Built for Humans, Consumed by Agents: The Next Decade of Sports Digital Platforms.

Welcome to another edition of The Sports Stack! As promised, I’m being intentionally sporadic with this year’s editions, writing only when I feel strongly about something. This is one of those times. 

The technology sector is evolving rapidly, and entire business models are being replaced with each new AI model release and a move to an agentic world. That is why it’s crucial to consider this question:

What will the future sports digital product actually be? The website and app we've spent a decade building, what are they for, in a world where fans increasingly don't visit them? 

I first wrote about AI agents in sports back in January with TSS #3, and again mid-year after getting hands-on with Perplexity Comet. Both editions ended with predictions. This one is about what sports organisations should actually do. 

This question might seem like a technology-focused issue, but in reality, it spans the entire sports organisation and should be a question for everyone involved in the business. The current sports websites and apps are key enablers for many different revenue models within a sports organisation. 

Quick note before we dive in: this is a long one. I also write everything you read here myself, which is why editions take time to produce (I have a full-time job to balance!). I have started to experiment with AI in my own workflows. After writing my final draft, I use a Claude skill I’ve developed to act as my newsletter editor, helping me verify my data and arguments so the content is as compelling as possible. Happy to share that skill with you. Just drop me a message.

P.S. To back up the point above about the rate of change, here are a few things that have happened in the 2-3 weeks I’ve been researching and writing this edition: 

In recent months, I’ve noticed that the way I consume content and information on the internet has shifted towards a more conversational style. There has also been a convergence in the number of apps and websites I use, with fewer overall and more frequent use of 2-3. I’ve started using my voice much more, using tools like the Perplexity voice shortcut on my phone and experimenting with Wispr to write my newsletter and capture ideas.

Wispr is so cool; you can give Wispr a try here. 

But it’s not just my behaviour that’s changing, and this is an important test; it can be very easy to forget, when you work in technology every day, that you operate in a bubble. Forgetting this can be troublesome, as you might build for yourself rather the people you’re actually building for. Thankfully, the data is starting to show that consumer behaviour is shifting online. 

Search is being answered, not visited. Bain & Company (February 2025) found that 80% of consumers now rely on AI-written results for at least 40% of their searches, reducing organic web traffic by 15–25%. More striking: 60% of searches now end without users clicking through to any website at all. Remember this as we work through this analysis, as it’s important. 

AI adoption has gone mainstream, fast. Deloitte's 2025 Connected Consumer Survey found that 53% of US consumers are now experimenting with or regularly using generative AI, up from 38% a year earlier. Regular users, those who have moved beyond experimentation, nearly doubled to 20% in a year, and workplace usage has jumped more than fivefold since 2023, from 6% to 34%. 

Fewer apps, deeper engagement (convergence). Global app downloads declined for the fifth consecutive year in 2025, reaching 106.9 billion, down from 109.8 billion in 2024 (TechCrunch). Yet consumer spending inside apps rose 21.6% to $155.8 billion over the same period. People aren't downloading more apps; they're doing more with fewer apps. 

The agent layer is forming quickly, and not just in consumer tech. Gartner forecasts that 40% of enterprise applications will feature task-specific AI agents by the end of 2026, up from less than 5% a year earlier. In a separate survey, 93% of IT leaders said they plan to introduce autonomous agents within two years, and nearly half had already done so. 

Have you heard of OpenClaw? (you probably have). It’s one of the clearest examples of what happens when you combine all of these shifts: instead of opening a browser to search or juggling a dozen apps. People are starting to wire AI agents like OpenClaw directly into the few tools they already live in and asking them to handle the work on their behalf.

Just this week, Salesforce launched Headless 360, transforming its CRM platform into a "headless" infrastructure optimised for AI agents by exposing nearly all capabilities as APIs, MCP tools, and CLI commands, eliminating the need for browser-based UIs. This shift supports agent-first workflows in which software agents handle tasks such as case resolution, sales deals, and approvals directly.  

There is an opportunity and a challenge for SaaS products that are already embedded into businesses and have years of data, by moving to a headless solution. Traditionally, per-seat models assumed users logged in, navigated UIs, and were upsold via feature gates, but AI agents bypass UIs entirely, making API calls for data, workflows, and actions. Therefore, for those first movers;  

I’m sure you’ve also seen the posts across social media claiming that SaaS is dead, and there may be an element of truth to that, especially for SaaS products that are more like tools than databases, e.g. Figma vs Salesforce. 

Each time a major AI company launches a new feature or product, e.g., Claude Design, it has a significant negative impact on some existing SaaS share prices as investors and the market try to assess its potential future impact. This is the convergence I spoke about. Consumers can do everything they used to do in one place. 

OK, so we now understand the driving forces behind these shifts. Before we look to the future and potential disruptions, I feel it’s important to review and understand what sports organisations’ digital and consumer strategies have been over the past 5-10 years. None of these are automatically redundant because of the introduction of AI, but they could be disrupted or become even more important. 

