The host asks ChatGPT to study existing HubSpot presentations and produce a written style guide. The resulting guide covers typography, colors, layouts, imagery, voice, and detailed formats for recurring slide types.

Existing public decks provide reference material for the research.
The guide identifies typefaces, weight, hierarchy, and color palette.
It describes layouts for agenda, problem, solution, data, and summary slides.
A company can create a useful presentation style guide in minutes.

“I had it just do a deep research project on existing HubSpot slide decks out on the internet.”
Host · 02:30

“it actually did a written style guide for the different formats and types of slides that would be in a deck.”
Host · 03:00

#chatgpt#style guide#branding#pitch decks

Explainer11:30

The Stronger Deck Connects ROI, Deployment, and Trust

The second deck moves beyond attractive design by presenting operational and commercial evidence. It explains setup, data connections, channel deployment, escalation, knowledge sources, security, and a concrete pilot offer in a sequence suitable for a real sales conversation.

The deck presents cleaner ROI data.
It promises a fast, no-code deployment path.
It explains how the agent uses knowledge bases, documents, PDFs, and websites.
Security and GDPR readiness address buyer concerns.
The final slide gives prospects a specific 30-day pilot action.

“Go live this week. No code required. Connect your data, train in minutes, deploy to your channels, escalate and learn.”
Host · 11:30

“It's GDPR ready, it's fully secure.”
Host · 11:30

#customer support#roi#sales enablement#ai agent

Story· 3

Story04:00

What Happened When GenSpark Researched and Wrote the Whole Deck

In the first test, GenSpark receives a short audience-and-product prompt plus the HubSpot style guide. It independently researches the customer-agent product and creates the slide content, producing a usable but imperfect branded deck.

The prompt specifies the product, audience, deck length, and style guide.
GenSpark performs the product research itself.
The output captures customer pain, benefits, ROI, and calls to action.
Brand fonts and colors are accurate, although some layouts are awkward.
The tool generates eight slides despite being asked for six.

“Gen Spark not only built the slides, but it went and did all the research and came up with all of the content for these…”
Host · 04:30

“But the fonts are right, the colors are right. It picked up the key features and benefits of the product really well.”
Host · 05:00

#genspark#sales deck#hubspot#ai research

Story09:00

The Deck Improved When O3 Wrote the Outline First

For the second test, the host has ChatGPT O3 prepare the slide structure, copy, design directions, and speaker notes before handing the material to GenSpark. This division of labor produces a stronger title, clearer value proposition, better pain framing, and a compelling product demonstration slide.

O3 supplies the outline, copy, design elements, and speaker notes.
The same researched style guide anchors the visual output.
The resulting deck communicates the core value proposition more clearly.
A simulated customer-agent interaction makes the product tangible.
The output still contains a small visual hallucination.

“I had it write its own outline.”
Host · 09:00

“I had ChatGPT give me the slides, the design elements, the slide copy, even speaker notes, everything that you would need to write a killer…”
Host · 09:00

#o3#chatgpt#pitch deck#ai workflow

Story13:30

A Google-Inspired Prompt Produced a Very Different Third Deck

The host converts Google's 60-page prompting guide into a reusable prompt-engineering assistant and uses it to generate a third deck. That version emphasizes the customer problem, annotated product visuals, results, and pricing, but omits the custom style guide and ultimately ranks below the O3-led second version.

Google's prompting guide is condensed into a prompt-engineering assistant.
The third deck leads with an overwhelmed-support problem.
It combines problems and solutions on one slide.
It highlights resolution rate, customer outcomes, and transparent pricing.
The host prefers the second deck created from O3's explicit outline.

“Basically, Google created a full 60-page guide on prompting, and I used O3 to turn that into a prompt engineer in a box.”
Host · 13:30

“I think I like the second one better.”
Host · 15:30

#prompt engineering#google#deck comparison#ai prompting

Tool· 2

Tool00:30

Why GenSpark Stands Out Among AI Slide Generators

GenSpark combines multiple AI models with image generation and a dedicated slide-creation feature. After testing roughly seven AI slide apps, the host considers it the first capable of producing a credible first draft for practical use cases.

