense pricing lays out the three bundles of its single license and the credit pool that expires unused, and if it is an account based platform for marketing and sales together, demandbase pricing splits the quote into its platform fee and its fee per user. If the quote is for intent data on its own, bombora pricing shows which plans already bundle the same topics.
The guide on this
Twelve questions to ask a vendor, in the order that saves you the most time, and the answers that should end the call.
The rest of the product, by job
The same product, entered from the job you came here to do. Each page states its own limits rather than repeating the pitch.
-
AI email writer
The drafting side on its own page: what it writes, and what it refuses to assert without a source.
-
Spam score checker
What the score is made of, line by line, and the two things it deliberately does not test.
-
Email outreach platform
The cap, the queue and the suppression list, which is the part of a platform nobody demos.
-
CRM automation
Activity logging, deal notes and next steps, written from the thread instead of from memory.
-
Cold email software
The four things a cold email has to get right before volume is even worth discussing.
-
Sales automation tools
What to automate, what to review, and what a person still has to decide.
Point it at your least famous account.
That is the test worth running. A tool that stays quiet about a company with no public footprint is a tool you can put in front of a buyer.
Your data stays yours. Prospect data is never used to train a model. The demo needs no card and no account.