Across sports, there has been a drive to move fans to owned-and-operated platforms to build a deeper understanding of them and to own the direct fan relationship. I’ve written about this in the past, and it’s been the topic of every sports conference in recent years. The rise of social platforms during that time boosted sports' overall brand value and drew more attention to the sport. However, some might say they also limited the ability to deeply understand fans, prompting sports organisations to seek change. We all recall the headlines about Barcelona and Spotify. 

Branded and sponsored digital content. Rights holders moved from selling static logo placements to packaging editorial, live match centres, score pages, and content series as named sponsor inventory, turning attention on owned platforms into measurable, attributable commercial assets. 

Digital ticketing and commerce. As teams look to move away from legacy third‑party marketplaces, digital ticketing is increasingly being built around owned or direct‑to‑fan infrastructure, turning each purchase from a one‑off transaction into a first‑party data event.

Membership and subscription tiers. Rather than competing with broadcasters for live rights, a growing number of rights holders have built tiered membership models that monetise access, community, and exclusivity; content that can't be aggregated or summarised elsewhere. 

Liverpool FC, for example, became the first global sports team to launch YouTube Memberships, offering premium shows, live reactions and behind‑the‑scenes content from £0.99 per month on top of its free channel. The club’s YouTube presence has since grown past 11 million subscribers, illustrating how even non‑rights‑holding content can support significant engagement and recurring digital revenue when layered with paid memberships and other monetisation on a scaled channel.

Fan data as a commercial asset. First‑party data collected across ticketing, apps, content and commerce has increasingly been packaged as a commercial multiplier, enriching sponsor valuations, enabling more precise audience targeting, and powering privacy‑centric data-clean-room partnerships with major brands.

This one is key. If the future does become more headless and API-focused, with agents acting on behalf of consumers, then sports organisations that have built large, robust audience databases will continue to prosper, as they can still maintain direct 1-to-1 communication with their fans. It’s hard to know whether an agentic consumer future leads to fewer opportunities to understand your audience, but the likelihood is that it will be more difficult. An agentic future will most likely lead to a more anonymous, privacy-first internet, except at moments when purchases occur, and you need specific first-party data such as names and addresses. 

Data licensing: the emerging Tier 1 model. The most commercially advanced rights holders have gone beyond fan profiles to monetise proprietary match and event data, licensing real‑time feeds to sportsbooks, media outlets and data providers, the same infrastructure that is increasingly used to power betting models, analytics and AI‑driven products. This model depends on genuinely scarce data systems and is largely confined to elite organisations.

The PGA Tour’s ShotLink system, for example, has tracked every shot on the Tour for more than two decades and is licensed to partners such as STATS  to power media coverage and betting markets. As the Tour put it in announcing a data partnership, it is “in a unique ownership position with the data collected at its tournaments through ShotLink” — a data moat that generates licensing revenue independently of how many fans visit PGATOUR.com on any given day.

As I have discussed throughout this article, more and more fans are using AI in their daily lives. That convergence is starting to happen. Consumers' behaviour will shift across all industries, so it’s not just sports that need to consider the future. The data shows that consumers want a single AI-aggregated platform that pulls together content from websites, search engines, and social media. (68% according to Capgemini’s 2025 retail research)  

Jarek Rozanski has written well on this, distinguishing three patterns in how AI is changing website visits. In the first, the visit happens, but the fan reads the AI summary and bounces; the page view is real, but engagement is gone. In the second, the visit never happens at all; the fan asks an AI, gets the answer, and moves on. In the third, the AI cites its sources, and the user clicks through, warmed up and engaged.

Rozanski's insight is that most fan queries end in pattern two, not three. I'd add one caveat for sport specifically: video still pulls a disproportionate share of pattern-three visits, which is why rights-holders with strong video IP retain more leverage than text-heavy publishers. 

If two out of these three patterns keep growing, and people who see your content never visit your site, what is the purpose of your platform?

Before going further, one important framing: sport, in most cases, has something valuable, a live event that fans will always turn out for, digitally or physically. None of what follows changes that. 

The disruption I'm describing happens in the spaces between: the fan on the move, checking where Rory McIlroy is on the leaderboard or how their football team is doing; the casual supporter deciding whether to buy a ticket for next weekend; and the new fan researching a team they've only just started following. These are the key moments that owned platforms were designed to capture, and these very moments are now beginning to be influenced by AI and ambient consumption. 

Here is a perfect example of a consumer who built their own dashboard for monitoring The Masters. I came across this on LinkedIn last week. He's pulling live data from ESPN's public API (which I'm not convinced ESPN realises is being used this way; more on that later), combined with feeds from Polymarket and elsewhere, and has assembled the whole thing using Perplexity's Computer product. 

