Workflow v1 — End-to-End

Source: Miro board "design" (uXjVH7iuOYc=)  ·  Extracted: 2026-07-14

Scope: Complete sales-call lifecycle — Call Prep → During Call → Post Call → Follow Up

4Stages
28Sub-Queries
8+9Commands
2Frameworks

Overview

The Cloud Latitude v1 sales-call workflow is a four-stage pipeline spanning from pre-call research through post-call follow-up. It is a human-in-the-loop, AI-augmented system that researches the prospect before the call, assists the sales rep in real time during the call via a copilot, processes the transcript into deliverables immediately after, and brands/saves/syncs all outputs to Drive and HubSpot.

AI action / process step
User interview / synthesis
During-call / copilot
Board transcription note
PRE-CALL                DURING CALL             POST CALL               FOLLOW UP
─────────────────────   ──────────────────     ─────────────────     ─────────────────
Form → AI Analyzes →   Real-time transcript   summary-to-email →    Pull design assets →
  7 query groups ×      + 8 copilot commands    1-pager → post-        Apply to content →
  4 sub-queries →       (discover, pain, ...)    slides → JBANT →     Save to Drive →
  synthesize →            transcription          MEDDPIC → homework    Update HubSpot ×4
  Brief hexagon →                              → update-CRM          (note/deal/task/files)
  3-stage interview
  loop → Save to Drive
1 PRE-CALL — Call Prep

Frame: "workflow v1 - call prep" — 72 items · 101 connectors

1.1 Form Input entry point

The user completes a structured form with 8 fields. A "Company URL" label sits above the form as the entry prompt.

#FieldType
1Team membersmulti-select dropdown
2Company info & contactssingle-line text fields
3Contextlong-form text input
4URLsURL field (click to add more)
5Questions to asksingle-line text fields (click to add more)
6Scenario(s)multi-select dropdown
7Audiencemulti-select dropdown
8Output formatsmulti-select dropdown
↓

1.2 AI Analyzes Input + Creates First Set of Research Queries

The AI ingests all form fields and generates the first set of research queries based on the company, context, team, scenario, and audience provided.

↓ fans out →

1.3 Research Query Fan-Out parallel · 7 groups × 4 sub-queries

Each group follows the same internal pattern: 4 sub-queries (A/B/C/D) run independently, then merge via a synthesize chevron.

⚠ Numbering on original board: There is no "New Query 6" — callouts jump 5 → 7 → 8. The last two groups have sub-queries mislabeled. Documented as-is.
New Query 11A · 1B · 1C · 1D → synthesize
New Query 22A · 2B · 2C · 2D → synthesize
New Query 33A · 3B · 3C · 3D → synthesize
New Query 44A · 4B · 4C · 4D → synthesize
New Query 55A · 5B · 5C · 5D → synthesize
New Query 7 ⚠6A · 6B · 6C · 6D → synthesize
New Query 8 ⚠7A · 7B · 7C · 7D → synthesize
↓ all synthesized outputs feed into ↓

1.4 Generate Brief of Findings + Interview User pivot point · hexagon

All synthesized query outputs feed into a central hexagon that compiles a brief of findings. This is the pivot point — the AI now has enough context to begin the interactive interview loop with the user.

↓ triggers 3 sequential stages ↓

1.5 Three-Stage Interview Loop sequential · 9 chevrons

Each stage is a 3-chevron row flowing left → right: AI analysis → confirm/elaborate → interview & synthesize

StageChevron 1 (purple)Chevron 2 (purple)Chevron 3 (green)
A — Topics & TrendsCreate topics, trends, and categories from collective analysisAsk User to confirm or elaborateAsk User questions and synthesize responses
B — Findings & InsightsSurface key findings and insightsAsk User to confirm or elaborateAsk User questions and synthesize responses
C — Content FormatsSuggest best content formatsAsk User to confirm or elaborateAsk User questions and synthesize responses; based on desired formats plan content outlines
↓

1.6 Save Output to Google Drive rectangle

1. Name folder        ['Account_Name'_plan]
2. Create sub-folder  'account_name-research'
3. Save files

Drive: Google Drive link

2 DURING CALL — Real-Time Copilot

Frame: "TRANSCRIBE" — real-time AI copilot with 8 commands

During the live call, the system provides a real-time transcription + copilot interface. The sales rep can invoke 8 slash commands to get AI assistance mid-conversation. Real-time transcription runs continuously, feeding the copilot's context window.

