AI Reality Check From a Former Meta AI Marketer (Our CMO)

In the last year, we saw companies increase their data analytics and customer insights investment by 54%. Yet despite this spending surge, most organizations are making a critical mistake: building AI on top of broken infrastructure.

In a recent conversation on the Martech Podcast, our CMO Noha Rizk sat down with host Benjamin Shapiro to discuss why legacy dashboards and delayed data are holding businesses back - and what it takes to build systems that deliver insights and AI-powered actions in real time.

First reality check? Infrastructure, and the importance of incrementality

"The modern data stack is fundamentally broken," Noha explained. "It's fragmented, slow, and costly. Legacy processes require staging and conforming data to outdated models, making reporting backward-looking and often obsolete."

The parallel to the early Internet days is striking. When browsers first appeared, companies built websites like digital brochures: pretty, but static. They had powerful new rails but weren't maximizing what those rails could do. Today's AI moment requires the same fundamental rethinking.

"You can't just slap AI on top of existing infrastructure," Noha said. "You need to ask: will this infrastructure actually help me maximize what I can do with this new technology?"

The biggest obstacle to adopting real-time data infrastructure? The math doesn't compute for many companies. Real-time systems are costly and time-consuming, making organizations worry about losing data fidelity. The total cost of ownership becomes prohibitive, and ROI calculations fall apart.

Noha's advice, drawn from her experience at Meta, centers on incrementality: "You have to really think through where in your organization there will be incremental value. Pick the pockets within your company where this will actually give you measurable returns, so that pain of migration becomes worthwhile."

Not every function needs live data. But identifying where it creates significant impact - whether in reducing waste, optimizing logistics, or improving customer engagement - is the homework that separates successful transformations from expensive failures.

Where do we see live data actually drive results?

Certain verticals see immediate returns from real-time analytics:

Retail & Food Service: One Incorta customer with 3,000 physical branches uses live data to optimize inventory waste. By understanding which items are moving at what times of day, they can push geo-targeted promotions to nearby consumers, maximizing revenue through strategic discounting rather than throwing products away. They can also move inventory between branches based on real-time demand signals.

Manufacturing: Companies track factory floor operations across multiple locations in real time, creating massive cost savings. For manufacturers where every cent counts, this translates to hundreds of thousands of dollars in operational improvements.

Healthcare & Retail: Imagine receiving a notification from your pharmacy that says, "We've seen a 75% increase in Tamiflu sales this week - a flu epidemic in your ZIP code is coming. Get vaccinated." That's the difference between generic marketing and data-driven personalization that actually resonates.

Perhaps the most surprising? Education. As universities become more tech-enabled, they're using live data to track student outcomes and behavior in near real-time, adjusting instruction, curriculum, and grading systems dynamically.

Creativity craves space for curiosity

One of the most compelling insights from the conversation centered on creativity versus control. While some argue that unlimited data access creates analysis paralysis, Noha makes a counterpoint: breakthrough moments require space to explore.

"The creative process requires enough space to go down rabbit holes and noodle and get lost a little bit," she explained. "If you close that door and become too rigid, only measuring certain KPIs, you might miss important insights."

She shared a powerful example from Meta's Marketplace product. When COVID hit and person-to-person exchanges stalled, only their deep data access and curiosity revealed that users were trying to find alternate transaction methods. This insight led to subsidized shipping, driving hockey-stick growth.

"You can't design for those moments," Noha said. "You can't put rigid rails around innovation."

The key is balance: maintain your North Star metrics and KPIs while enabling the freedom to dig deeper when trends change and ask why.

Lessons from Meta

Noha's most important takeaway from her time at Meta? Data helps you move faster.

"Bureaucratic processes that slow companies down typically exist to mitigate bad decisions," she explained. "Meta flattened decision-making as much as possible, putting decisions as close to the work and the signal. To enable that, Meta had to be one of the most data-driven companies in the world."

The mantra wasn't just "move fast and break things" - it was move fast with confidence, powered by data fidelity and real-time insights. At Meta's scale, every tweak and fine-tune impacts millions of people. That requires data you can trust, delivered instantly.

Killing the "awareness" metric

Noha shared a controversial take: awareness as a marketing metric needs to die.

"There's no inherent value in just being known," she said. "Someone hearing your name or recognizing you - that's a vanity metric. You need stronger value signals: breakthrough, message comprehension, intent signals."

Awareness matters for top-of-funnel, but its usefulness dies almost as quickly as you measure it. "It's the fruit fly of metrics," as the conversation concluded.

Instead, the focus should be on insights that drive action. "Analytics is nothing without insights," Noha emphasized. "Raw data doesn't mean much. How you make connections between data: that's where value lies. Question the why behind the data constantly."

Right person, right message, right moment

Looking ahead, Noha sees AI enabling marketing analytics to become laser-focused on hitting the right message to the right person at the right time through hyper-targeted, cross-medium personalization.

"We often contend with having the right message for our ICP, but hitting them at the wrong moments," she explained. "Maybe they're seeing it in the morning when they should see it in the evening, or on LinkedIn when they should see it on TikTok. What AI analytics will enable is getting really, really good at that precise timing."

Think 'Minority Report's hyper-personalized holographic ads - but actually useful and contextually relevant.

Companies scrambling to add AI capabilities without addressing their broken data infrastructure are building on quicksand. The winners will be those who do the homework, identify where live data creates incremental value, and build systems that can unlock creativity, rather than constrain it.

As Noha put it: "The question isn't whether you can afford to use live data. It's whether you can afford not to."

Listen to the full conversation on the Martech Podcast to hear more about Noha's experience at Meta, her approach to marketing analytics, and why she believes websites as we know them might be dying...