24 April 2025

Using ChatGPT o3 to Plan our 2025 Marketing Campaign

2Frameworks

8Insights

Frameworks in this episode

Marketing7 steps

Deep Research-to-Creative Campaign Pipeline

Turn researched business context into campaign-ready creative through rapid iteration

Read →

Marketing7 steps

One-of-One Marketing Model

Personalize each campaign asset to one company instead of one broad segment

Read →

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Hot Take· 3

Hot Take03:30

Near-Zero Creative Cost Makes Disposable Originality Possible

AI image generation changes creative economics by making unusual visual ideas cheap enough to create once and discard. The hosts argue that marketers no longer need to rely as heavily on reusable templates when each video or campaign can contain original imagery.

Template reuse was partly a response to production cost
AI makes unique imagery practical for individual pieces of content
Distinctive B-roll can help short-form video stand out
Generated images can also be animated with other AI tools

“the cost of creative goes to near zero.”
03:30

“you can actually have everything one of one, and then just throw it away and do something brand new for the next time.”
04:00

#creative-economics#originality#short-form-video#ai

Hot Take20:30

Sora Is Still at the Early DALL-E Stage of Video Generation

After an unsuccessful image-to-video experiment, the hosts compare Sora's current maturity with early DALL-E. They expect video quality to improve but argue that computing resources and model development currently favor image generation.

The demonstrated video transition fails to preserve a coherent concept
Current video generation resembles an early product stage
GPU allocation may constrain rapid video improvement
Image generation currently offers more reliable marketing value

“Oh, God, hearing sore is Dolly.”
20:30

“That's where we are with video.”
21:00

#sora#video-generation#dall-e#ai-models

Hot Take25:00

Using More AI Can Increase Demand for Better AI Tools

The hosts reject the idea that learning about AI naturally reduces human participation or demand. Their experience is that each new capability creates more ideas, more experimentation, and stronger demand for APIs, reusable GPTs, and better production tools.

New capabilities generate additional use cases rather than satisfying a fixed demand
Rapid experimentation makes one idea build on another
Users want to consolidate successful workflows into reusable agents
API access would let many external tools adopt the new image capability

“the more you learn is crazy.”
25:30

“The more I learn about AI, the more I want to use it”
25:30

#ai-adoption#experimentation#gpts#api

Explainer· 1

Explainer01:00

Five Non-Obvious Ways Marketers Can Use AI Images

The hosts review five image-generation concepts proposed by o3, ranging from hyperlocal advertising to interactive lead-generation carousels. The examples show how AI can apply brand context to formats that previously required substantial creative production.

Generate hyperlocal ads using landmarks, slang, street art, and weather
Show enterprise prospects a personalized future state
Prototype positioning through movie-style concept posters
Use historical-era product imagery as distinctive short-form video inserts
Build branching visual carousels around a reader's chosen problem

“people click faster on ads that make them feel that feel made for their neighborhood”
01:30

“before you spend a bunch of money on doing full video shoots actually generate movie style posters.”
02:30

#image-generation#creative-ideas#advertising#b2b

Tool· 2

Tool05:00

ChatGPT Memory Can Preserve Brand Context Across Creative Sessions

The hosts distinguish ChatGPT's cross-chat memory from project document storage. They argue that remembered conversations can improve outputs and shorten prompts because the model can reuse previously developed style context.

Memory covers prior chats but not necessarily every uploaded document
Remembered brand context can improve later creative outputs
Prompts can explicitly reference a previously developed style guide
Projects remain useful when shared document context is required

“It's only all of your chats.”
06:00

“it's changed how I prompt.”
06:30

#chatgpt#memory#brand-context#prompting

Tool13:00

Why ChatGPT's Image Library Acts Like a Lightweight Design Studio

The image library collects generated visuals in one place and links them back to their original conversations. It supports basic editing, revision, and visual comparison, making it easier to manage a large volume of campaign experiments.

View generated images together instead of searching individual chats
Open the original thread when an asset needs further iteration
Perform basic text-directed image edits
Compare visual consistency across a brand's generated assets
Starting from the library may improve multi-image workflows

“now you can have a library of all of the images you've made in one place, in chat, GPT”
13:00

“this just looks like a creative library that a team have worked really diligently on.”
19:00

#chatgpt-library#image-editing#design-workflow#creative-assets

Takeaway· 2

Takeaway17:00

AI Image Generation Is Strongest at Prototyping—For Now

Despite impressive creative quality, the hosts encounter unreliable variation handling and limitations around precise final production. They position the current tools as excellent for prototyping, lightweight assets, and campaign tests rather than every high-stakes deliverable.

Requests for separate variations may produce one combined image
Repeated rendering is sometimes needed to obtain the desired result
The tools are already adequate for many small-business ads
High-stakes assets still need human production review
Real campaign tests can reveal whether AI work matches specialist creative

“It's still not quite there.”
17:30

“I think prototyping is the best use of it right now.”
18:00

#prototyping#limitations#creative-testing#image-generation

Takeaway22:00

o3 Works Best as a Context-Rich Creative Thought Partner

The hosts conclude that o3 can contribute strategic ideas as well as execution prompts, especially for unfamiliar features or open-ended marketing problems. Its usefulness rises sharply when the marketer supplies rich business and brand context.

Use o3 for creative inspiration when exploring a new capability
Provide information about the business before requesting ideas
Ask for execution prompts alongside the ideas
Treat generated prompts as first drafts that can still be improved
More context produces more specific and useful recommendations

“03 is a really great marketing assistant.”
22:00

“the more context you give it, the far better the results.”
22:30

#o3#thought-partner#marketing-strategy#prompting