With nearly two decades of experience driving innovation at the intersection of AI, data engineering, and marketing technology, Angshuman Rudra has led product teams through major digital transformations at companies including Adobe, Yahoo, and TapClicks.
As Director of Product Management at TapClicks, a $45M ARR marketing analytics company, he has spearheaded the development of AI-powered marketing infrastructure that serves over 5,000 agencies and brands.
Angshuman’s approach combines deep technical expertise in data platforms with a strategic focus on AI agents, ETL solutions, and scalable MarTech architecture.
In this interview, he shares insights from building widely used products that redefine how marketers work with data, his perspective on the convergence of AI and marketing operations, and how organizations can modernize their data infrastructure to drive measurable business outcomes.
Q. How are AI agents transforming marketing operations, and what are the most significant improvements you’ve observed in organizations that have successfully implemented them?
Angshuman Rudra: There are two key directions where AI Agents will transform Marketing Operations: insights and automation.
For insights, we’re already observing substantial value as AI agents leverage their extensive knowledge to provide a deeper understanding of customer interactions. They excel at tasks such as summarizing messages and content, personalizing content to individual preferences, and enriching customer profiles with more granular details and predictive understanding. This capability extends to distilling unstructured data, such as call recordings or chatbot transcripts, into structured, actionable insights that traditional methods might miss.
For automation, we’ve begun to see AI agents capable of automating specific workflow segments, particularly tasks such as summarizing messages and personalizing content, and then taking downstream actions like pushing to a CRM or triggering an alert. Organizations are increasingly achieving this, with greater success evident in those that already possess a robust data infrastructure, enabling AI models to access data readily. Companies at a higher level of data and analytics maturity are already benefiting from this more. One thing to note here is that at the enterprise level, a critical focus remains on establishing or refining underlying infrastructure to ensure the security, reliability, and data governance necessary for AI agents to deliver optimal outcomes.
The future will be marked by AI agents that can solve higher-level goals, reason, and interact with each other to complete those goals, ultimately leading to higher levels of autonomy for various workflows.
Q. What role does modern data engineering play in enabling effective MarTech solutions, and how has this landscape evolved over your 19+ years in the industry?
Angshuman Rudra: The MarTech ecosystem has expanded significantly over the past 10-15 years. The first Marketing Technology Landscape by Scott Brinker in 2011 listed 150 products, and by 2024, that number had grown to over 14000. Large enterprises can have as many as 100 MarTech software programs.
Each software have/had its own database with a certain data architecture. And I think one of the biggest problems to solve for the last few years has been extracting this data from the various data silos and getting to a single unified view that can then be used for additional insights. I will say that is the single most important role data engineering has played and will continue to play in the coming years.
Some of the native warehouse-related architectures has been refreshing and more and more companies are adopting a cloud data warehouse to house all of their data into a centralized data repository. And, for this future to succeed a data engineering has a huge role to play as well.
Q. What are the biggest challenges organizations face when building scalable marketing data platforms, and how can they overcome common pitfalls in ETL and data transformation?
Angshuman Rudra: Data silos and not having a way to connect data across systems to take informed decisions is one of the key challenges. Even if there is a single unified data layer, maintaining it and making sure the right data models are built is still a huge challenge.
Another challenge I see is inconsistent data hygiene, not having data dictionaries, and not enriching a data set (for it to actually add value).
Q. How should marketing teams approach integrating AI capabilities into their existing workflows, particularly when it comes to campaign optimization and performance analysis?
Angshuman Rudra:
There are two key aspects to this:
Data Strategy: First, let’s talk about your data. AI is only as powerful as the data it’s fed, so you need to think comprehensively. Consider what data you need across your campaigns, audiences, budgets, and outcomes. How is that data structured? It’s crucial to have clean architecture, clear documentation of where the data comes from, and consistent rules for how it’s transformed. Also, decide when and how often data flows – whether it’s real-time or in batches – and ensure consistency across all your platforms. A solid data foundation is essential for trustworthy, AI-driven insights.
