GTM AI Podcast 2026 #1-AI Agent live build to increase 30% SQLs
It’s 2026 people! We are excited to announce some changes this year to the podcast and the newsletter.
Let’s start with the podcast.. first we have had some amazing guests who have shared powerful insights, in fact we put together a NotebookLM with amazing insights from both the past year AND trends going into 2026.
Moving forward we are updating the format so that along with the interview, our guests will be doing a deep dive SHOWING exactly behind the scenes of automations, building AI agents, and actual strategy. Screen sharing and showing you EXACTLY what is going on with companies and their AI implementation. Today we start with the interview with Justin Parnell of JustinGPT doing a deep dive into his 6 step AI agent.
The newsletter will change a bit to now instead of doing a once a week update, we will be doing posts here on Substack and a once a week article going into AI trends applied to GTM professionals and leaders to help make it consumable and easy to implement and use.
Justin has given away a ton of goodies you can get here below and use for yourself!
You can go to Youtube, Apple, Spotify as well as a whole other host of locations to hear the podcast or see the video interview.
This episode features Justin Parnell (JustinGPT) demonstrating how to transform static lead magnets into dynamically personalized assets using AI automation. The core insight: most companies gate generic PDFs behind forms, leaving 70% of conversion potential on the table. Justin’s solution automatically customizes sales collateral based on prospect inputs, driving 30% increases in MQL-to-SQL conversion rates for vertical SaaS clients, at a cost of roughly 20 cents per personalized asset.
Why This Matters for GTM Leaders
Traditional lead magnets follow a one-size-fits-all approach: prospect submits form → receives static PDF → sales follows up days later with generic outreach. This workflow ignores critical context the prospect just provided and fails to demonstrate understanding of their specific use case.
The automation Justin built solves three problems simultaneously:
Speed: Custom proposals delivered in under 60 seconds, not 48 hours
Personalization: Content dynamically adapted to prospect’s industry, role, and stated objectives
Scale: Sales teams receive qualified leads with pre-built, personalized collateral already in prospect’s hands
The downstream impact extends beyond conversion rates. When prospects receive immediately relevant content, sales conversations start from shared context rather than discovery basics. One client reported reducing first-call duration by 40% because prospects arrived educated and aligned.
The Technical Architecture
Stack Components & Strategic Rationale
Airtable (Data Layer)
Justin chose Airtable over Google Sheets for three reasons: robust API flexibility, atomic data structure that enables cross-agent workflows, and native integration with Claude Desktop for rapid synthetic data generation during testing. As he explains: “The customization and atomic nature of this to combine into other molecules... just makes it a lot more powerful.”
Make.com (Orchestration)
Critical choice: Make.com offers native Gamma integration that exposes text preservation modes and theme controls not available in n8n or Zapier. Justin notes: “Sabrina Romanov tried to do the same thing with n8n and she ended up telling her users to use Gamma with Make because there was a native one there.”
Claude Opus 4.5 (AI Processing)
Specifically Opus 4.5, not Sonnet. Reason: superior adherence to design standards and brand guidelines. Lighter models produce inconsistent outputs where “your logo is in a wonky spot on page five.” Cost differential is negligible compared to conversion impact.
Claude Skills (Custom Instructions)
This is the secret weapon. Skills are markdown files that give Claude exact instructions for content transformation. Justin developed his by conversing with Claude, iterating until output matched requirements, then uploading the skill to platform.claude.com. The skill eliminates prompt engineering on every run.
Gamma (Design Engine)
Handles presentation creation with deterministic brand consistency. Justin emphasizes: “When I let Claude develop the PDF, it is a little janky... I highly recommend using Gamma with a prebuilt template to generate this content... you will see very consistent, very deterministic results.”
Step-by-Step Implementation Process
Phase 1: Foundation Setup (60 minutes)
Build Your Airtable Base
Create a new base with these fields: First Name, Last Name, Email, Company, Title, Business Objective, Current Process, Use Case Title, Submission Time, Gamma Output URL. Enable form view for external submissions.