GenSpark is designed as an AI super agent.
Its backend uses models from OpenAI, Anthropic, Google, and DeepSeek.
The slides feature can research a topic and generate a complete presentation.
The host rates it above the other AI slide tools he tested.

“I have tried about seven different AI slide apps, and none of them up to this point were good enough that made me feel like…”
Host · 01:00

“it is the closest, and I think it is ready for some use cases”
Host · 01:30

#genspark#ai slides#presentations#ai tools

Tool05:30

How GenSpark Handles Slide Edits and PowerPoint Export

GenSpark supports conversational revisions and selection-based editing of slide elements. Users can alter text and fonts directly, ask the agent to repair selected areas, or export the result to PowerPoint for changes the interface cannot handle.

Users can request revisions through natural-language prompts.
Select-to-edit supports text, font, deletion, and duplication changes.
Element repositioning appears limited in the demonstrated interface.
Generated decks can be exported as PDF or PowerPoint.
PowerPoint provides a fallback for unresolved layout problems.

“you can type and ask it to update the slides by typing.”
Host · 06:00

“you can export it to PowerPoint and do some basic manual editing in PowerPoint as well”
Host · 08:00

#slide editing#powerpoint#genspark#workflow

Takeaway· 3

Takeaway07:00

What GenSpark's Free Tier Actually Covers

The free version appears sufficient for approximately one or two presentations, depending on their length. Beyond that allowance, the next tier costs $20 per month, which the host frames as reasonable when it replaces hours of manual presentation work.

GenSpark offers a free version.
The free allowance may cover one or two presentations.
Longer or repeated deck creation is likely to require an upgrade.
The demonstrated paid tier costs $20 per month.

“About one or two presentations you can do for free, depending on how long they are.”
Host · 07:00

“20 bucks is not crazy for saving hours and hours of presentation time”
Host · 07:30

#genspark#pricing#free ai tools#productivity

Takeaway12:00

A Strong On-Brand Deck Took Only 20 to 30 Minutes

The host estimates that the style guide, O3 outline, and GenSpark generation process can produce a strong product deck in 20 to 30 minutes. The generated version still needs minor corrections, but it is polished enough to test with a small sales team.

The generated draft needs roughly 10 to 15 minutes of tweaking.
Total production time is estimated at 20 to 30 minutes.
AI generation removes much of the image and layout work.
The result is considered suitable for a limited sales-team test.
The host believes it exceeds the quality of many existing company decks.

“And all in, it will probably have taken me about 20 to 30 minutes.”
Host · 12:30

“I think this looks better than most companies' pitch decks today.”
Host · 13:00

#time savings#sales enablement#pitch decks#generative ai

Takeaway16:00

Why Detailed Context Beats a Bare AI Prompt

Across three generated decks, the best result comes from combining brand constraints, a researched content outline, and a dedicated slide-generation tool. The comparison demonstrates that stronger instructions and clearer source material improve both the relevance and polish of agent-generated work.

The unstyled third deck remains usable but is less brand-specific.
The style guide materially improves visual consistency.
O3's researched outline improves content quality and structure.
GenSpark is most effective as the presentation-production layer.
Clearer instructions help the agent produce better work.

“the more instruction and the more details and clarity you can give an AI agent the better work it can do with you.”
Host · 16:00

“the best version was taking a style guide that I created with O3 in ChatGPT having O3 do a deep research create an outline and…”
Host · 16:00

#prompting#context engineering#ai slides#brand consistency

Frameworks in this episode

AI Deck Quality Scorecard

AI Slide Style Guide Reverse Engineering

Brand-Grounded AI Pitch Deck Pipeline

Prompt Engineer in a Box

Specialist Model Handoff

Insights & moments

✶Explainer· 2

ChatGPT Can Reverse-Engineer a Brand's Pitch-Deck Style