.svg)
Introduction
Sreenath Reddy:
Welcome back to the Intentwise Connect podcast. I’m your host, Sreenath Reddy, founder and CEO of Intentwise. Today’s episode is coming to you live from the floor of Amazon unBoxed—Amazon’s marquee advertising conference of the year. We’re diving into the latest innovations from Amazon Marketing Cloud, and for that, I’m joined by a very special guest: Aby Angilivelil, Director of AMC. Aby, thanks for being here.
Aby Angilivelil:
Great to be here. Hello everyone—I’m Aby. I’ve been with Amazon for over 25 years and have spent the last several years leading the Amazon Marketing Cloud organization. I started as a software development engineer, moved into the ads business six years ago, and now lead the AMC team. We’ve been inventing a lot on behalf of advertisers, and I’m excited to talk through what we’ve launched.
AMC’s High-Level Strategy
Aby:
At a high level, our focus is on making AMC easier to use, expanding the signals available, and helping advertisers combine Amazon data with their own first-party data. Advertisers come to AMC for the richness of signals—from impressions to conversions to shopper-level interactions. Our job is to simplify access to those signals and make it easier to extract value.
Major Announcements: The Evolution of Ads Agent
Sreenath:
One of the big innovations this year is AMC’s Ads Agent. Walk us through how it developed.
Aby:
We’ve been iterating on this for more than a year. We launched the first version at CES with natural-language support for audience creation. We started there because audiences are finite and it was an easier training set—though still a lot of hard work.
Next, we added product-related questions. And at unBoxed, we announced the biggest step: natural-language measurement queries. Measurement is the most complex area—multiple signals, complex joins, and nuanced definitions—so getting this right was a major milestone.
Sreenath:
Our team jumped on this when it launched. Early accuracy was around 50–60%, but it’s improved dramatically.
Aby:
That improvement comes from building a strong “truth set” of valid queries and partnering closely with AWS and our sales organization. Analysts reviewed outputs, gave feedback, and helped refine the model sprint after sprint.
How Advertisers Are Using the Ads Agent
Aby:
Since launching, more than 70% of audiences are now created through natural-language prompts. That shows both ease of use and strong repeat engagement.
We’re still early in AMC adoption overall—many advertisers are just beginning. Common questions we hear include: “Where do I start?” and “Which table do I use?” Simplification solves that.
The Role of Analysts and What’s Next
Sreenath:
Can users fully trust the output, or should analysts still verify queries?
Aby:
In the short term, verification is a good practice. But in the long term, the goal is: enter a prompt, and the system executes the query automatically with no intervention.
Today, Ads Agent reduces time-to-query from days to minutes. Long term, we want to eliminate SQL checks entirely.
Expanding Signals: Retail, Traffic, and More
Aby:
This year we launched one of the most impactful datasets: five full years of Amazon retail purchase data. It allows advertisers to understand long-term lifetime value, repeat purchase behavior, and upgrade cycles—especially valuable for 1P vendors who’ve never had this visibility before.
We also just announced the expansion of traffic signals from 13 months to 25 months. That’s a major unlock for longitudinal analysis. We plan to expand other datasets over time, balancing usefulness with storage and processing costs.
We’re also investing in vertical-specific signals. Last year we launched Amazon Garage for automotive advertisers to understand vehicle ownership and upgrade windows. More are coming.
How Brands Should Get Started
Aby:
If you advertise on Amazon, this is a no-brainer—you should already be using AMC. Access is now self-service. Just go to the left navigation in Ad Console, open Measurement & Reporting, and select Amazon Marketing Cloud.
You’ll find getting-started templates for sponsored ads, prebuilt queries, and now natural-language support. For deeper analysis, partners like Intentwise play a crucial role in helping brands operationalize the data.
The Role of Partners Like Intentwise
Aby:
Advertisers often ask: “What’s next?” or “How do I become an advanced user?” The answer is personal—each advertiser has unique goals and data structures. That’s where partners matter.
Amazon focuses on the common building blocks: signals, APIs, clean room capabilities, modeling frameworks. Partners like Intentwise extend that into custom business logic, multi-brand rollups, advanced ML models, and first-party data integrations. That’s where a lot of real value is created.
Closing Thoughts
Sreenath:
We’ve had a fantastic partnership with your team and appreciate the pace of innovation—it keeps us sharp and unlocks tremendous value for our clients.
Aby:
It’s been a great journey, and it’s still just day one. We have much more coming as we continue to simplify AMC, expand signals, and help advertisers unlock deeper insights.
Sreenath:
With that, we’ll wrap. Thanks for joining us live from unBoxed, and safe travels home.