Pro tip: Connect Airtable to Claude Desktop and generate 5-10 synthetic test records by asking Claude to populate realistic data for your industry.
Create Your Gamma Template
Log into Gamma → Settings → Themes → Create Custom Theme. Either input your website URL (Gamma will auto-extract brand colors and fonts) or manually enter hex codes. Build a 10-slide template matching your standard deck structure. Save as your default template.
Generate Your Claude Skill
Open Claude.ai and describe your content transformation goal
Upload your current lead magnet PDF
Ask Claude to create a skill that personalizes this content based on form inputs
Iterate on examples until output meets quality standards
Download the skill package (zip file)
Navigate to platform.anthropic.com → Skills → Add Skill
Critical: Drag and drop the zip file directly into the upload square (clicking upload won’t work)
Justin’s process: “We just had a back and forth conversation about how to build that. And it gave me a sample output... and we went back and forth a couple of times until I got it to give me what I wanted.”
Phase 2: Workflow Construction (45 minutes)
Make.com Scenario Setup
Trigger Module: Airtable → Watch Records → Select your base and table → Set trigger field to “Submission Time” → Limit to 1 record per execution (adjust based on volume)
AI Processing Module: Claude → Generate Content → Select Opus 4.5 model → Set effort level to Medium → Allocate 50,000 tokens (won’t use all, prevents cutoff) → Input type: Single String
Map Form Data: Pull Airtable fields (name, company, objectives, current process) into user prompt variables
System Prompt Template:
You are creating a customized proposal for [COMPANY] based on their stated objective: [BUSINESS_OBJECTIVE].
Their current process: [CURRENT_PROCESS]
Use the uploaded skill to transform our standard methodology into a personalized recommendation addressing their specific situation. Maintain professional tone and our brand voice.
Skills Configuration: Select “Custom” skill type → Choose your uploaded skill → Select latest version → Enable skill in processing
Gamma Generation Module: Gamma → Create Presentation → Input text: Map Claude’s output → Select “Preserve text structure” → Choose your custom theme → Set to 10 cards (adjust based on content depth) → Enable AI slide breaks → Image source: AI Generated (Nano Banana Pro recommended) → Format: 16:9 → External access: View → Export options: Enable PDF
Email Delivery Module: Gmail → Send Email → Recipient: Map Airtable email field → Subject: “Your Custom [Use Case] Proposal, [First Name]” → Body: Include Gamma view link and PDF download link
Notification Module: Slack → Send Message → Channel: #new-leads → Message: “New request from [First Name] at [Company] for [Use Case Title] - Deck sent: [Gamma URL]”
Database Update Module: Airtable → Update Record → Field: Gamma Output URL → Value: Map Gamma link
Phase 3: Testing & Optimization (30 minutes)
Run 3-5 test submissions with varied inputs. Check for:
Brand consistency across all slides
Appropriate image generation matching content context
Proper variable substitution (no [PLACEHOLDER] text in output)
Email deliverability and link functionality
Slack notification accuracy
Cost monitoring: Each execution consumes approximately 3,300 output tokens and 17,000 input tokens from Claude (~$0.05-0.10) plus 10 Gamma credits. Gamma Pro ($20/month) includes 8,000 credits—sufficient for 200+ executions monthly.
