Put agents to work today
Transform your campaign workflows with powerful AI that delivers measurable results.
Book a meeting
Every media organization has data, but getting that data into a shape that people, systems, and AI can actually use can feel like a Sisyphean task.
Campaign data arrives from different platforms, partners, markets, and teams. Naming conventions vary, taxonomies drift, fields don't line up, definitions change, and before anyone can answer a client question or put an AI workflow to work, someone has to reconcile it.
For enterprise media companies, that reconciliation compounds quickly, creating a growing backlog of messy data, consuming analyst time that could be spent on analysis, delaying answers to client questions, and forcing AI initiatives to inherit the same inconsistencies.
The problem isn't that the data isn't there.
It's that too much of the work happens after the data arrives.
This issue is by no means unique to a single media company. In fact, in 2026, the Interactive Advertising Bureau (IAB) launched Project Eidos to address fragmentation and inconsistency in how campaign data is structured, classified, and connected across the industry.
IAB's work describes an all too familiar campaign data problem: brands, agencies, and platforms all collect campaign data, but they don't structure it the same way. The same inventory can be labeled differently across systems, classifications don't match, and naming conventions vary by organization, team, or tool. Before meaningful analysis can happen, teams spend significant time translating and reconciling data across sources.
IAB has since finalized Campaign Data Standards 1.0, establishing a shared framework for describing advertising experiences and inventory across platforms, channels, and campaign systems. The goal is to improve consistency and comparability across digital advertising.
While it's an important step, standards alone don't solve what happens to the data already flowing through a large organization.
Enterprise media companies still have years of campaign data, client-specific taxonomies, proprietary systems, vendor feeds, and market-level variations to contend with. So while the question of ‘what should the standard be’ is a good starting point, it is by no means a conclusion.
The industry needs to ask how an organization gets the data it already has to conform to a more consistent structure.
When campaign data arrives with different naming conventions, taxonomies, and structures, someone has to make it consistent before analysis can begin.
Analysts map fields, normalize names, resolve duplicates, and reconcile definitions, often across data from different platforms, partners, and markets, and because new data keeps arriving, the work doesn't really end.
This, in essence, reconciliation tax. The organization isn't just paying for the original data infrastructure; it's also paying people to translate between systems that don't speak the same language.
And the cost isn't limited to cleanup.
It's the analysis that gets pushed back, the client questions that have to wait, the one-off reporting processes created for a particular market or client, the models that need their own version of the data, the AI workflow that requires another layer of transformation before it can actually work.
IAB's 2026 research points to many of these same structural issues, identifying fragmented operations, inconsistent data, conflicting signals, and siloed AI adoption as foundational challenges for modern measurement. The organization’s State of Data report argues that incremental fixes are no longer enough and that shared standards and trusted data foundations are needed for AI-enabled measurement to scale.
For media companies, data preparation has become the work instead of the work enabling something else.
For an organization operating across client environments, data readiness has another consequence: access.
Clients increasingly expect faster, more direct answers from the data their agencies manage, but putting conversational access or AI on top of inconsistent data doesn't make the underlying problem disappear.
It makes the problem visible.
A client asks a simple question.
One market uses a different naming convention.
Another uses a different definition.
A field is missing.
Two systems describe the same thing differently.
The answer now requires someone to investigate the data before they can answer the question.
The bottleneck moves from reporting to trust.
And that matters more as media organizations give clients more direct access to their data. A dashboard can hide a lot of complexity. A human analyst can compensate for inconsistencies they know are there.
A conversational interface can't rely on either.
If a client can ask a question directly, the underlying data needs to be structured well enough for the system to understand what that question means and return an answer the client can trust.
This is why the industry's push toward common campaign structures matters. IAB's Project Eidos is explicitly focused on creating shared structures, classifications, and data conventions that reduce fragmentation and the need for downstream normalization and reconciliation.
The argument for data readiness isn't new, but what organizations are asking their data to do has.
AI systems don't just report on data; they use it to answer questions, generate recommendations, build models, trigger workflows, and increasingly take action, making consistencies far more consequential.
Historically, an analyst could act as the translation layer between messy data and the person asking the question. They knew which fields to use, which definitions didn't line up, and which exceptions needed to be accounted for.
That knowledge often lived in the people and processes surrounding the data, but AI needs that context to be much more explicit.
If the underlying data is inconsistent, an AI system can surface those inconsistencies faster than a manual workflow ever could. And when AI moves from answering questions to taking action, the consequences of getting the underlying data wrong become larger.
IAB's 2026 State of Data research makes a similar point: AI can improve data preparation and measurement, but without shared standards, transparent governance, and trusted data foundations, it can also reinforce existing problems and introduce new concerns around accuracy and data quality.
The lesson for media organizations is straightforward: AI doesn't eliminate data preparation. It makes the quality of that preparation more consequential.
Consider what happens when inconsistent data is allowed to persist.
Analysts reconcile data before answering a question.
Teams hesitate to expose data directly when they can't trust that the underlying definitions are consistent.
Chat and agents surface inconsistencies that were previously buried in manual workflows.
Client-specific models remain isolated from the workflows where they could actually be used.
Different definitions and structures make it harder to connect outputs across channels, methods, and markets. But these aren't five separate data problems. Rather, they're downstream consequences of the same one: data isn't consistently ready for use.
And that creates another problem at enterprise scale: duplication.If one team transforms campaign data for reporting, another transforms it for a model, and another builds a separate mapping for an AI workflow, the organization may be solving the same underlying data problem multiple times.
The more systems and workflows an organization adds, the more places that logic can live. And the more places it lives, the harder it becomes to maintain a consistent definition of the data, which is how a data-quality problem becomes an infrastructure problem.
The answer isn't to replace everything that already exists. Large media organizations have spent years building data platforms, client environments, models, workflows, and proprietary tools. The opportunity is to make that infrastructure more useful. This means moving the transformation work closer to where data enters the system, rather than asking every downstream workflow to solve the same problem again.
The goal isn't to replace the systems an organization has already invested in, but to make the data moving through those systems more usable. Instead of repeatedly paying the reconciliation tax downstream, the organization can address more of the problem at the foundation.
Every hour spent reconciling is an hour not spent analyzing, every client question that requires another manual reporting cycle is a slower answer, every AI workflow built on inconsistent data carries that inconsistency forward, and every model that stays disconnected from production represents work that never reaches the workflow it was built to improve.
The goal isn't perfect data.
It's data that's consistent enough, governed enough, and usable enough to support what comes next.
That's what makes AI infrastructure useful at scale.
To learn more about Akki’s approach to data transformation, request a meeting today.
Transform your campaign workflows with powerful AI that delivers measurable results.