/discoverSurface discovery insights from the prep research
/painIdentify and highlight pain points mentioned by the prospect
/successReference success stories, case studies, or proof points
/connectConnect the prospect's pain points to Cloud Latitude solutions
/buyingAnalyze buying signals and intent indicators
/what's that?Get an on-the-fly explanation of a term or concept
/smartSmart reply suggestion — AI generates a response based on context
/next stepsSuggest next steps and close the conversation forward
3 POST CALL — Processing

Frame: "POST CALL" — 9 processing commands

Immediately after the call ends, the transcript is processed through 9 sequential commands that transform raw conversation data into structured deliverables.

/look againRe-examine the transcript for missed details
/summary-to-emailConvert call summary into a polished follow-up email
/1-pagerGenerate a one-page account summary document
/post-slidesCreate a post-call slide deck for the prospect
/resourcesCompile relevant resources (docs, links, case studies)
/JBANTRun the JBANT qualification framework against call data
/MEDDPICRun the MEDDPIC qualification framework against call data
/homeworkGenerate action items and homework for both parties
/update-CRMPush call data and qualification results to the CRM

Qualification Frameworks

BANT

BBudgetDoes the prospect have budget allocated?
AAuthorityIs the contact a decision-maker?
NNeedWhat pain or problem needs solving?
TTimelineWhen do they need a solution?

MEDDPIC

MMetricsWhat measurable outcomes?
EEcon BuyerWho has final authority to sign?
DDecision CriteriaWhat criteria to evaluate?
DDecision ProcessWhat procurement process?
PPaper ProcessWhat documentation/legal?
IIdentified PainWhat validated pain points?
CChampionWho advocates internally?
4 FOLLOW UP — Persist & Sync

Frame: "workflow v1 - follow up"

4.1 Get Cloud Latitude Design Assets source

Pull the brand design source material from the linked Google Drive folder.

GET CLOUD LATITUDE
DESIGN.MD, DESIGN SYSTEM,
UI KIT FROM THIS FOLDER >
Verbatim from board https://drive.google.com/drive/folders/1o41Z-2jYiW_d_mN6z9TvO7yvZxZgca?usp=drive_link
⚠ Transcription note: The Drive URL on the original board appears slightly mangled (rendered across line breaks). Documented as-is; verify folder ID before relying on it.
↓

4.2 Apply to Content apply

Take the Design System and UI Kit and apply them to the post-call deliverables (one-pagers, slides, summary emails). This is the brand-application step.

APPLY TO CONTENT
↓

4.3 Save Outputs to Google Drive persist

Verbatim from board Save all outputs to [Account_Name_plan] in google drive

Convention: one [Account_Name]_plan folder per account — the single source of truth for that account's generated materials.

↓

4.4 Update HubSpot integration · 4 actions

Mirror the call's outputs into the CRM record. Four sequential write operations:

#ActionVerb on boardWhat it does
4aadd noteadd noteCreate a note on the contact/deal record capturing the call summary
4bupdate dealupdate dealUpdate the deal record (stage, fields, BANT/MEDDPIC, next steps)
4ccreate taskcreate taskCreate a follow-up task (e.g. next-step reminder)
4dattach filesattach follow up items as filesAttach generated artifacts as files to the record
Verbatim from board update hubspot add note update deal create task attach follow up items as files

Totals

4Stages
8Form Fields
28Sub-Queries
17Commands
2Frameworks

Notes

  1. Two Google Drive folders are referenced: one for call-prep saves, one for follow-up design assets. The follow-up URL appears mangled on the original board.
  2. Numbering inconsistencies in the original call-prep frame (missing Query 6, mislabeled sub-queries under Queries 7 & 8). Documented as-is.
  3. JBANT extends the standard BANT framework — the "J" prefix likely represents Cloud Latitude's custom qualification variant.
  4. HubSpot action order is intentional: note → deal → task → files.

Workflow v1 — End-to-End  ·  Extracted from Miro board uXjVH7iuOYc=  ·  2026-07-14