AI Marketing Decision Intelligence: POMO Founder Praneet Dutta on Autonomous Marketing Operations and the Future of D2C Growth

At CXQuest, we are publishing an in-depth conversation with Praneet Dutta, co-founder of POMO and former Google DeepMind engineer who worked on foundational AI systems including Imagen and Gemini. The discussion centers around a major emerging shift: AI marketing decision intelligence.

A Period of Operational Overload 

Marketing organizations are entering a period of operational overload. As brands expand across digital channels, marketplaces, quick-commerce ecosystems, CRM systems, social platforms, and AI-native search environments, the number of daily decisions required to sustain customer engagement has increased dramatically. Yet most marketing teams remain structurally under-equipped to manage this level of complexity in real time.

This challenge is especially visible among modern D2C brands operating with lean teams while simultaneously managing rising customer acquisition costs, fragmented consumer attention, platform volatility, and rapidly shifting competitive environments. In many cases, marketers are spending more time reacting to dashboards than identifying which decisions actually matter most for customer growth and retention.

The emergence of AI marketing decision intelligence represents a broader shift away from static reporting systems toward continuous signal-to-action orchestration. Rather than simply generating analytics, newer AI-native systems are beginning to prioritize actions, recommend execution pathways, coordinate workflows, and compress marketing response cycles from weeks into hours. That evolution could fundamentally reshape how customer journeys, personalization, campaign strategy, and growth operations are managed.

About Praneet Dutta, Co-founder of POMO 

In this CXQuest interview, we speak with Praneet Dutta, co-founder of POMO, an AI-native platform focused on autonomous marketing decision systems for modern brands.

Before building POMO, Praneet spent years working on foundational AI systems at Google DeepMind, contributing to projects including Imagen and Gemini. He co-founded POMO alongside Joe Cheuk, whose background includes experience at Databricks, Meta, and Google Cloud.

The conversation explores how AI marketing decision intelligence may evolve beyond dashboards and recommendation engines into continuously operating decision systems capable of prioritizing, coordinating, and executing marketing workflows across fragmented digital ecosystems. It also examines the implications for customer journeys, personalization, operational agility, and the future of AI-driven marketing organizations.

Fragmented Platforms,  Signals, and Dashboards 

Q1. Marketing teams today are overwhelmed by fragmented platforms, signals, and dashboards. What structural problem did you originally set out to solve with POMO?

PD: The original problem was simple to describe and hard to fix: marketing teams had more data than ever and less clarity about what to do next.

A modern team is looking at Instagram, Facebook, Google, TikTok, Shopify, the CRM, attribution tools, competitor activity, social trends, and a stack of weekly agency reports. Every system tells part of the story. None of them says, “Here are the three decisions you should make this week.”

That gap between information and action is what we built Pomo to close.

Pomo is built as an agentic AI marketing platform that sits as a decision layer on top of the marketing stack. It listens across performance data, customer behavior, competitor signals, and demand shifts, then helps the team prioritize what actually matters.

The goal was never to add another dashboard. It was to help marketers move from looking at data to making decisions, faster.

Autonomous Decision Layer

Q2. You describe POMO as an “autonomous decision layer” rather than a traditional analytics platform. What is the fundamental difference between the two models?

PD: Analytics platforms are great at telling you what happened. Pomo is focused on what should happen next.

That sounds like a small distinction, and operationally it’s the whole game. A dashboard can tell you ROAS is down, CAC is up, or a campaign is underperforming. But then someone still has to ask the second question: is this creative fatigue, a channel issue, a competitor move, a weak offer, or a customer segment problem? And even deeper: what do we do about it, and where do we start?

Pomo  is built for those two questions. It connects the signals, explains what likely changed, ranks what matters, and helps the team prepare the next action.

So I don’t think of Pomo as replacing analytics. It’s the layer after analytics. An AI marketing decision layer that turns insight into decisions, and decisions into execution.

Modern Growth-stage Companies 

Q3. Why do you believe marketing is becoming one of the highest-stakes decision functions inside modern growth-stage companies?

PD: Because marketing now controls more than campaigns. It controls growth efficiency.

For a growth-stage company, every channel, creative, offer, audience, and lifecycle program is really a capital-allocation decision. Where does the next rupee go? Which segment deserves more attention? Which message is actually working? And, which channel is quietly becoming inefficient?

Get those calls wrong and you don’t just run a bad campaign. You burn capital, lose momentum, and learn the wrong lessons about your customer.

That pressure shows up clearly across D2C, CPG, and growth-stage B2B teams right now, where headcount is lean, acquisition costs are climbing, and growth expectations are still high. Those teams don’t need more reporting. They need faster, sharper decision-making. That’s exactly the kind of team Pomo is built for.