Most sports platforms were built on the premise that they are content destinations. Fans come to read, watch, and discover. That logic is breaking down, and it was always somewhat fragile, because the content itself was rarely scarce enough to be worth a dedicated visit. In a world where AI agents & crawlers increasingly consume your content on a fan’s behalf - without the fan ever visiting, the human (fan) visit becomes something organisations need to work harder to earn. So before we look forward, it's worth going back to the beginning and asking a more fundamental question.

If a fan visited your platform today with no intention of watching live content, what would they find that they couldn't get faster, better, or more conveniently elsewhere?

For most sports organisations, that question is harder to answer than it should be. That's not a criticism of the work that's been done; the investment in owned platforms over the last decade has been real and meaningful. It's a diagnosis of what that investment was optimised for: reach, traffic, and content volume. 

First, let’s take the panic level down a bit :) 

None of what I’ve described above happens overnight. Sports will likely shift more slowly than the wider consumer internet, and within sport, the pace will vary significantly depending on your audience demographics, your commercial maturity, and how digital your fan relationship already is.

The PwC Global Sport Survey found that the median age of fans across the major US sports is 47-51 for golf, 48 for baseball, 46 for the NFL and NHL, and 43 for the NBA. These fans are the least likely to be wiring OpenClaw into their daily lives, asking Claude to summarise the weekend's action, or buying tickets through an agent. My own behaviour at the top of this piece is not representative of the average sports fan in 2026, and it would be a mistake to plan as though it were. 

The truth is that the agentic shift may take seven to ten years to fully take hold in sport, not the three to five years most retail and media analysts are forecasting. The owned platform will remain more valuable for longer in sports than in almost any other content category.

But three things keep this from being an argument for inaction. Younger fans, the 18-34 cohort that most sports are already losing on other metrics, are moving fast, and the agentic shift will accelerate that loss, not slow it. Commercial partners don't have demographic drag; sponsors, sportsbooks, and media companies are already being rewired for an agentic world and will reshape sports' commercial conversations, whether or not fans have caught up. Decisions that are most critical, such as rebuilding contracts, restructuring teams, and negotiating with AI companies, typically take years rather than months.

Begin with what you can, but if you want four key areas to start, this is what I’d recommend: 

Point 4, however, is a tension worth calling out. Throughout this analysis, I’ve described AI platforms as a disruptive force that sports organisations need to prepare for, design around, and, in some cases, defend against. And now, in the same analysis, I’m suggesting you license your most valuable content assets to the very companies driving that disruption. That contradiction deserves to be acknowledged. 

The commercial logic is sound in the short term, but the long-term risks are less clear. Providing AI companies with access to high-quality sports content that makes their products better, their answers more accurate, and their platforms more valuable to the fans you're trying to keep. You may take a licensing fee today and spend the next five years watching that content fuel a product that makes your owned platform increasingly redundant. The media industry has had versions of this conversation before with Google, with Facebook, and with aggregators of every kind, and the outcomes have rarely favoured the content creator as much as the initial deal suggested they would. 

The question of leverage also arises here. These commercial structures are relatively new, meaning that legal frameworks, valuation models, and deal precedents are still developing. AI companies possess a much clearer understanding of their technology and its future potential than most sports organisations. This knowledge gap generally favours the party with a better grasp of the asset's future worth, which, at present, is not the rights holder. 

Currently, I am unaware of any sports organisation that has publicly announced an AI content licensing agreement like this. This could be because such discussions are happening privately, or because the risks involved, legal issues, reputational damage, and the challenge of creating something with no prior example, are too great compared to the potential benefits. 

My opinion: if sports are going to negotiate with AI companies, they should do so collectively. The media industry's mistake was that individual publishers cut individual deals before anyone fully understood what the asset was worth. Sport has the chance to learn from that and, if it moves collectively rather than competitively, approach it with strategic clarity rather than commercial desperation, know what you're licensing and what you're giving up, and expect to regret the terms if you sign them alone. 

One final question, and it's the one this whole piece builds toward: who in your organisation actually owns this decision?

This isn't a technology question. It's a commercial, organisational, and strategic question, deeply influenced by technology. If the digital shop-front is becoming a set of APIs consumed by agents rather than a website visited by humans, the decision about what to build and what to sunset can't rest with the CTO alone. It involves commercial, legal, partnerships, and the board.

And if sports organisations move toward a more platform-first model, the technology team itself will look structurally different. Fewer front-end designers. More backend and developer talent. More ‘all-rounders’, orchestrators and maybe fewer do-ers. The people who used to build for humans now build for agents that build for humans. The great thing about this is that it is not a doom-and-gloom prediction about the future of jobs. I see it as a positive development. Throughout history, job roles have changed, and we now have the chance to do the same if we wish. Access to tools that help solve complex problems is now so widespread, and the barrier to entry is so low, that the learning curve for people is much less. 

The visible part of the problem lies in the strategic decisions, but the unseen part concerns whether your organisation is appropriately structured to implement them.

Thanks for reading this piece. As mentioned, it’s one of the most in-depth editions I have written, and it required extensive research and learning. If you have enjoyed this article:

Buy Me A Coffee (100% optional):