28 May 2024

Ai Agents: The Future Marketers You Can't Afford to Ignore

2Frameworks

11Insights

Frameworks in this episode

Innovation6 steps

Agent Automation Job Map

Map customer jobs, rank automation potential, and deploy agents progressively

Read →

Marketing5 steps

AI Search Geographic Split Test

Compare US and international traffic to isolate AI search effects

Read →

Insights & moments

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

Hot Take· 3

Hot Take05:30

AI Overviews Could Sever Google's Publisher Bargain

Kieran argues that complete answers placed above conventional organic results will reduce the need to visit publisher pages. He expects a substantial decline in click-through rates, extending the traffic effects previously observed with featured snippets.

AI answers occupy the most valuable search-result space
Citations may not generate clicks when the answer is already visible
Featured snippets provide an early warning about click displacement
Organic acquisition strategies face a potentially monumental change

“if you have your answer if you have the question answered why do you ever need to go and actually click on those links”
Kieran Flanagan · 06:30

“I can only imagine the decrease in clickthrough rate on organic listings when the entire top top half of the page is going to be…”
Kieran Flanagan · 07:00

#seo#google#ai overviews#organic traffic

Hot Take16:30

Generic Advice Is the Content Format Least Safe From AI

Multimodal AI can answer questions using the user's exact circumstances rather than publishing one response for thousands of people. The hosts expect agents to combine broadly applicable knowledge with a customized final portion that conventional articles and videos cannot economically reproduce.

AI can work across text, video, and audio
Personalized answers weaken generic how-to content
Most of an answer may be common while a crucial portion is individual
Agents enable personalized automation rather than broad rule automation

“one of the things that's not safe is kind of generic advice”
Kipp Bodner · 16:30

“the last 20% may be very different than what everybody else needed”
Kipp Bodner · 17:00

#personalization#content#multimodal ai#automation

Hot Take21:00

Agents Could Pull More of the Customer Journey Into Marketing

Kieran predicts agents will infer and deliver each prospect's best next step, whether that means answering a question, providing information, or opening a product demo. As these interactions become automatable, marketing may retain ownership further into the journey before handing prospects to another team.

Agents can infer an individual's most useful next action
The next step can be generated and delivered automatically
Demand generation and conversion assistance may converge
Human handoffs may occur much later in the customer journey

“the AI agent is going to be able to figure out how to provide that person with the next step”
Kieran Flanagan · 21:00

“over time more and more of the customer Journey collapses into marketing and marketin owns more of this”
Kieran Flanagan · 21:30

#customer journey#demand generation#marketing automation#conversion

Explainer· 4

Explainer01:30

The Difference Between an AI Agent and a Copilot

A copilot assists a person through repeated prompting, suggestions, feedback, and analysis. An agent instead works toward an assigned outcome with less guidance, executes the underlying tasks, monitors results, and adapts its future work.

Copilots require ongoing human prompting and assembly
Agents receive an outcome rather than a sequence of instructions
An agent can research, create, publish, measure, and improve
The appropriate user experience for supervising agents remains unsettled

“the agent does not need constant guidance”
Kipp Bodner · 03:00

“it actually goes back and starts to look at the data over time and improve itself without you really needing to give it much Braden…”
Kipp Bodner · 03:00

#ai agents#copilots#automation#content marketing

Explainer10:00

Google Is Moving From a Search Engine to an Action Engine

Traditional search finds information and leaves execution to the user. The emerging action-engine model lets an agent acquire information and complete tasks such as bookings or orders through connected applications.

Search historically separated discovery from task completion
Agents push search activity into the background
Google and OpenAI are competing to own the personal-assistant layer
The shift may create new buyer and consumer behaviors

“Google is moving from a search engine to an action engine”
Kieran Flanagan · 10:00

“there is a layer on top of search that pushes the search back that pushes the search into the background so you never really have…”
Kieran Flanagan · 11:00

#search#action engines#google#gemini#ai agents

Explainer17:30

The Marketing Agent Roles Most Likely to Emerge

The episode surveys agent roles spanning email assistance, advertising, analytics, project management, visual storytelling, content, search, and prospecting. Advertising is highlighted as a major shift because AI can produce and manage hundreds or thousands of personalized creative variants.