Human Prompting and Interpretation

Q4. Many AI tools today still rely heavily on human prompting and interpretation. Why did you choose to build a continuously listening and prioritizing system instead?

PD: Because the most important marketing signals usually appear before anyone thinks to ask about them.

Prompt-based tools are useful, but they put a lot of burden on the marketer. You have to know what to ask, when to ask, which data to include, and how to interpret the answer. Marketing doesn’t move that neatly. Creative fatigues quietly. Competitors change offers. Demand shifts. A channel starts behaving differently. A customer segment starts responding to something new.

So we built Pomo to listen continuously and surface the few things that actually deserve attention. The product idea isn’t “generate more stuff.” It reduces the number of decisions the team has to manually dig for.

A good AI system in marketing shouldn’t make the team busier. It should make the next decision clearer.

Fragmented Digital Channels

Q5. How does AI marketing decision intelligence change the way brands manage customer journeys and personalization across fragmented digital channels?

PD: AI marketing decision intelligence helps brands stop treating the customer journey as a set of disconnected channels.

A customer might see an Instagram video, search the product on Google, compare it on Amazon and Flipkart, visit the brand’s site, abandon cart, see a Meta retargeting ad, get a WhatsApp message, and then buy during a promotion. Most tools split that journey apart by channel. Each one optimizes its own slice.

Pomo reads those signals together.

That changes what personalization actually means. It stops being about using a customer’s first name in an email. It starts with understanding what that customer probably needs next: education, urgency, social proof, a replenishment reminder, a bundle, a different offer.

For brands operating across paid media, CRM, commerce, marketplaces, and social all at once, that connected view stops being a nice-to-have. It becomes the only way personalization actually works.

Compressing Marketing Decision Loops

Q6. POMO references compressing marketing decision loops from weeks to hours. Operationally, what does that acceleration actually look like inside a fast-moving D2C brand?

PD: It means a team can notice, understand, and act on a signal inside the same business day.

In the old workflow, a team waits for a Monday report, sees performance drop, asks someone to investigate, checks campaign data, looks at competitor activity, debates it in a meeting, then decides what to change. By the time that loop closes, the market has often already moved past the signal.

With Pomo, the system is watching continuously. If creative is fatiguing, if spend is becoming inefficient, if a competitor changes messaging, if a demand signal is emerging, the team gets a clearer read much earlier in the cycle.

Operationally, that might look like pausing a creative on day two instead of day five, shifting budget across Meta and Google before the wasted spend compounds, launching a new test the same afternoon a competitor moves, or rewriting a CRM segment before a campaign goes out.

The point isn’t speed for its own sake. The point is acting while the signal is still useful.

Improved Marketing Execution

Q7. Without naming clients, can you share a real-world implementation example where the platform materially improved marketing execution or customer engagement outcomes?

PD: The pattern repeats across both B2C and B2B companies we work with: lean growth teams with plenty of information, but no clean operating rhythm.

One B2C deployment is a consumer brand running performance marketing across Meta, Google, TikTok, and marketplaces. The team was trying to understand creative fatigue, spend efficiency, competitor movement, and customer behavior at the same time. Before Pomo, those signals were showing up across separate dashboards and weekly reviews. By the time the team agreed on what to do, the signal was already a few days old.

With Pomo, those signals get turned into a prioritized action queue. The team can see which campaigns need attention, which creative themes are weakening, which audiences are worth testing, and where competitor pressure is starting to show up.

On the B2B side, the pattern is similar but the signals are different. A SaaS or services company might be watching Google and Instagram Ads, content performance, lead quality, sales feedback, and pipeline movement. Pomo helps connect those signals earlier, so marketing isn’t waiting until the end of the quarter to realize that lead quality shifted or a channel mix stopped working.

In both cases, the biggest change is usually inside the meeting itself. The conversation moves from “what happened?” to “what are we doing next?”

In practical terms, the team starts making decisions earlier in the week instead of spending the week trying to understand the problem.

That sounds like a small shift. It’s a major one. It means less dashboard-hopping, faster diagnosis, clearer execution, and a team that ends each week with a list of decisions made rather than a list of dashboards opened.

AI-driven Decision Fatigue

Q8. Modern marketing increasingly depends on first-party data, competitor intelligence, demand signals, and rapid experimentation. How do you prevent signal overload or AI-driven decision fatigue?

PD: By being ruthless about prioritization.

The risk with AI in marketing is that it can quietly create more noise. More summaries. More alerts. Then, more charts. More recommendations to consider. That isn’t intelligence. That’s automation overwhelming you.

Pomo is designed to do the opposite. It looks at signals together and asks the only question that matters: does this actually matter to the business right now?