Automation/API Strategy: Once you’re generating insights, the next step is acting on them. This is where automation comes in: AI shouldn’t just provide insights and a plan; it should initiate actions as well. To make this happen, you’ll need a strong underlying API layer that allows AI agents or rule-based engines to trigger tasks. These tasks should be directly tied to your real campaign workflows, like reallocating budgets, prioritizing channels, or updating creative.
Q. From a technical perspective, what are the key considerations when architecting AI-powered marketing analytics platforms that need to handle massive data volumes from multiple channels?
Angshuman Rudra There are three main technical pillars I like to focus on:
Data Strategy: First, you need a robust data strategy. In today’s world, your marketing and revenue data and other relevant data can be locked into various tools and silos that do not talk to each other – CRM, MAP, Ad Platforms, Product Analytics, Web Analytics tools, etc. Key elements here include establishing a way for the data to talk to each other. Some simple ways to start are consistent naming conventions and ensuring your data transformation logic aligns with your business rules. Don’t forget data lineage for full traceability, and flexible data refresh frequencies to accommodate both real-time and batch use cases. Without this strong data foundation, your AI insights simply won’t be accurate or actionable enough.
Automation Strategy: Next up is your automation strategy. The goal here is for AI to drive action, not just provide observations. Architect your system to intelligently detect insights, such as pacing issues or drops in performance, and then automatically trigger actions through APIs or workflow engines. This approach enables continuous optimization across your campaigns, reducing/eliminating human intervention when it’s not needed. This is still not a common occurrence because of additional complexity surrounding data governance and permissions in larger enterprises, as well as a lack of the necessary infrastructure in smaller organizations.
Meta-data Layer as the back-bone: This is a corollary of the data layer but absolutely critical – the metadata layer. AI models require context to truly shine. Therefore, structuring your data with rich metadata – including taxonomy mappings, platform-specific nuances, and campaign hierarchies – is very important for generating meaningful insights. Consider this metadata layer as the essential translator between your raw data and intelligent decisions. It allows you to unify logic across fragmented systems, making your AI explainable and, most importantly, trustworthy. There are different ways you can achieve this, and you can start by documenting some of the business logic and workflows and making sure your team and your AI Agents know about it.
Q. How is the convergence of AI, data engineering, and MarTech reshaping the way agencies and brands approach campaign management and client reporting?
Angshuman Rudra: Some of the key changes I am seeing in campaign management are – hyper-personalization, automated workflows for the ad-buying process and other workflows and improved targeting and segmentation. What is happening even now with AI is teams can analyze vast customer data to deliver real-time, tailored content, recommendations, and ads across channels, with generative AI creating copy and visuals. Entire workflows of Ad buying and media planning are getting automated and this trend will continue to grow.
Another aspect that falls between campaign management and reporting is how was success measured and used to determine next steps. The old metrics of click-through rates, shares, and views are no longer sufficient. Advertisers are increasingly focused on real financial returns, demanding to know “how much sales did my ad spend deliver?” And Ad platforms are using (and will use) AI to provide financial attribution.
In client reporting, the core change comes from leveraging massive, integrated datasets. And generating insights from that data and summarizing these large reports to provide very specific recommendations. A lot of this is possible because of the underlying data layer and data engineering, through cloud data warehouses/lakehouses, which provide this universal data layer, bringing together both structured and increasingly unstructured data (like call transcripts or social messages). AI then analyzes this wealth of data, offering richer insights and enabling optimization of return on marketing spend. The end result is much more context-aware reports with insights.
Q. What metrics and KPIs should organizations focus on when measuring the success of their AI and data engineering investments in marketing?
Speed of Execution and Experimentation: AI accelerates the marketing cycle by reducing time-to-market, enabling faster campaign launches, and increasing the frequency of testing. KPIs like iteration velocity, average time to launch, and number of experiments per quarter reflect this agility. This aligns with the principle of compound marketing – where small, fast gains compound into major outcomes over time.
