Customer-Engagement Prep Pack.

Long-running AI chats are expensive coworkers. They accumulate token cost, lose context, and bury tacit knowledge that nobody else on the team can see. The Customer-Engagement Prep Pack replaces that with four inputs you can hand to a fresh chat in five minutes — and a maintenance loop that keeps the inputs sharper over time.

The procedure exists for three reasons. Reduce reliance on mature chats — durable artifacts in SharePoint beat a 200-message thread nobody else can read. Capture each associate's unique knowledge — insider analysis, EDA 3.0 receptivity reads, cultural-impasse observations belong in a dossier any colleague can load into a fresh chat months later. Wire every artifact to HubSpot — linking dossiers and GTM materials to Company / Contact / Deal records compounds the intelligence and positions us for the day HubSpot itself can generate prep on top of the connected corpus.

The structure isn't decorative. Four named inputs with explicit roles prevents the model from collapsing them into mush. Asking for a one-page brief, three questions, two objections, one priority, and a follow-up is concrete enough that the output is usable as-is. Anchoring the follow-up to HubSpot deal stage closes the loop — the prep produces the next CRM action, not just talking points.

Before the meeting (≥ 24 hours out)

1. Confirm the account is dossiered. Open Target_Accounts / <Account> / 05_Internal_Enablement. If a current Strategic Dossier is there, proceed. If not, escalate to the dossier owner; for prospects without a dossier, prepare a stub from the template and flag the gap.
2. Run the opportunistic field-intelligence sweep if the dossier is > 30 days old and the account has had material HubSpot activity since the last refresh. This is part of the four-input assembly, not a separate task.
3. Confirm the client GTM set is current. GTM Strategy and Playbook should be dated within the current quarter. If not, escalate to the client account lead before the meeting.
4. Pull each attendee's LinkedIn bio. Full "About," "Experience," recent activity. Into prep notes — not SharePoint.
5. Pull the HubSpot context. Via connector or export. Note duplicates and hygiene issues, but redact those from any client-facing artifact.
6. Open a fresh AI chat. Not a reused one. Paste the standing prompt and attach / paste the four inputs.
7. Generate the prep brief, then critique it. The model is a force multiplier on the inputs, not a substitute for judgment. Adjust where it has missed the dossier's insider read.

At the meeting

• Take notes against the four prep deliverables — exec brief, questions, objections, priority.
• Capture what surprised you, what contradicts or strengthens a dossier thesis, and any named third parties (competitors, partners, acquisition targets) the contact mentioned.
• Do not commit to anything that contradicts the GTM Strategy without an offline check with the client account lead.

After the meeting (≤ 48 hours)

1. Log activity in HubSpot. Meeting note, attendees, key takeaways, next steps with dates and owners. Update deal stage if warranted. Field intelligence goes in the note body — this is the input that feeds the next dossier maintenance pass.
2. Decide: Path A or Path B for each insight (see §05). Judgments go to the dossier directly. Facts go to HubSpot and the next sweep promotes them.
3. If the prep was unusually strong or weak, note it in the SOP feedback log. Continuous improvement of the standing prompt depends on field signal.

Two triggers

• Opportunistic, before each prep. If the dossier is > 30 days old and there's been material HubSpot activity since, run the sweep before assembling the prep. Catches the high-priority case: accounts you're actively engaging with.
• Monthly catch-up sweep. First business day of each month, across every dossiered account with HubSpot activity since the previous sweep. Operations owns the cadence; the dossier owner per account owns the actual update. Catches the lower-priority case: accounts that have generated intelligence but no recent meeting.

Client scoping rules

• Rise DA vs. Glide. Rise DA is a separate AI Tech Sales client from Glide. Never include Rise DA accounts, deals, or campaigns in any Glide-facing report or pipeline analysis, even if they appear associated with Glide in HubSpot.
• InPsy = InPsyTech. When generating any InPsyTech-facing artifact, never include data, deals, contacts, tasks, or activity belonging to other AI Tech Sales clients. Cross-check task / note context before attributing.
• Client-facing vs. internal. When building any client-facing dashboard or report, redact internal data-hygiene issues, CRM cleanup tasks, duplicate-record flags, email tracking gaps, and AI Tech Sales internal process notes.
• Sweep scoping. The maintenance sweep prompt operates on a single account's HubSpot notes. Never feed cross-client notes into a sweep.

AI / tooling discipline

• Fresh chat per engagement and per sweep. The prep pack provides everything the new chat needs.
• Treat mature chats as expensive coworkers. Use sparingly, only where accumulated history is genuinely the value.
• Ask for a handoff brief before retiring a long chat. Do not copy old chats into new ones. The handoff brief is reviewable and editable; raw transcript is neither.
• Never paste client-confidential material into an unsanctioned tool. Approved tooling only.
• Save the brief, not the chat. The chat is ephemeral. The brief attaches to the HubSpot meeting record.