A small CTR drop on one campaign may not. But if that CTR drop is happening alongside rising CPMs, weaker conversion, a competitor discounting in the same category, and inventory pressure on a hero SKU, then it probably does. That’s the kind of pattern Pomo is built to recognize and surface.

The job of AI here isn’t to show every signal. It’s to help the team understand which signal is worth acting on this week, and which can wait.

The teams that feel this pain most are usually lean growth teams managing too many channels with too little time. That’s where Pomo tends to create the most leverage.

Human Oversight and Governance

Q9. Many organizations worry about losing control as AI systems become more autonomous. How does POMO balance execution speed with human oversight and governance?

PD: We don’t think serious brands want black-box automation.

A brand shouldn’t wake up on a Monday and find that an AI system changed budgets, published campaigns, or made product claims overnight without approval. That isn’t a responsible model, and it isn’t a model marketing leaders will trust at scale.

Pomo is built around human-governed autonomy. The system can monitor signals, recommend actions, draft campaigns, prepare tests, and lay out next steps, but the brand defines the boundaries. Tone, claims, visual rules, budget limits, channel permissions, approval flows. All set by the team.

So the AI accelerates the work around the decision. The human team still owns the judgment.

That’s the balance that actually matters: faster execution, not less control.

Capital-allocation Function

Q10. You’ve described marketing as evolving into a “real-time capital-allocation function.” How should marketing leaders rethink decision-making in that environment?

PD: They need to think of marketing less as a campaign calendar and more as a portfolio of bets.

Every day, a growth leader is allocating attention and capital across Instagram, Google, TikTok, creators, CRM, marketplaces, WhatsApp, retention, content, and a long list of experiments waiting in the queue. Each one is a position with a different expected return, a different time horizon, and a different decay curve.

The old rhythm of monthly planning and weekly reporting was built for a slower market. Today, a creative can fatigue in days. A competitor can change the economics of a category overnight. A demand window can open and close inside a weekend.

So the better question isn’t “how did the campaign perform?” It’s “where should the next rupee, the next creative, and the next hour of team attention actually go?”

That’s the operating model Pomo is built around. And the one we think the best growth leaders are already moving toward.

AI-native Marketing Operations

Q11. India’s D2C ecosystem is scaling rapidly while operating with lean teams and rising acquisition costs. Why do you believe India is particularly suited for AI-native marketing operations?

PD: India has the exact combination that makes AI-native marketing operations valuable: complexity, speed, lean teams.

An Indian D2C brand might be managing Meta, Google, Amazon, Flipkart, quick commerce, WhatsApp, influencers, regional-language campaigns, festival moments, and its own website, all at once. That’s a lot of surface area for a five-person team.

At the same time, acquisition costs are climbing and teams are being asked to grow more efficiently. The problem isn’t “how do we create more content?” It’s “how do we make better decisions with fewer people and more channels?”

That’s where AI creates real leverage.

India rewards teams that can sense demand quickly and act before the window closes. That makes it a very natural early market for what we’re building with Pomo. The operating model that gets perfected here will travel.

Orchestration and Execution

Q12. Looking ahead, how do you see the role of marketers evolving as AI systems increasingly move from assistance toward orchestration and execution?

PD: Marketers move up the stack.

A lot of marketing work today is operational: pulling reports, checking dashboards, coordinating briefs, writing variants, chasing updates, translating spreadsheet rows into next steps. AI is going to take on more and more of that work. That’s healthy. Almost no marketer joined the field to do that part.

What doesn’t go away is judgment. If anything, it gets more important. Setting direction. Defining the brand. Understanding the customer. Choosing the right bets. Deciding when to trust the system and when to override it.

The best marketers won’t be replaced by AI. They’ll become better directors of AI systems.

That’s the future Pomo is built for: marketers spending less time operating the machine and more time deciding where the business should go.

AI marketing decision intelligence: Next Phase of Enterprise AI 

The next phase of enterprise AI may not be defined solely by model capability, but by operational integration. As digital ecosystems become more fragmented and customer attention becomes increasingly volatile, organizations are under growing pressure to make faster, smarter, and more adaptive decisions across every stage of the customer journey.

AI marketing decision intelligence reflects a broader transformation already underway across modern business operations. Rather than functioning as passive analytics layers, AI-native systems are beginning to act as continuously operating coordination engines capable of prioritizing signals, orchestrating workflows, and accelerating execution within defined governance structures.

For customer experience leaders, this shift raises important strategic questions around personalization, trust, oversight, responsiveness, and organizational design. The brands that succeed may ultimately be those capable of combining AI-driven operational speed with human judgment, customer empathy, and disciplined governance at scale.

The post AI Marketing Decision Intelligence: POMO Founder Praneet Dutta on Autonomous Marketing Operations and the Future of D2C Growth appeared first on CX Quest.