Outcome Quality and Marketing ROI: This is the one I am most excited about. I think a lot of metrics that get measured and provided to Executives today are lower-level metrics. With AI lowering the cost of ideation and production, teams can test more ideas, leading to a higher volume of winning campaigns. I think we will get closer to getting financial attribution and be able to track metrics like success rate of campaigns, conversion uplift, CPA, LTV, and marketing ROI
Operational Efficiency and Decision Intelligence: AI reduces manual, repetitive tasks – freeing teams to focus on strategy and creative thinking. Metrics should include time saved on content creation or reporting, reduction in operational overhead, and improved decision accuracy from AI-driven insights and personalization. Ultimately, this reflects AI’s role in making marketing more intelligent and less burdensome.
Q. What trends do you see shaping the future of marketing technology, and how should organizations prepare for the next wave of innovation in AI-driven marketing?
Consolidation – In the next 2-3 years, I expect consolidation across the MarTech landscape. AI will be a major driver – whether through standalone native AI platforms or existing products embedding AI capabilities deeply into their workflows. I think composability and cloud warehouses will be a factor too. We can already see consolidation in the CDP market as a lot of the CDPs’ system-of-records capabilities are taken over by cloud data warehouses, and the system of actions features will get consolidated.
Composable – Warehouse-Native Architecture – I am a huge fan of Composability and excited to see it becoming more mainstream. Tools that integrate natively with cloud data warehouses and allow teams to build modular, interoperable stacks will have an advantage (It does require a data engineering and some architect to think this through and many organizations are still not at that maturity yet). I personally believe companies should adopt this approach, as it enables greater flexibility, improved cost control, and faster innovation cycles.
Unified Data Layer will be a Competitive Advantage: The most overlooked trend is the growing importance of mapping decisions back to clean, contextualized data. Organizations need to invest in incrementally improving how data is captured – across campaigns/channels, touchpoints, product analytics, CRMs, etc. – and how it’s unified into a consistent layer.
This unified data layer becomes the foundation for AI, personalization, attribution, and optimization to actually work.
Q. What best practices would you recommend for marketing teams and agencies looking to modernize their data infrastructure and implement AI-driven solutions?
Angshuman Rudra: To modernize data infrastructure and implement AI-driven solutions, marketing teams and agencies should adopt a focused, iterative approach.
Begin by defining clear business objectives and the specific decisions that need to be made to achieve them. Next, identify the evidence required to support these decisions, and subsequently, determine the precise data needed to build that evidence. This top-down mapping ensures efforts are aligned with strategic goals.
Once this foundational understanding is established, proceed to build a Minimum Viable Product (MVP) of your data infrastructure, focusing on a unified data layer or system of records. Validate this system by observing improvements in your “Time to Insight” and your ability to make data-driven decisions that were previously impossible. This iterative process, continuously adding layers of data to your unified data layer, is crucial for success. The specific tools or processes can be adjusted later; the key is to prioritize foundational thinking and build incrementally for optimal outcomes.
Q. Based on your experience helping to growing TapClicks from a startup to a $45M ARR company, what advice would you give to other product leaders building MarTech solutions in today’s competitive landscape?
Angshuman Rudra: Marketing will always have opportunities; marketing will always be a complex and evolving problem. Businesses constantly need to acquire new customers and retain or grow existing ones. Because there’s no single formula for success, there’s always room for innovation. Every shift in platforms, behavior, or regulation creates opportunities for new players to solve newly emerging problems.
Following Trends: From small businesses to large enterprises, marketing needs vary significantly based on the target audience. Add to that the speed of change in consumer behavior and tech (as an example – ChatGPT disrupting local search intent from Google) – product leaders in this space must stay on top of consumer and tech trends.
If a new category is emerging (e.g., creator-led commerce, AI-generated content, zero-party data), align your product with it. Paradoxically, a crowded space often signals a strong market need; I consider this an opportunity for newcomers, not a red flag.
Focus on your ICP: This may sound cliché, but it’s foundational.
Start by defining your Ideal Customer Profile and gaining a deep understanding of the problem you’re solving for them. Pick a problem that: Is high-priority for your ICP, they’re actively trying to solve (with or without you), and they’re willing to pay for.
Immerse yourself in that space by talking to users, studying their existing workflows, and building solutions to address those problems.
In our case, the ICP was mid-market marketing agencies trying to solve data management problems at scale. We kept immersing ourselves in the space, attempting to solve their issues and continuously expanding our Product Market Fit.
