Data has always been out there, buried in billions of bid requests flowing through DSPs every second. The real problem is not about data availability but rather accessibility: turning data into actionable intelligence fast enough to actually influence decisions.
“What frustrated me most was how slow and fragmented supply intelligence was. In programmatic, the data exists… But ironically, accessing actionable insights from that data was still incredibly manual.” – Tomás Yacachury, Global Supply & Partnerships Lead at Kayzen and Product Owner of kAI Explorer.
For too long, media planners have built campaign strategies on static reports, BI dashboards or worse, assumptions carried over from previous campaigns.
kAI Explorer was built to change that.
We sat down with Tomás Yacachury to unpack the thinking behind kAI Explorer. Not just what it does, but why it exists in the first place, what planning problem it was built to solve, and what it says about how programmatic planning should work today.
The Planning Problem Nobody Was Solving
Before kAI Explorer existed, the frustration, Tomás explains, wasn’t about data scarcity. It was about turning that data into insights that inform decisions.
“Programmatic is a real-time, highly dynamic market, but the planning tools were operating like static quarterly research reports.”
The workflow was painfully familiar to anyone who worked in media planning or programmatic teams. Someone asks whether there’s enough rewarded inventory in Japan. Someone else pulls a report. A few hours, if not days, later, the team gets a partial answer. By then, the moment may have passed.
Running campaigns without properly validating inventory data can lead to poor performance: targeting that has almost no reachable inventory, CPM expectations misaligned with the market, or creative formats that lack sufficient scale.
For Tomás, the risk is straightforward: “If you don’t validate supply first, you can easily design a campaign that the market simply can’t support … These issues don’t always appear until the campaign is live. At that point, you’re already burning time and budget trying to debug something that was fundamentally a planning problem.”
The irony is that the intelligence needed to avoid all of this exists, but it just wasn’t accessible in a way that matched how planning actually works.
The Insight that Started All
The idea behind kAI Explorer didn’t emerge from a product roadmap discussion. It came from an unexpected but regular day-to-day situation.
The aha moment came while Tomás was working on Kayzen’s Programmatic Inventory Index. The report aggregates a significant volume of inventory intelligence across the global in-app ecosystem; format distribution, OS share, and regional inventory patterns. It proved that the underlying data was genuinely valuable but also revealed a structural limitation.
The report was static and published annually. Marketers could read it and understand macro trends, but when they needed to answer practical questions, such as how scalable is rewarded inventory in Brazil, what CPM range should we expect for Android in Germany, or which formats dominate inventory in Southeast Asia, they still had to go back to dashboards, spreadsheets, or manual analyses.
For Tomás, that gap made the problem obvious: “What if marketers could explore the programmatic supply landscape the same way they ask questions in a conversation?”
That question became the foundation of kAI Explorer: taking the intelligence captured in the Programmatic Inventory Index and turning it into something interactive, real-time, and instantly explorable.
Not a better dashboard, but something fundamentally different.
The Data That Powers kAI Explorer
At its core, kAI Explorer is built on the bid request data that flows through Kayzen every day. In Programmatic advertising, every ad opportunity begins with a bid request sent from an exchange to a DSP. Each request contains structured information like: app, country, device type, operating system, ad format, and placement context.
A global DSP like Kayzen receives billions of these requests over time, spanning hundreds of thousands of mobile apps across 180 countries, and multiple ad formats and device environments. Aggregated across a 90-day rolling window, it becomes an extremely rich view of how the global in-app inventory landscape actually behaves.
As Tomás explains, “Instead of seeing billions of raw auction opportunities, kAI Explorer translates that into market intelligence about supply, reach, and pricing dynamics.”
But aggregating that data was only half of the challenge. As Tomás frames it: “One of the biggest parts of building kAI wasn’t the AI itself. It was the data normalization layer. Structured data gives the illusion of being clean, but in programmatic, it rarely is … Building an AI agent on structured data isn’t about feeding it everything—it’s about designing a hierarchy of relevance. With kAI, we learned to start small”, Tomás explains, a challenge he expands on in his article on building AI agents on structured data in programmatic.
When dealing with planning intelligence, precision matters most. That principle has guided the work behind kAI Explorer.
The Dream Team and Formula for Success
When it comes to approaching a new market, Tomás says most planning questions tend to revolve around the same core uncertainties: how much inventory actually exists, what formats dominate the supply, and whether the pricing expectations match reality.
Over time, Tomás has seen a simple workflow emerge from answering those questions. The best workflows he has seen play out in four steps, each one compressing what used to take hours into minutes:
- Validate inventory scale. Is there enough inventory to support the campaign?
- Check format distribution. Does the market support the creative strategy?
- Benchmark expected pricing. Are CPM expectations realistic?
- Define main channels: Which ad exchanges and apps support my campaign strategy?
The output isn’t a report to file away but a way to identify opportunities backed up by data, not assumptions. Within a few minutes, a team can move from guessing about a market to understanding it.
Three groups benefit most from this:
- Media planners who need to validate supply before pitching.
- Agencies approaching new markets that need to understand reach and price dynamics quickly.
- Growth teams scaling internationally without deep local expertise.
The common thread is anyone who needs to move from assumptions to evidence, fast.
The impact becomes clear in real workflows. Z2A Digital, for example, reported that planning scenarios that once required hours of manual analysis across datasets can now be explored in minutes using kAI Explorer.
Where It Goes From Here
Looking ahead, Tomás sees two main directions shaping the future of campaign planning intelligence.
The first is depth. Today, kAI Explorer already answers the whats and the wheres of programmatic inventory. The next step is helping marketers understand the why, bringing more contextual insights built on a richer data foundation.
The second direction is improving the experience of exploration itself. While kAI Explorer already provides powerful market intelligence, Tomás wants the interaction to feel less like operating a tool and more like working alongside a strategist.
“We want the system to feel less like querying a tool and more like collaborating with a strategist.”
The broader opportunity, as he envisions it, is a shift in how marketers relate to the markets in which they operate. Execution tools for bidding, targeting, and optimization have been sophisticated for years. Planning, on the other hand, has often relied on experience and instinct, when it could have relied on evidence.
“If we get this right, marketers won’t just run campaigns more efficiently. They’ll actually understand the market they’re operating in. And that changes everything.”
In the end, that shift may be one of the most important evolutions in programmatic planning. When market intelligence becomes instantly explorable, feasibility questions no longer slow teams down, and planning decisions can start with evidence rather than assumptions.
For anyone working in programmatic today, that’s a capability worth experiencing firsthand. If you’re curious about how supply actually behaves across markets, formats, and platforms, kAI Explorer is available to explore for free.
