Call-to-Custom-Deck Automation
Convert every recorded sales call into a reviewed follow-up deck
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 6
- Confidence
- 99%
Record a sales conversation with a note-taking tool such as Granola, Gong, or Fireflies, then use Zapier or Make to pass the transcript or summarized notes into an AI presentation generator. Generate the result from an approved proposal template so every output begins with consistent structure and branding. The call provides the prospect’s needs, objections, decisions, and next steps, while the template supplies the repeatable sales narrative. The deck appears automatically in the presentation workspace for the representative to inspect at the end of the day. Crucially, automation stops before delivery: a person reviews facts, positioning, and anything the representative said before sending the deck to the prospect.
Origin
Kristen Frackett demonstrated a Granola-to-Zapier-to-Gamma flow that created a proposal deck from a synthetic Ridgeline Software call. Extracted from Marketing Against The Grain.
Core principles
- 01Let the call supply the customer-specific context
- 02Standardize the output with an approved proposal template
- 03Automate repetitive production rather than customer judgment
- 04Keep a human approval gate before external delivery
- 05Capture explicit next steps from the conversation
How to run it
- 1
Capture the conversation
Record and transcribe the sales call using the team’s approved note-taking system.
Pro tip Use consistent folders or labels to identify calls that should trigger deck creation.
Watch out Confirm that call recording and transcription comply with applicable consent and privacy requirements.
- 2
Trigger the workflow
Use Zapier or Make to detect a completed call transcript or set of notes and pass it into the deck workflow.
Pro tip Filter out internal calls and conversations that do not require follow-up.
Watch out A broad trigger may create irrelevant or duplicate decks.
- 3
Apply the proposal template
Create the deck from an approved template containing the standard sales structure and brand system.
Pro tip Include mandatory slides for needs, proposed solution, proof, and next steps.
Watch out Generating without a template can produce inconsistent positioning across representatives.
- 4
Personalize from the call
Use the transcript to populate the prospect’s priorities, relevant solution details, and agreed next actions.
Pro tip Tell the model to distinguish explicit customer statements from inferred needs.
Watch out Summaries can erase qualifiers or turn tentative discussion into apparent commitments.
- 5
Review the deck
Have the representative verify customer facts, quotations, promises, and next steps.
Pro tip Compare sensitive claims directly against the transcript.
Watch out Do not let polished slides bypass normal sales approval standards.
- 6
Deliver the follow-up
Send the approved deck with a concise message linked to the conversation and agreed action.
Pro tip Deliver promptly while the call context is still fresh.
Watch out Do not automate direct emailing until the workflow has demonstrated reliable accuracy.
In the wild
A representative completes ten discovery calls recorded in Granola. Each transcript triggers Zapier, which generates a prospect-specific proposal from the company’s standard template. At the end of the day, the representative reviews the ten drafts, corrects any questionable details, and sends the approved versions.
→ Every prospect receives prompt visual follow-up without requiring ten manual deck builds.
A software salesperson’s call notes identify the buyer’s integration concern, purchasing timeline, and requested proof. The automation places those details into the relevant template sections and adds the next meeting date.
→ The follow-up reflects the buyer’s actual priorities and makes the agreed process easy to understand.
Common mistakes
Auto-emailing unreviewed output
A transcript may contain inaccurate summaries or statements the representative should not formalize in writing.
Ignoring the template layer
Using only raw call notes can produce decks that are personalized but inconsistent with approved sales positioning.
Treating inference as fact
The model may convert an implied need into a definitive customer statement unless prompted and reviewed carefully.
Is it for you?
Best for
Sales teams handling enough daily calls that manual personalized follow-up is slow or inconsistent.
Not ideal for
Sensitive negotiations where incomplete transcripts or nuanced commitments make automatic summarization unsafe.
From the transcript
“I record on granola, my notes from all my first calls. I put those in this folder, and then automatically I'm creating a DAC in…”
“There's no prompting. This is all automated. It's just taking in the transcript.”
“I don't trust AI that much yet. It's not the AI I don't trust. It's like if your rep says some things that you don't…”
From the episode
This AI Workflow Turns Every Sales Call Into a Custom Deck