09 March 2023

OpenAI’s ‘Foundry’ Release Will Change Marketing Forever...

1Frameworks

10Insights

Frameworks in this episode

Strategy4 steps

Back-to-Front AI Adoption

Apply AI to existing customers before expanding into acquisition.

Read →

Insights & moments

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

Myth Buster· 1

Myth Buster28:30

More AI Conversation Does Not Automatically Improve Conversion

Kieran cautions that open-ended chat can send new visitors down an information rabbit hole when a conversion page should make the next action obvious. Kipp expects AI eventually to become a better facilitator, but agrees that current tests must optimize for outcomes rather than novelty.

Conversion depends on quickly clarifying the next action.
Open-ended chat may create distraction.
Engagement time is not the same as conversion success.
AI interfaces should be trained toward a defined customer outcome.

“I think it could be a distraction for people.”
Kieran Flanagan · 29:00

“But what you're basically saying is like conversion rate, optimization is about facilitating versus distracting.”
Kipp Bodnar · 29:30

#conversion optimization#chat#websites

Hot Take· 2

Hot Take11:00

The Bigger AI Shift May Be How People Operate Software

Kipp argues that generative AI is not merely changing search or content creation. Prompt-driven interaction may replace portions of the graphical interface across many software categories, changing what users expect from applications.

AI changes interfaces as well as outputs.
Text prompts can replace navigation through complex menus.
The shift is expected to extend beyond search products.

“Is that the thing that people are not talking enough about is how AI will change how we use technology and use software.”
Kipp Bodnar · 11:30

“And the move from a graphical user interface to a more text prompt based interface.”
Kipp Bodnar · 11:30

#user interface#software#chat

Hot Take25:30

When AI Makes Great Customer Experience a Commodity

If AI enables most companies to deliver responsive, low-friction customer experiences, experience alone may stop being a durable differentiator. Kipp argues that companies will then need to accelerate product innovation or pass lower operating costs to customers.

AI may lift the baseline customer experience across competitors.
Common AI tooling could make experiences increasingly similar.
Product innovation offers a stronger differentiator than price competition.
Automation-driven margin gains may place downward pressure on prices.

“And if that's the case, then you're not gonna be able to differentiate on great customer experience.”
Kipp Bodnar · 27:00

“I think it comes down to either your pace of product innovation has to greatly accelerate or your costs have to come way, way down.”
Kipp Bodnar · 27:00

#customer experience#differentiation#pricing#innovation

Explainer· 3

Explainer04:00

AI Replaces Tasks Before It Replaces Entire Jobs

The hosts distinguish jobs from the repeatable tasks that compose them. Automation can remove enough tasks to transform a role or sharply reduce the number of workers needed, even when the occupation itself continues to exist.

Jobs are bundles of repeatable tasks.
AI can alter a role without eliminating its title.
Farm automation illustrates how worker demand can fall faster than industry activity.

“And I think what Nathan made a really good distinction on, is it doesn't replace jobs, but replace tasks. Like repeatable tasks.”
Kieran Flanagan · 04:00

“So it doesn't replace an entire job. But it replaces a lot of tasks that are part of a job.”
Kieran Flanagan · 04:30

#automation#jobs#future of work

Explainer10:00

AI Will Combine Content Creation, Research, and Localization

The hosts expect newer language models to improve the complete content workflow rather than writing alone. Research, repurposing, and translation become connected capabilities, enabling marketers to adapt material across formats and markets more quickly.

AI writing quality is expected to improve substantially.
Research becomes part of the same content workflow.
Localization can make existing content useful across more markets.
Repurposing expands the value of source material.

“So like creating content, repurposing your content, researching your content, and actually translating your content.”
Kieran Flanagan · 10:30

“One of the ones, that I hadn't really even spent time thinking about is, how quickly you can actually localize all of your content through…”
Kieran Flanagan · 10:00

#content#localization#research#gpt-4

Explainer12:00

Proprietary Customer Data Is the Real AI Unlock

Public content can help a model imitate a company's writing, but the hosts expect much greater value when AI can use private customer and product context. That combination can make responses, recommendations, and experiences specific to a person's actual needs.

Public web data limits differentiation.
Customer history can improve relevance.
Product expertise and customer context are complementary inputs.
Private data requires controlled, company-specific model access.

“But you know what's gonna be way better. Way more important, is when you train that model with all of the data that you have…”
Kipp Bodnar · 12:30

“So like it's an expert on your product, it's an expert on the customer.”
Kieran Flanagan · 20:00

#customer data#personalization#foundry

Story· 2

Story18:00

HubSpot's AI Chat Test Beat Rule-Based Automation

HubSpot connected GPT-3 to its automated chat experience instead of relying solely on static rules. Kipp reports materially stronger engagement and conversion, while preserving the pattern in which automation can transfer a conversation to a person.

Effective chat often begins with automation and escalates to a human.
Language models reduce the limits of rigid rule trees.
HubSpot observed higher engagement and conversion from AI-enabled chat.
The experiment required API development expertise.

“And what we found is a much higher conversion rate and engagement rate for folks who were interacting with the AI enabled chat versus a…”
Kipp Bodnar · 19:00

“It was like the results that the AI returned were exceptionally good, and the engagement from the users on the chat was exceptionally good.”
Kipp Bodnar · 19:30

#chatbots#conversion#hubspot#experiments

Story24:00

AI Sales Emails Can Recommend the Right Product Package

HubSpot tested automated sales emails that went beyond inserting a recipient's name and employer. AI used role and company context to recommend a particular product package and explain its relevance.

Traditional personalization relies on shallow merge fields.
AI can recommend a product rather than mention one generically.
Messages can explain value in the context of the recipient's role.
The same capability can support sales and customer-success teams.

“One of the things that happens in an AI world, those messages get so much more contextualized.”
Kipp Bodnar · 24:30

“And like that level of personalization and recommendation that can be delivered through these AI models is exceptionally powerful”
Kipp Bodnar · 25:00

#sales emails#personalization#recommendations

Takeaway· 2

Takeaway07:30

AI Productivity Gains Depend on Rapid Worker Reskilling

Kipp challenges the assumption that automation automatically improves life for every worker. He argues that technological disruption creates new opportunities unevenly, making accessible mid-career reskilling essential.

Automation benefits are not distributed evenly.
Displaced workers need practical routes into newly demanded roles.
Traditional education systems are poorly designed for mid-career transitions.

“You need the ability to re-skill your workers at a really fast rate.”
Kipp Bodnar · 08:00

“We built this really rigid education system, that was really good for people coming right outta high school. But was really poor for people who…”
Kipp Bodnar · 08:30

#reskilling#education#labor

Takeaway22:00

AI Shrinks the Marketing First-Mover Window

Early adopters of an effective marketing strategy can gain disproportionate benefits, but rapid AI tooling development shortens that advantage. Teams must learn and capture value quickly because successful tactics will be copied and commoditized faster than before.

Effective new channels produce outsized early returns.
The duration of an advantage depends on innovation speed.
AI accelerates imitation and commoditization.
Being merely early may no longer be early enough.

“Is there is a huge first mover advantage in marketing.”
Kipp Bodnar · 22:00

“We are in a period where you have to literally be very first and get all of the benefit you can because it's going to…”
Kipp Bodnar · 22:30

#first mover#competitive advantage#innovation