Offers › Buyer guide
Buyer guide · Updated 14 October 2026
Which marketing research and AI copilot plan do you need?
We make these products and earn the full sale price when you buy. This guide recommends between our own products, says what each one's proof does and does not show, and names the free or cheaper option where it is enough.
In shortThe Marketing & Sales Library is a 62-volume research system with 7,440 source-trailed records and a copilot that answers only from them, citing each source. The public layer is free. Operator is $49 a month for one person; Pro is $149 for consultants whose work goes to clients; Team is $499 for five named internal seats. In a measured test, 60 of 60 citations were valid and it refused every question outside its library.
What has been shown
It has the most published proof of anything we sell, and the proof measures what buyers worry about with AI tools: whether the answer is made up.
| Proof | What it measured | Result |
|---|---|---|
| Citation accuracy, 2 Sept 2026 | Whether each cited record exists and supports the sentence citing it, scored by a different model from the one that answers | 60 of 60 citations valid. 53 of 59 claims supported (89.8%). 8 of 8 out-of-library questions refused. 3.5 citations per answer, 6.0 seconds on average. |
| Agency run, 15 Sept 2026 | A licensed run on the questions an agency asks while onboarding and running a client | 16 of 16 in-scope questions answered with sources. 2 of 2 follow-ups handled in context. 2 of 2 out-of-scope questions refused. 1 minute 53 recording. |
| Discord transcript, 13 Sept 2026 | Every reply from the live bot's engine, published | 30 questions, 0 wrong or unsourced answers, 5 refused where the library does not cover them. |
What that proof does and does not show
- We make this product and ran these tests ourselves. They are not independent.
- 89.8% means about one claim in ten was not fully supported by the record it cited. That is the honest figure. An earlier 96.7% was withdrawn because it measured retrieval, not accuracy.
- Refusing questions outside the library is a feature and a limit: the copilot will not answer what the 62 volumes do not cover.
- None of these tests measures a business result from using the library.
Which one to choose
| If you are | Choose | Price |
|---|---|---|
| Exploring, or you need an occasional answer | The free public layer | Free, no card |
| One person making your own marketing decisions | Growth Operator | $49 a month |
| A consultant, agency lead or growth hire whose work goes to clients | Growth Pro | $149 a month |
| An in-house team of up to five people | Growth Team | $499 a month |
| A Team workspace that has used its monthly queries | 500-Query Copilot Top-Up | $199 |
Who should buy none of these
Someone who needs answers on a subject the library does not cover. The copilot refuses rather than guesses, which is the point of it, so check the free layer covers your questions first.
The honest alternative
A general AI chatbot answers anything, instantly, at no cost. It also invents sources. Choose the library when you need to show where a claim came from; choose a general chatbot when you do not.
The main drawback
It covers marketing, sales and growth, not every subject, and the monthly query caps (100, 500 or 1,500 pooled) are real limits on heavy use.
Prices and what you get
| Plan | Price | What it includes | Terms | |
|---|---|---|---|---|
| Growth Operator | $49 a month | The full 62-volume research and product layers, the Growth Engine self-audit, copilot with 100 queries a month, 1 seat, internal-use exports | Cancel anytime on monthly plans | Buy |
| Growth Pro | $149 a month | Operator plus the Diagnostic Evidence Engine, the benchmark reference database, branded client-ready exports with limited delivery rights, copilot with 500 queries a month | Cancel anytime on monthly plans | Buy |
| Growth Team | $499 a month | Pro tooling for five named seats, shared diagnostic history, copilot with 1,500 queries a month pooled, internal-use rights. No paid client delivery, no API | Cancel anytime on monthly plans | Buy |
| 500-Query Copilot Top-Up | $199 | 500 more pooled copilot queries for a Team workspace | Buy |
Prices and terms checked against each checkout on 14 October 2026. Prices are in US dollars.
Questions buyers ask
Does the copilot make things up?
It answers only from the library and cites every record it uses. In the published test, 60 of 60 citations pointed at a real, retrieved record, and about nine in ten claims were fully supported by what they cited. When the library does not cover a question, it refuses.
What is free?
The public layer of all 62 volumes, with no card required.
Should I choose Operator or Pro?
Pro if your output goes in front of clients: it adds client-ready exports, delivery rights and the evidence engine. Operator if you are deciding for yourself.
Can I use the Team plan for client work?
No. Team is for internal use. Paid client delivery needs Pro.