A Month with Muse
In September Meta released Muse, a consumer-friendly AI assistant built for everyone. Initial response was strong with over 5 million downloads of the app in the first month. This outpaced OpenAI, Grok, and Claude in adoption rate. This rapid adoption is also rapidly changing the way consumers find their new favorite brands, restaurants, and widgets. We’ll get into what you should be doing to make your brand more Muse-friendly. But first, let’s walk through the human experience.
Help or hype?
I represent one of those 5,000,000 Muse downloads. I’ve spent the past month getting to know the app and it is impressive. Muse’s ability to navigate mundane web tasks does make it very useful. I have used it to book appointments, secure elusive restaurant reservations, and help plan an itinerary for a recent trip to California. Muse was able to help with my rental car and hotel on that trip, as well as reminding me to leave in time to make a dinner reservation.
There are definitely still some sticking points. Not all websites are Muse-friendly. So the human has to help Muse log in, which can defeat the point of an assistant. But for the most part, with some minimal training, Muse can navigate most sites.
Personal Shopper
As I became comfortable with Muse navigating the web for me, I contemplated the next step. Was I ready to hand over my credit card for Muse to make purchases? This is a big selling point for Muse. It claims to be able to find the best deals on the web and have your purchases delivered to your doorstep with a quick approval from its human boss.
This is actually causing a bit of tech drama. Amazon has cut Muse off as it fears losing the revenue created by cross-selling impulsive humans browsing the site. Also, your Muse agent can look past all the ads that Amazon has sold into its site. It will be interesting to see how other retailers react. For now, it seems that Walmart is embracing Muse agents.
Despite lingering concerns with Meta and user data, I finally acquiesced and gave Muse access to my credit card to book an appointment requiring a form of payment. Financial connections are made through the Link system, which does make everything feel a little more secure.
Thinking Like an Agent
AI chatbots have already transformed internet search results and brand visibility, driving the need for Answer Engine optimization (AEO). You can read more about that here. But AI agents and personal assistants are a different breed of digital entity that should be thought of in a new way.
The three main differences between established AI search summaries and Agentic search are:
- Personal Context
Agentic recommendations are tailored to what the agent already knows about its human. In the case of Muse, it also has access to the preference data that Meta has collected on a particular user. For instance, an agentic recommendation for a new skillet may differ between a 50-year-old culinary enthusiast and a college student moving into their first apartment. - Proactive
Agents can make recommendations before the user makes a request. For instance, a published discount on a product aligning with user activity may trigger an agent to suggest it. - Operational Ability
This is a big one. Where traditional SEO/AEO bots simply crawl the web and read the code, agents have the ability to open a browser and interact with the web. This is an important distinction as these agents now have the ability to open drop downs, fill out forms, and complete transactions.
How Marketers Should Prepare
The good news is that the Answer Engine Optimization work you’re already doing (right?) to improve brand visibility for LLMs and AI search is also the foundation for establishing agentic awareness. A good AEO platform helps you create and refine content to ensure the bots and agents are not only finding your brand, but also assigning it credibility and preference.
Once this baseline visibility is established, there are some further steps to fine-tune your brand for Muse and the flood of personal agents to come.
- Make sure your site is ready.
Agents can autonomously interact with your live site. So it’s important to remove any agent roadblocks. The easiest way to do this is to experiment with an agent and ask it to directly interact with your site: browse and compare products, fill out forms, complete a purchase. This will help identify any obstacles that an agent may face. - Apply user context where you can.
Is your product geared for an experienced user (50-year-old culinary enthusiast) or a beginner (college student grilled cheese enthusiast). Characterize your product or service with as much user context as possible. Side-by-side comparisons are a great tool to help an agent decide which product is best for their user. - Make your catalog available.
If your site uses a catalog system such as Shopify or WooCommerce, make sure that the catalog is connected to both Meta and Google. This ensures that your products are readily available for agent consideration.
