Even if copilots replace search engines as the main discovery interface, they may not reproduce the organic referral traffic marketers previously received. Reliable answers can satisfy users without a click, while the opaque behavior of large language models makes consistent ranking optimization difficult.

Copilots can answer questions without requiring website visits
Traditional links may become a secondary verification tool
Model recommendations are difficult to reverse-engineer
Organic optimization may not transfer cleanly from search engines to LLMs

“the answers are so reliable that it answers the question within the co-pilot experience so you never have to click on anything to go to…”
Kieran Flanagan · 11:30

“no one really knows why the llm model picks the answer it does”
Kieran Flanagan · 12:30

#organic traffic#search disruption#llm discovery#seo

Hot Take· 3

Hot Take09:00

AI Could Insert Synthetic Host-Read Ads Into Videos in Real Time

Future video systems may reproduce a host's voice and appearance from uploaded footage, then generate advertiser-specific reads inside existing videos. The insertion could be created at viewing time from an advertiser's script and bid, extending programmatic advertising into synthetic host endorsements.

Video may become a native response format inside copilots
Models could reproduce hosts from existing footage
Advertisers could bid for dynamically generated insertions
The ad script and synthetic performance could be assembled in real time

“they're going to create real time ad insertions that are essentially host R into those videos”
Kip Bodnar · 09:30

“the AI will be able to create the ad from the script that that third party uploads”
Kieran Flanagan · 10:30

#synthetic media#video ads#programmatic advertising#ai video

Hot Take21:00

Content Teams May Need Creator-Style Pay and IP Contracts

The hosts argue that companies will need channel specialists rather than treating every content role as generic blogging. Compensation may combine salary with channel-growth incentives, while contracts explicitly determine whether a departing creator can take a show or must sell their interest back to the company.

Podcasting, YouTube, and short-form video require distinct expertise
In-house creators may receive performance-linked compensation
Contracts must clarify ownership of creator-led shows
Companies may need buyout mechanisms for successful content properties
Creator identity and company intellectual property currently overlap awkwardly

“they actually have to set their content teams up like a team of creators”
Kieran Flanagan · 21:00

“the way that you set up your contracts to work with creators in the future will be the most important part of your content strategy”
Kieran Flanagan · 22:00

#creator economy#content teams#compensation#intellectual property

Hot Take27:30

AI Could Make Product Bundling and Dynamic Pricing Routine

The hosts briefly predict software that makes cross-brand bundles easier to construct at checkout, including in B2B markets. They connect this to a broader future in which AI supports dynamic pricing, packaging, and commercial combinations that are difficult to operate manually today.

Consumer-style product bundles could expand into B2B
Software may simplify partner integration at checkout
AI could support dynamic pricing and packaging
Pricing expertise remains important even if execution becomes easier

“in the future what happens is we get software that helps us to bundle more easily with other brands”
Kieran Flanagan · 27:30

“it's G be way easier to bundle do Dynamic pricing”
Kip Bodnar · 28:00

#bundling#dynamic pricing#packaging#b2b

Explainer· 3

Explainer04:00

Why AI Copilots Could Become Affiliate Marketing 2.0

The hosts predict that similar foundation models will push copilots toward subscriptions supplemented by advertising. AI could assemble offers, forms, checkouts, targeting, and attribution inside the copilot, effectively acting as a real-time affiliate that brokers customer introductions.

Model similarity could create price competition among copilots
Advertising may supplement subscription revenue
AI could generate complete transactional ad experiences inside the copilot
Brands may compete through bids and budget more than execution quality
Learning affiliate marketing now could prepare teams for the emerging format

“this is like affiliate marketing 2.0”
Kip Bodnar · 06:30

“AI is the affiliate actually for everyone”
Kip Bodnar · 06:30

#ai copilots#affiliate marketing#advertising#monetization

Explainer12:00

The Emerging Black-Hat Tactic of LLM Model Poisoning

Kieran describes an attempted analogue to old link-spam tactics: flooding the web with content that characterizes a brand in a desired way before a model's next training cycle. The intent is to influence how future models perceive or recommend the brand, although the hosts do not present it as a reliable or sustainable practice.

The tactic mass-produces favorable brand descriptions
It targets future model training rather than a live search index
Its effectiveness is uncertain because model outputs are opaque
It resembles earlier attempts to manipulate search rankings

“there's this thing called poisoning which is basically create content on mass to influence the AI llm model”
Kieran Flanagan · 12:00

“you want to poison the web with a huge Avalanche of content that describes your product and brand in certain way”
Kieran Flanagan · 12:30

#model poisoning#black hat ai#llm training#brand manipulation

Explainer28:30

The AI CRM Could Make More B2B Purchases Touchless

An AI-powered CRM could collect signals across the customer journey and personalize video, email, chat, and website experiences. If those interactions become sufficiently useful, customers may accept AI mediation and involve a human salesperson only when the transaction genuinely requires one.

AI can clean CRM data and analyze complete journeys
Personalization can span video, email, chat, and websites
Marketing may control more of the pre- and post-sale experience
More B2B transactions could happen without a salesperson
Marketing skills may spread into sales and customer success

“in the future marketers own more the customer journey and they have an AI powered CRM”
Kieran Flanagan · 28:30

“more B2B transactions go touchless because why talk to a human unless you need to talk to a human”
Kieran Flanagan · 29:30

#crm#personalization#touchless sales#customer journey

Takeaway· 2

Takeaway25:00

When Code Gets Cheap, Leadership and Execution Become the Constraint

The discussion predicts that AI-assisted development will release a large backlog of viable marketing and product ideas previously blocked by scarce engineering capacity. As technical implementation becomes easier, the differentiator shifts toward selecting worthwhile ideas, leading projects, refining details, and sustaining iteration.

Organizations already have more ideas than available engineers
AI may supply additional engineering capacity without adding headcount
Cheaper implementation increases the number of ideas teams can test
Human leadership and persistence become more important as technical barriers fall

“we've lived in an era where there's more ideas than Engineers”
Kieran Flanagan · 25:00

“it's just going to come like down to a human's ability to lead and see that project through and execute it and iterate on the…”
Kip Bodnar · 26:30

#ai coding#execution#leadership#innovation

Takeaway36:00

AI May Shift Marketing's Advantage From Tooling Back to Craft

Data, analytics, tooling, and marketing science accelerated careers in the previous era, but AI is increasingly capable in those areas. The hosts expect positioning, message-making, building, and craft to regain importance because AI can assist with analysis more readily than it can originate and finish distinctive work.

Analytical and tooling skills powered the previous marketing era
AI is well suited to data analysis and easier software operation
Positioning, messaging, and creative construction become stronger differentiators
AI-generated material still requires human judgment before shipping
Future marketing leaders need both craft and the ability to build

“AI is incredible at data analysis AI is going to make Tulin much much easier”
Kieran Flanagan · 37:00

#marketing careers#creative craft#ai skills#positioning

Frameworks in this episode

Emotional-Cue Segmentation

Multimodal Storytelling Advantage

Single-Use Premium Tool Loop

Insights & moments

⚡Myth Buster· 1

Why AI Copilots May Not Replace Lost Search Traffic

◆Hot Take· 3