What this guide is for
O*NET describes this role as selling wholesale or manufactured goods and requiring substantial knowledge of what is sold. That makes product facts, claims, availability, price, terms, and buyer-facing communication work for approved people and sources—not a general-purpose model. Use AI here as a local preparation clerk: it can make the evidence gaps easy to see before a human decides what, if anything, may happen next.
Make one review packet, not a synthetic sales move
needs_review.Bound the scenario and authority.
Name a fictional or organization-approved scenario, product/category owner, account owner, pricing/claim reviewer, and prohibited actions. Stop for private data, an unapproved source, a requested sales conclusion, or an instruction to contact or update a system.
Inventory approved public material only.
AI may capture source labels, owner, version/date, stated product fact, and permitted use. It must not infer buyer intent, need, authority, product fit, availability, or price.
Make unknowns visible.
If a fact, version, approval, price, availability, claim, or authority is missing or conflicts, label it
unknown. Do not fill the gap with model memory or a plausible sales story.Draft questions, never promises.
Prepare neutral questions for named product, pricing, marketing, legal/compliance, privacy/security, and account reviewers. Do not draft a quote, proposal, discount, negotiation, or message.
Hand off and stop.
Create the review packet and name people who can approve, revise, reject, or escalate it. The packet never authorizes a customer-facing or system action.
Give AI a clerk job, never a sales-rep job
Organizes approved public-source labels, dates, owners, and stated permitted uses.
Flags missing, stale, duplicate, unsupported, or conflicting material without resolving it.
Drafts neutral reviewer questions tied to named human authority.
Safe preparation vs. approval-required work
| AI may prepare | Human authority must decide or do |
|---|---|
| Inventory approved public source labels and metadata | Access private contacts/accounts or choose a source for business use |
| Flag missing, duplicate, stale, unsupported, or conflicting support | Assess buyer intent, fit, authority, lead status, product suitability, or account strategy |
| Produce an unknown list and neutral review questions | Approve product facts, claims, availability, price, discount, quote, terms, or proposal |
| Assemble a local review-only receipt | Negotiate, message, update CRM/order/account records, or take any external action |
FTC guidance offers broad advertising and data-security context. It does not establish a particular claim, consent basis, privacy/security posture, industry rule, contract obligation, or legal outcome. Obtain current organizational review.
Five ways a preparation tool becomes an unsafe sales tool
- Calling a summary a buyer insight. A tidy source list is not evidence of intent, need, authority, fit, or readiness to buy.
- Turning an unknown into a product promise. Product facts, availability, price, discounts, claims, and terms require current approved sources and people.
- Silently blending source versions. Preserve owner, date, stated use, and conflict rather than normalizing them into a single confident statement.
- Adding contacts “just in case.” Minimize inputs; do not paste private contacts, CRM notes, order history, contracts, or account data into a general-purpose tool without separately authorized review.
- Letting a draft become outreach. The packet ends at
needs_review; it cannot recommend, quote, negotiate, message, create, update, or send.
Account Brief & Review Desk
Enter only fictional or organization-approved, minimum-necessary public text. This form stays in your browser and produces a review-only receipt; it does not send or save anything.
What this manual is built on
- O*NET: Wholesale & Manufacturing Sales RepresentativesOccupation and task context; not a sales procedure.
- BLS: Wholesale and Manufacturing Sales RepresentativesBroad U.S. role context; not a demand or sales-outcome prediction.
- NIST AI 600-1Voluntary generative-AI risk-management framing.
- FTC: Advertising FAQsMarketing-claim context; obtain applicable review.
- FTC: Protecting Personal InformationData-minimization and security context; not a compliance determination.
The cited sources and the guide boundaries should be rechecked before real-world use.
Connect this desk without blurring authority
Use the Sales Engineer Technical Discovery Desk for technical pre-sales evidence, the Customer System Guide for broader approval-gated customer work, and the AI Agent Workflow Builder to keep preparation roles least-privileged. These are related guidess only until an approved integration verifies them.