Five Additional Use Cases Beyond Lead Magnets
1. Post-Demo Follow-Up Automation
Trigger: Sales rep checks box in Salesforce after demo
Process: Pull call recording transcript from Momentum.io → Claude summarizes key discussion points and prospect pain points → Gamma generates custom one-pager highlighting relevant features discussed → Auto-sent within 5 minutes of demo completion
Impact: Reduces sales cycle length by maintaining momentum; prospects receive tailored recap instead of generic deck
2. Competitive Displacement Campaigns
Trigger: Prospect identifies current vendor in form
Process: Airtable lookup matches vendor to competitive intelligence database → Claude generates comparison highlighting switching benefits specific to their use case → Gamma creates battle card → Email includes ROI calculator pre-populated with their business metrics
Impact: Arms prospects with business case before first sales conversation
3. Multi-Stakeholder Buying Committee Assets
Trigger: Champion uploads list of decision-makers with roles
Process: For each stakeholder (CFO, CTO, VP Sales), Claude generates role-specific value proposition → Gamma creates individual one-pagers addressing their priorities → Batch email to champion with all assets
Impact: Enables champions to evangelize internally with professional, personalized materials
4. Event/Webinar Custom Recaps
Trigger: Attendee submits post-event feedback form indicating interest areas
Process: Claude extracts relevant content segments from event transcript matching stated interests → Gamma generates personalized “Your Custom Event Recap” with action items and resource links
Impact: Increases post-event engagement and content consumption
5. Sales Proposal Generation from Discovery Calls
Trigger: Rep completes discovery call notes template
Process: Claude transforms notes into structured proposal following company framework → Gamma generates branded deck with pricing scenarios, implementation timeline, success metrics → Auto-added to CRM opportunity
Impact: Eliminates 3-5 hours of proposal writing per deal; ensures consistency across sales team
Economics & Scale Considerations
Per-Execution Cost Breakdown:
Claude API: $0.05-0.10 (Opus 4.5)
Gamma Credits: $0.03 (assuming Pro plan)
Make.com Operations: $0.06 (6 steps at $0.01/step on base plan)
Total: ~$0.20 per personalized asset
ROI Calculation (based on client data):
Baseline MQL-to-SQL conversion: 15%
Post-implementation conversion: 19.5% (+30%)
Average deal size: $50,000
Cost per 100 MQLs processed: $20
Additional revenue from 4.5 converted SQLs: $225,000
Return: 11,250x
Justin’s client insight: “Some of my clients have seen their MQL to SQL conversion rate increased by 30% by putting in just this simple, customize your lead magnet content into their workflow... that goes down the funnel to 30% more revenue.”
The Future-Forward Implication
Justin and I discussed an emerging paradigm: agent-to-agent GTM. Within 18-24 months, buyers will deploy AI agents to evaluate vendors, submit RFPs to multiple providers simultaneously, and compare proposals—all without human involvement until final decision stage.
As Justin predicts: “The next paradigm that we’re fast approaching for go-to-market teams is designing for agents... buyers will have agents that filter content for them, prepare and analyze content for them, and marketing teams will be tasked with marketing to an agent.”
This shifts content strategy from visual appeal to structured data optimization. Instead of beautiful PDFs, winning vendors will provide machine-readable schemas, API-accessible evaluation criteria, and Claude Skills that enable buyer agents to extract exactly what they need.
The modular automation Justin demonstrated that it’s training ground for the agent-native GTM era. Teams building these systems now are developing the muscle memory and technical infrastructure they’ll need when agent-to-agent becomes the default buying process.
Access Resources
Connect with Justin Parnell:
Schedule consultation: justingpt.ai
YouTube technical deep-dives: JustinGPT Channel
Request custom agent blueprint: Airtable Form
View sample output: Gamma Example
Download Make.com blueprint: JSON Template
Implementation Note: The Make.com blueprint includes all prompts and module configurations. You’ll need to add your own API keys and customize prompts to your business, but the infrastructure is production-ready.
Closing Insight
The automation demonstrated is deceptively simple—six workflow steps, 20-cent execution cost, 60-second runtime. But simplicity is the strategy. As Justin emphasizes throughout: “You want to create agents that do specific things and chunk those things out into a longer job to be done... if we combine these together it would look like a big spider graph... but this is how I recommend tackling these things.”
The companies winning with AI in GTM aren’t building complex mega-agents. They’re building focused micro-agents that execute single jobs exceptionally well, then chaining them together strategically. Start with one high-impact use case, prove the conversion lift, then expand the system methodically.
Your lead magnet personalization agent is the first link in that chain.
Let me know what you think!