When the math stops mathing: Fandom signals in the agentic AI era

Artists understand their audiences largely through their actions. Following, liking, subscribing, opening an email, buying merch, buying a ticket, waiting in line – each of these efforts creates attachment and signals a different level of commitment. This helps marketers draw the lines between casual listeners and dedicated fans.

But what happens when that friction disappears? As technology makes it easier to automate actions, the audiences behind them may become harder to understand.

This could be the outcome of an emerging category of personal AI assistants. The buzziest so far, Instinct, can be used for tasks like booking hard-to-get restaurant reservations, scheduling tedious appointments, and even negotiating prices with vendors. As brand strategist Barr Balamuth argues in a recent piece for the newsletter Sociology of Business, this is great for consumers but difficult for brands, which are losing the ability to use gated access to forge bonds and segment audiences. Once AI agents can complete unlimited tasks on everyone’s behalf, effort becomes less useful as a signal of fandom.

This is a shift the music industry should pay attention to. Imagine an agent directed to find tickets for Charli XCX under $200 anywhere in London, enter every presale, monitor resale sites continuously, join every waitlist, and buy immediately when something becomes available. RAYE just experienced a nefarious version of this, receiving 1.8 million pre-sale sign-ups for a 5,000-seat show after AI agents created “hundreds of thousands” of fake registrations for ticket brokers (per Billboard). While platforms continue to fine-tune technology to weed out this kind of “bot” behaviour, it’s likely that real fans will increasingly want to use agents, too. Rather than blocking agents entirely, it is more likely that platforms will have to learn to interact with them.

Of course, a fan sending Instinct to wait in a virtual line for Charli XCX tickets still signals higher interest than the average consumer. Yet as these types of hacks become more easily automated and widespread, the audience picture marketers have spent years polishing is about to get blurrier. This is not to mention a further shift yet to come: when agents begin making decisions for users, beyond just fulfilling their requests. For example, imagine an agent decides to hit play on wind-down music at evening-time, before the user even asks for it. The distinctions between actions, decisions, and intentions are getting harder to read.

When the math stops mathing

The real story here goes far beyond Instinct. As algorithms and AI agents increasingly mediate consumer behaviour, it’s generally been getting harder to identify the signals within the noise of user metrics. 

Examples abound. Tour promoters are finding that high social followings do not necessarily translate to high ticket sales. Users don’t even have to follow accounts anymore to be served their content automatically, so many simply don’t. Songs can generate millions of TikTok views without listeners necessarily knowing who the artist is. Algorithmic recommendations drive more consumption, but make it harder to distinguish what audiences seek out from what is simply served to them. The result is a growing disconnect between consumption and commitment – and, perhaps more strikingly, between consumption and cultural capital. This helps explain why, at the same time as streaming subscriptions and listening hours are growing, MIDiA data finds that music is playing a diminishing role in younger generations’ identities.

Measuring relationships, not moments

The music industry has spent the last decade-plus systematically removing friction. While too much complexity can be a barrier to developing fandom, we’re now starting to see why some friction was actually valuable, both to strengthen connections (e.g., playing a vinyl record forges a tangible relationship to music) and to identify and measure fandom. 

As digital behaviour becomes less legible, we are moving from the attention economy to the intent economy. Or, as a recent MIDiA Letter From The MD put it: “We need to develop the mechanisms for understanding what people want, not simply what they do.”

This means moving away from a hyper-focus on measuring individual actions, and towards measuring the strength of relationships. What could that look like in practice?

• Focusing less on isolated actions and more on longitudinal relationships – understanding patterns of behaviour over time
• Prioritising outcomes over engagement signals – not “who signed up for the pre-sale” but “who attended multiple concerts”
• Improving the ways we understand and measure physical, IRL interactions
• Recognising the value of intentional friction – creating experiences that require meaningful participation, rather than optimising every interaction for convenience

In all, this could be a sign for the music industry to soften its obsession with metrics like streams, view counts, and chart positions – or at least, to evolve how they read these signals. The industry has become exceptionally good at measuring consumption, but understanding connection is another matter.