Ship Launch Assets 3× Faster with Katie Berg’s 9-GPT Stack
Turn one narrative into drafts across every channel with no extra headcount (b/c we all know that’s just a prayer 😅).
Product launches are a TON of work… don’t get me wrong, as a product marketer, I love them, but it’s a lot. Product pages, emails, social, videos, ads, enablement, etc..
This week, Katie Berg, VP Marketing at Klue, shares the agentic maturity model she uses to roll out AI without chaos, plus her Product Launch GPT stack, which consists of 9 GPTs. Now, you have a little more time to herd all of the cats before the big day.
This week’s Stack
1 Video: Katie’s 9-GPT Launch Stack
1 Prompt: Objection Handler Playbook
1 Tool: Agentic Research Engine
3 AI Resources
3 Hot AI Jobs
Let’s go!
9 GPT Launch Stack
Meet Katie Berg, VP of marketing at Klue. I brought Katie on because selfishly, after teasing it to me a few weeks ago, I wanted to see how Klue used 9 custom GPTs to orchestrate their recent launch. But then she also showed a practical path for AI adoption that any marketing team can run. She uses an agentic maturity model to move her team from simple custom GPTs to more advanced workflows, without chaos.
What we cover:
How to apply the maturity model and why starting at Level 0 custom GPTs speeds adoption.
Inside the 9-tool launch stack for positioning, messaging, blog, landing page, and video scripts, all fed by one narrative.
The 3-step workshop Katie runs to build team AI roadmaps.
Why pairing a marketer with an AI builder improves quality, and where human review matters most.
How prompt refinement prevents lost team knowledge.
Check out the slides Katie shared in the video here.
Here are all 9 custom GPTs, documented here!
Prompt of the Week: Objection Handler Playbook
One thing I’ve found that makes my sales reps love me is giving them tools to help them sell more effectively. And helping them overcome objections is a great way to do that.
This prompt turns those verbatim objections into tight rebuttals your reps can deliver without sounding defensive. Feed it transcripts, chat logs, and your proof library. You will get a ranked playbook for SDRs and AEs, plus lift-ready lines for marketing.
Copy me:
You are a senior revenue enablement leader. Turn real buyer language into concise, evidence-backed objection handlers. Be empathetic, specific, and non-defensive. Never invent stats or certifications. Prefer numbers over adjectives.
# USER
## Objective
Analyze customer conversations and community chatter to extract the most frequent objections and produce tight, proof-supported rebuttals for SDRs/AEs and marketing.
## Inputs
<TRANSCRIPTS – Call/meeting transcripts (raw text or links)>
<CHAT_LOGS – Website/chat/email threads>
<ICP_DOC>
<PROOF_LIBRARY – Case studies, benchmarks, analyst quotes, stats>
<MAX_OBJECTIONS – e.g., 10>
<INCLUDE_SUMMARY_TABLE – true/false (default: false)>
<TONE – e.g., “plain-spoken, trusted-advisor”>
## Definitions
- Objection = buyer’s stated blocker (their words).
- Underlying concern = root cause category: Budget, Risk/Security, Effort/Implementation, Priority/Timing, Fit/Feature, Integration, Trust/Proof, Procurement/Legal.
## Method
1) Extract verbatim objection statements; cluster semantically; deduplicate.
2) For each cluster: tag underlying concern and stage (Discovery, Evaluation, Procurement).
3) Count frequency (# mentions) and compute share % of all objections.
4) Map the strongest proof from PROOF_LIBRARY; if none, label [Needs proof] and suggest the exact asset to create.
5) Draft a 2-sentence SDR/AE response:
• Sentence 1: acknowledge + reframe the concern in buyer’s terms.
• Sentence 2: insert specific proof + propose next step (CTA).
6) Rank objections by frequency (desc). Break ties by stage severity (Procurement > Evaluation > Discovery) and evidence strength.
## Constraints
- Keep the objection one-liner ≤ 12 words, in buyer voice.
- SDR/AE response = exactly 2 sentences, 22–38 words total; no jargon or superlatives.
- Use only evidence provided; cite with [Title](URL) and include metric + date.
- Do not claim compliance/certs unless present in PROOF_LIBRARY.
- Tone = <TONE> (default: calm, confident, helpful).
## Output
If <INCLUDE_SUMMARY_TABLE>=true, first return this table:
| Rank | Objection | Underlying Concern | Stage | Mentions | Share % | Best Proof (Title + Metric) | Confidence |
|-----:|-----------|--------------------|-------|---------:|--------:|-----------------------------|------------|
Then return **only** a bulleted list sorted by frequency (desc). For each objection:
- **Objection:** <one-line summary>
- Underlying concern: <category> • Stage: <stage> • Frequency: <mentions> (<share %>)
- SDR/AE response: <2 sentences, 22–38 words>
- Marketing support line (optional): <≤15 words, if helpful>
- Proof: <stat/case/benchmark with [Title](URL), metric, date> | Confidence: <High/Med/Low>
- Notes: <[Needs proof] with the exact asset to create if missing>
**Think step-by-step internally, but show only the specified outputs.**
I’ve added this to my The Ultimate ChatGPT Prompt Library for B2B Marketing Leaders notion doc. Check it out for 55+ more prompts.
Know someone who might find this prompt useful? Share it with them!
Tool of the Week: Agentic Research Engine
GenSpark calls itself “AI-powered search engine,” but I feel like it’s selling itself short. I’ve found that it's more of an agentic workspace that turns a single prompt into deep research, briefs, decks, docs, and media, so you can go from idea to packaged assets in one place.
This can run web auto-research and orchestrate tasks across tools.
Here are some other ways to use GenSpark:
Auto-research a market/account and return a cited brief
Convert a whitepaper or PDF into a polished slide deck for sales enablement
Generate ad/social visuals and short video cuts to A/B test creative angles fast
Build simple ROI calculators, campaign trackers, or content calendars in AI Sheets
Use my referral link and we both get 1000 additional credits. Or just go directly to Genspark.ai
AI Resource Roundup
Is Role Prompting Effective? Clear take: roles mostly shape tone and style, not accuracy, on modern models. For marketers: lock brand voice with roles, then use task-specific prompts and evaluation when correctness matters.
n8n starter walkthrough: My favorite Clay influencer, Patrick Spychalski show how n8n wires LLMs to your stack with drag-and-drop flows, templates, and API nodes so you can ship GTM automations fast. It’s a little in the weeks, but it’s good.
How to Test AI Models: Christopher S. Penn’s guide to building your own test suite tied to real use cases, plus why to test in both the UI and the API with identical prompts for apples-to-apples results. For marketers: create a 5-prompt go or no-go pack for ad copy, briefs, and analytics summaries, then re-run it whenever a model changes.
Hot AI Jobs
Here are three stand-out roles at the intersection of marketing and AI, perfect if you're looking to shape AI-fueled strategies at growing B2B SaaS companies:
Product Marketing Lead, AI at Snowflake
Location: Menlo Park, CA
Pay: 175K to 249.9K per year, full-time
Senior Director, Product Marketing, AI at Airtable
Location: San Francisco, CA
Pay: 308K to 360K per year, full-time
Product Marketing Manager, AI at Sentry
Location: San Francisco, CA, hybrid
Pay: 150K to 190K per year, full-time
Wow, I feel like that was a lot of product marketing this week, but I hope you liked it!
If this helped you out-execute the AI hype cycle, do a fellow marketer a solid and tap that forward. Here’s to stacking smart workflows and scaling real pipeline. See you in the next drop.
– Brandon