Telemetry: the key to solving AI’s valuation and measurement crisis
Opinion
OpenAttribution’s director discusses the measurement crisis in the AI economy, how legislation is only part of the solution, and why telemetry is key to ensuring quality content fuels AI systems, cites sources and understands outcomes.
As AI’s grip on online activity tightens, the entire digital economy faces a valuation and measurement crisis. For years, LLMs and AI agents have scraped with impunity, taking the content and data that fuel their algorithms without asking – let alone remunerating – the creators or owners of that material.
Brands and publishers of all kinds (from traditional media to creators and affiliates) are finally waking up to the fact that they’re being ripped off. Yet stakeholders across the supply chain still struggle to decide how to value content for AI systems and how to understand the new outcomes shaping this information era.
These two issues are troubling the very top of the UK Government. In May this year, the House of Lords Communications and Digital Committee published its report on AI’s impact on the creative industries. It concluded that a “licensing-first” approach would end the extractive AI era.
The UK Government’s action is reflected globally, including in the EU and US, with all concluding that telemetry is the critical infrastructure required for the AI economy to thrive in a safe, legal, and positive way.
But governments can debate legal structures until they’re blue in the face. Ultimately, to protect the value of content and understand the outcomes it influences, we must agree on a measurement layer that shows what happens to content once ingested into an AI platform; that layer is telemetry.
Safely and confidently
A strong AI economy requires AI platforms and LLMs to produce decent, accurate answers to user queries. Recently, A.G. Sulzberger (chairman of The New York Times Company) reported that OpenAI has acknowledged it would be “impossible to train today’s leading AI models without using copyrighted materials”.
As a result, quality publishers and creators have practical leverage in their relationship with AI solutions because their content powers these platforms.
Through a telemetry solution, all parts of the equation can gain from the transaction, whatever form that might take. Creators and publishers get financial reward for their work, while AI developers get licensed, traceable inputs, allowing them to build safely and confidently without the risk of a lawsuit.
To accurately measure the value quality content offers AI systems, material must be treated as a premium ingredient with a traceable supply chain; telemetry provides that.
Telemetry solution
A telemetry solution eliminates the probabilistic guesswork and scattered reporting that currently plagues methods used to measure outcomes. It creates a reporting signal that is standardised; a kind of counter for when content is cited and retrieved.
Just as a supermarket manager knows exactly how many units of a product are sold each day, telemetry gives an AI system a unit of measurement that lets it track how much content it consumed and, subsequently, how it was used.
When ads.txt was introduced in 2017, it successfully combatted fraud in digital advertising by creating a visible record of authorised sellers. Telemetry will do the same for AI platforms, creating an irrefutable ledger to measure outcomes and provide accurate citations.
Agentic middle layer
In the new era of information, understanding outcomes will be key for both AI platforms and publishers. The old “clicks-for-traffic” tension that has driven digital advertising for decades is now redundant. About 60% of searches no longer lead to a click, and Google’s AI Overview kills 61% of clicks on organic results.
Furthermore, traditional subscription and click models can’t survive when content supply is effectively infinite. One BBC story can be replicated thousands of times in an instant by any number of publishers or AI platforms, making it near impossible to measure the genuine outcomes and impact the piece has.
The rise of the agentic middle layer is further muddying outcome measurement. LLMs such as Amazon’s Rufus or Walmart’s Sparky now exist between consumers and brands. Their growing popularity drives outcomes across the customer journey, yet it’s still difficult to understand and measure where the content fuelling their answers comes from.
Measuring the right metrics
As such, simply measuring metrics like ‘visits’ no longer provides an accurate picture of content performance. Instead, we should look at a piece of content’s influence on consumers when delivered by an LLM, for example, and, through this, cite the material’s creators and publishers. Only then can we understand a piece of content’s full value, and the outcomes it has driven across the consumer journey.
In 2026, we stand at a crossroads. If the AI economy is to be sustainable, and not purely extractive, we need a citation mechanism that operates across the whole digital space.
We need to decide whether to follow a healthy format in which the value of a piece of content is acknowledged, its influence recognised (and cited) and the outcomes it drives accurately measured. This way, premium content can be safely licensed, and everyone from AI platforms to creators and consumers benefits.
The alternative is to preserve an extractive status quo: one in which AI systems absorb and reproduce whatever content attracts the most engagement, rewarding scale, sensationalism and outrage rather than accuracy and originality.
The largest platforms capture the value, while the creators of high-quality material go unrecognised and uncompensated – and users are left with a less trustworthy information ecosystem.
By following the former path and adopting an AI telemetry citation solution, we can better understand how AI platforms obtain and use quality content. In doing so, we can re-establish the value exchange that AI has destroyed in recent years, creating a better-performing, fairer internet for all.
Alex Springer is director of OpenAttribution. To hear more on telemetry, listen to The Media Leader podcast this Monday.