Email agents can arrange meetings, travel, and files
Advertising agents can generate large volumes of personalized creative
Analytics agents can surface insights from data
Project-management agents can handle menial coordination
Prospecting agents can research individuals and craft outreach

“that's going to move to hundreds or thousands of VAR ad ad and creative variants over the next few years”
Kipp Bodner · 18:30

“you're going to get better conversion rates from it so it's going to be worth managing”
Kipp Bodner · 19:00

#marketing#advertising#analytics#ai agents

Explainer24:30

Why Context Capture May Pull AI Companies Toward Hardware

Agents become more effective when they can observe the context surrounding a user's daily work. Kieran connects browser extensions and desktop applications to a possible move into AI-native phones or laptops that can learn recurring behavior and support agents acting on the user's behalf.

More context enables agents to reproduce complex work
Browser extensions capture information from web activity
Desktop software can potentially observe broader workflows with permission
Hardware ownership could give AI companies a richer context stream

“Agents will get better the more context they get”
Kieran Flanagan · 24:30

“the app can start replicating that and have agents that do that on your behalf”
Kieran Flanagan · 25:00

#context#hardware#desktop ai#ai agents

Takeaway· 4

Takeaway04:00

Why Agents Represent the Next Stage of Business Automation

The hosts position agents as a new automation layer resembling skilled interns that can complete multistep assignments. Early systems will likely remain subject to review and approval until their capabilities and reliability mature.

Agents automate less rigid and more skilled tasks
Initial deployments will retain human review and approval
Business-focused agent innovation is expected to accelerate
Capability will grow with stronger models and better data access

“agents unlock a new layer of automation”
Kipp Bodner · 04:30

“It's almost like you had a skilled intern who can go do a bunch of things for you and then come back and you can…”
Kipp Bodner · 04:30

#automation#ai agents#productivity

Takeaway12:30

Businesses Must Start Designing for Robot Customers

As agents begin taking actions for users, companies will need systems and information that machines can access and interpret. Businesses that preserve human-only processes may deliver a dramatically worse experience than competitors designed for agent interaction.

Customer expectations will shift toward action-oriented experiences
Companies will need machine-accessible data and integrations
Some proprietary back-end information may require safe public interfaces
Serving agents will become part of serving customers

“most companies are going to in some ways become technology companies because they're going to need to make it easy for an agent instead of…”
Kipp Bodner · 13:30

“those companies that are really slow to move towards serving robots in addition to humans are going to become obsolete”
Kipp Bodner · 14:00

#agent commerce#customer experience#integrations#automation

Takeaway23:30

Agents Turn Audience-Level Guesses Into Individual Predictions

Human marketers normally generalize about an entire audience because one-to-one analysis does not scale. Agents can apply those broad hypotheses to each prospect's context, potentially unlocking better personalization and conversion performance.

Marketers usually optimize around aggregate audience assumptions
Agents can reason at the individual prospect level
Better model quality and data access will expand agent capability
Early agents may disappoint before the underlying systems mature

“agents are better at guessing at a onetoone prospect level than a marketer can ever be”
Kipp Bodner · 24:00

“agents are going to be able to take your guess on what the audience cares about and personalize it to that specific person”
Kipp Bodner · 24:00

#personalization#prospecting#marketing#productivity

Takeaway27:30

Start Measuring Whether LLMs Surface Your Brand

Kieran recommends maintaining conventional search optimization while beginning to measure brand visibility inside LLM-generated answers. Even when those answers produce few clicks, repeated inclusion can reinforce the brand across product-related questions.

Traditional organic search remains valuable
LLM search behavior and click-through rates remain uncertain
Brands should inspect where they appear in different LLM engines
Content and other signals may influence future model answers
Visibility without a click can still reinforce brand recognition

“be early to the llm strategy”
Kieran Flanagan · 27:30

“even if someone doesn't click on the link it's still a great reinforcement of brand”
Kieran Flanagan · 28:30

#llm optimization#brand visibility#seo#search