I’ve spent decades analyzing shifts in marketing, but the rise of AI agents is the most disruptive one yet. These systems are quickly taking over the user journey — but unlike humans, they evaluate structured data, analyze backend specifications, and make decisions in milliseconds.
Major players are already taking note and adapting. For example, Adobe recently introduced AI agents that brands can use to help consumers navigate through their websites. Businesses can then enable personalized marketing based on real-time user behavior and unique customer attributes.
From a practical perspective, however, what does it actually look like when AI agents take over key stages of the user journey? And, what does that mean for marketers? To find out, I put three leading AI agent-powered tools to the test by assigning them real marketing tasks. Here’s what happened and what every marketer needs to do next.
What are AI agents, and why are they important in marketing?
AI agents are autonomous systems that research, analyze, and take action on behalf of users. Unlike traditional AI-powered tools that assist with isolated tasks, AI agents actively manage workflows, interact with software, and execute complex processes — without direct human input.
In marketing, these agents are quickly becoming the new gatekeepers, deciding which brands, products, and services get surfaced and used. Instead of persuading consumers directly with creative campaigns or paid advertising, marketers must now optimize for AI-driven decision-making. So, brand content needs to be structured, clear, and machine-readable.
3 Real-World Examples of AI Agents in Marketing
AI agents are already changing how marketing works. Here’s how three leading tools are taking over research, onboarding, and execution — and what that looks like from a user perspective.
1. Gemini Deep Research: the End of Customer Discovery
What’s one of the most overlooked changes in marketing? The customer research and discovery phase is slowly vanishing. That’s not because customers are skipping it, but because AI agents are doing it for them.
To test this, I used Google’s Gemini Deep Research, part of the Gemini 2.0 Suite, and asked it a simple question: How do I add a chatbot to HubSpot’s website? Instead of giving me a list of links or summaries, Gemini scanned 37 websites, synthesized the steps into a single tutorial, and delivered it in a format I could instantly use. No ads, no searching, no clicking around.
This change is subtle — but it means that if you’re still optimizing solely for human eyes, you’re risking irrelevance. AI agents don’t browse your blog or evaluate your brand voice; they look for verifiable information and clear, objective reliability. Even a superior product can be overlooked if your content isn’t presented in a way that agents can parse and evaluate.
How to Stay Ahead
- Format your content for extractability by using structured headers, ordered steps, and scannable product summaries.
- Create citation-worthy documentation by ensuring your product information is clean, factual, and consistently described across all digital properties.
- Choose product clarity over cleverness by skipping branded language or marketing jargon in favor of clear answers to functional product questions.
2. Google Stream Realtime: the AI Onboarding Partner
While Gemini Deep Research is reshaping how product information is discovered and gathered, Google Stream Realtime, part of Google AI Studio, is changing the way users learn to use a product.
Continuing with my experiment to add a chatbot to HubSpot’s website, I tested Stream to see how it would assist me in navigating the setup process. Instead of directing me to a help article, Stream observed my screen, analyzed my actions, and provided real-time, step-by-step guidance. Every recommendation was context-aware, adapting to exactly where I was in the process.
What stood out was that Stream doesn’t just react to inputs — it anticipates needs. As I navigated the interface, Stream learned how I was interacting with different elements and adjusted its guidance accordingly. This creates a continuous feedback loop where Stream teaches users while learning from their behavior. Onboarding then becomes more efficient and personalized.
How to Stay Ahead
- Clearly define and map out user journeys that follow a predictable structure and design flow, so AI agents can walk users through each step without confusion.
- Label interface elements clearly and consistently by making buttons, menus, and actions machine-readable with standard, unambiguous language.
- Provide structured help content like tooltips, walkthroughs, or embedded tutorials that AI agents can surface instantly when guiding users through your product.
3. Claude AI: Extending agent capabilities through tool integration
Claude AI, developed by Anthropic, represents the next step in AI agent capabilities through its Model Context Protocol (MCP), which allows the agent to utilize external tools and operate with greater independence.
For example, you can give Claude access to tools like Brave Search, productivity apps, or CRMs through secure connections. Once authorized, it can pull reports, generate content, trigger workflows, or even connect data across platforms — all without the user lifting a finger.
During the experiment, I found it particularly exciting that the agent doesn’t hand off tasks to the user — it completes them on the user’s behalf. This means your product needs to be accessible by both human users and AI operations.
How to Stay Ahead
- Build clear, well-documented APIs (like Hubspot’s API Guides) so agents like Claude can understand and act on your product’s capabilities.
- Use secure authentication tools like OAuth to allow safe agent access without compromising user trust or requiring constant credential re-entry.
- Structure your product and data for automation by creating well-defined actions, clear data models, and reliable feedback mechanisms that enable agents to trigger, complete, and validate key operations.
Adapting to Agents
AI Agents don’t scroll through your website, engage with your ads, or respond to emotional storytelling. Instead, they look for structured information and gather those insights for a human user.
Cutting-edge companies will restructure their approach to marketing so they resonate with AI gatekeepers and get their offerings in front of human decision-makers. If you haven’t begun experimenting with AI agents, it’s time to dive in, or you might just get left behind.
To learn more about how AI agents are reshaping your marketing strategy, check out the full episode of Marketing Against the Grain below:
This blog series is in partnership with Marketing Against the Grain, the video podcast. It digs deeper into ideas shared by marketing leaders Kipp Bodnar (HubSpot’s CMO) and Kieran Flanagan (SVP, Marketing at HubSpot) as they unpack growth strategies and learn from standout founders and peers.