Use AI to improve preparation and follow-through. Keep customer understanding, factual accuracy, and permission at the center.
Prepare to learn, not to perform familiarity
AI can help organize legitimate information about a prospect and prepare relevant questions. Supply approved notes and publicly available business information you have checked. Ask what is known, what is uncertain, and what you should learn in conversation.
Do not ask it to invent a personal connection, claim you read something you did not read, or infer sensitive details about an individual. “I saw your team offers three service tiers” is useful only when you actually verified that information.
Turn features into discovery questions
Before discussing your offer, explore the current process, its friction, its consequences, and who is involved in a decision. If you sell help with lead follow-up, ask how new inquiries arrive, how ownership is assigned, and how unanswered inquiries are noticed.
A fictional example: a studio says leads “sometimes disappear.” That is a starting point for discovery, not proof of lost revenue. Ask to trace a recent inquiry and understand what happened before proposing an automation.
Use role-play to practice listening
Ask AI to play a skeptical buyer who has reasonable concerns about time, cost, or change. Practice one question at a time. Then review whether you interrupted, assumed the problem, or answered a question the buyer never asked.
Role-play is rehearsal, not customer research. A convincing simulated objection does not establish what your actual buyers believe. Use it to prepare better questions and return to real conversations for evidence.
Draft follow-ups from what was agreed
After a conversation, provide approved notes and request a concise recap: the situation, the outcome sought, unresolved questions, and the agreed next step. If there was no agreed next step, ask a clear question rather than inventing a commitment.
For example, “You mentioned difficulty tracking inquiries. Would it be useful to review one recent example together?” is more accurate than “As agreed, we will begin your implementation Monday” when nobody agreed to that.
Review every draft before sending. Check names, product fit, commitments, pricing, and whether the recipient has a reasonable expectation of hearing from you. Generating a draft does not authorize a system to send it.
Make pipeline review more useful
Use AI to organize approved opportunity notes into next actions, missing information, and stale assumptions. A useful review distinguishes “no response yet” from “not interested” and “proposal sent” from “decision scheduled.”
Track outcomes that matter: qualified conversations, clearly agreed next steps, time to follow-up, and conversion through defined stages. Do not optimize only for the number of messages sent. A smaller number of relevant, respectful conversations can teach you more than automated volume.
Prompts to put it into practice
Replace the brackets with your context. Use information you have permission to share.
Prepare discovery questions
Using these verified business notes, suggest six discovery questions about the current process, its consequences, and the desired outcome. Mark assumptions and avoid leading questions that presume my service is the answer. Notes: [notes].
Draft a faithful follow-up
Create a short follow-up using only these meeting notes. Include confirmed needs, unanswered questions, and the agreed next step. If a date, promise, or decision is missing, leave it unresolved. Do not invent familiarity or urgency.
Try it on your work
Use an anonymized past sales conversation. Draft a follow-up, compare it with what was actually agreed, and identify one question that would have improved discovery.
Check the result
- Personalization is accurate and relevant.
- The draft preserves the customer’s actual words and commitments.
- A person approves outreach and respects contact preferences.
