How one expert explains the shift from raw data to real strategy with the help of AMAI

When it comes to understanding audiences, there’s a quiet revolution underway. It’s not about finding more data or building fancier dashboards. It’s about clarity, cutting through complexity and surfacing what really matters. And at the center of this shift is machine learning, paired with tools like Softcrylic’s Audience Match for AI (AMAI), which are helping brands reimagine how they segment, strategize, and communicate.

Morgan DiPietro, a Consulting Manager  deeply involved in AMAI, has watched this transformation unfold firsthand.

“At the core of this, it’s really understanding what sets this customer or group of customers apart from others,” he explains. “What makes them uniquely different? Is this customer budget conscious? Are they looking for a premium experience?”

It’s a deceptively simple question, but one that marketing teams struggle to answer with confidence. Not because they lack data, but because they lack a clear way to make sense of it. According to DiPietro, that’s exactly what AMAI helps to solve.

“We have so many data points around customers,” he says. “Being able to distill those few core things that separate one from another in a concise enough way to maintain a mental map about these audiences is really important.”

Its not about adding more complexity. It’s about subtracting the noise. AMAI’s output can be distilled into simple heuristics, quick and memorable rules that help marketers keep each audience clear in their minds. With its LLM capabilities, AMAI can help create those heuristics, making it easier to translate complex data patterns into actionable strategies.

Behavior, Not Just Belief

Twenty years ago, most companies relied on surveys to understand their customers. But those surveys had limitations since they were subjective, sample-based, and often slow. Today, it’s a different game entirely.

“We’re now working in a more revealed preference world versus stated preference world,” DiPietro explains. “So customers are revealing their preference to us via their behaviors rather than telling us their preference via survey.”

That shift, powered by digital analytics tools capturing millions of clicks per day, has created an opportunity. But it also created a challenge: how do you organize all that behavioral data in a way that’s usable, fast, and strategic?

“These modeling tools actually came first,” he says of clustering techniques like hierarchical models and neural nets. “Then we applied them to customer science once the digital ecosystem and those data capture tools matured.”

From Clicks to Clarity

Some of the most useful indicators for segmentation, DiPietro says, are surprisingly aligned with what companies already care about. These aren’t just interesting observations, they are the data points that the model relies on to separate audiences into distinct clusters. Take the airline industry, for example.

“A metric that has really helped our model separate groups [of customers]… would be something like your premium mix of seats. Out of all the routes that someone’s flown, what percentage were flown in a premium cabin?”

This behavioral metric ties directly to profitability and strategy. In other industries, the pattern holds. For wine and spirits distributors, digital channel mix (online vs. offline) has emerged as a key indicator. And across verticals, customer engagement with specific website sections, like payments, search, or curated product areas, often reveals how each group prefers to interact.

DiPietro says. “That turns out to be incredibly useful information for the model, but also really insightful information for our clients.”

Letting the Machines Do the Heavy Lifting

One of AMAI’s biggest advantages is that it automates what used to be a technical, time-consuming task. The clustering itself is push-button, powered by auto feature engineering and models designed to maximize contrast between groups and cohesion within them.

“Getting to that clustering output is only half the battle,” DiPietro says. “The next half is really where we create the value… that’s where we build a story around these audiences and where we start to build those audience heuristics into our mental maps.”

This is where the platform’s integrated language model becomes crucial. It helps analyze and summarize aggregated customer behavior across clusters. Translating data into strategy. And it does so in a way that  less-technical teams can use.

Speed to Strategy

Before tools like AMAI, audience modeling might take months and require an entire data science team. Now, the same process can unfold in days and often by less technical users.

“Some of these exercises in the past have taken potentially a very long time,” DiPietro recalls. “Audience Match reduces that time to market… and opens up the field to less technical users so more people can do it.”

Even better, the strategy doesn’t end with segmentation. Once clusters are formed, marketers can use AMAI’s LLM to explore what types of campaigns might resonate, which segments to prioritize, and how to activate them quickly.

“It really helps with getting the juices flowing and strategizing against these clusters,” he says. “We can actually ask questions about which clusters might be the most beneficial to target for certain types of campaigns.”

What Success Looks Like

In one case, a premium upsell campaign for an airline used our traditional modeling framework to target customers who had previously flown in economy class and to measure the results with confidence by using precisely matched control groups. The outcome was clear: increased premium seat upgrades, more loyalty redemptions, and higher average spend.

“Those numbers might seem small,” DiPietro notes, “but the amount of dollars at scale that those numbers can generate for a business can actually be really big.”

Smart Personalization Without Overkill

One of the most common pitfalls in personalization is over-messaging. With the data AMAI shows , and the ability to explore it conversationally, marketers can spot potential saturation points and adjust their strategies accordingly.

“If a cluster’s already sitting at 80% of their flights in a premium cabin,” DiPietro says, “we probably don’t need to consistently message them about our premium cabin offerings.”

By surfacing these kinds of insights, AMAI empowers marketers to explore adjacent opportunities, like promoting brand partnerships or exclusive lounge experiences, keeping high-value audiences engaged without overwhelming them.

Trust, Transparency, and Control

Because AMAI leverages generative AI to analyze clusters, Softcrylic has built in clear guardrails to ensure model outputs are trustworthy.

“We had to be very careful,” DiPietro says. “Users can export and verify the data. They’re also presented the summarized data for each cluster.”

Additionally, AMAI restricts the LLM to only view aggregated data, preventing hallucinations and maintaining customer privacy.

“That really helps to reduce or even eliminate hallucinations,” he says. “We’ve confirmed that through numerous tests.”

A Foundation Built on Unified Data

As with any machine learning system, quality input matters. That’s why Softcrylic emphasizes building strong data foundations before segmentation begins.

“Our data has to be in the right format. It has to be accurate and accessible,” DiPietro says. “If we can get the data and the engineering set up correctly, then we can create that unified customer group across many different channels and touch points.”

Through marketing data foundations and cross-channel stitching, AMAI enables consistent segmentation across platforms like web, email, and social.

Looking Ahead

Softcrylic is continuing to evolve AMAI with features like multimodal generation, where the platform can generate not just campaign ideas and copy, but full creative assets. Agentic AI workflows, which enable the platform to plan and execute campaigns on behalf of the user, are also on the horizon.

“Agentic workflows could end up adding a lot of value to this work stream in the future,” DiPietro says. “And multimodal capabilities would be really cool.”

The Bigger Picture

At its core, AMAI reflects a simple idea: marketers don’t need more data, they need better ways to use the data they already have. From segmentation to activation to measurable results, AMAI is helping marketing teams work faster, smarter, and with greater confidence than ever before.

“It’s one of the coolest things I’ve seen in audience science in quite a while,” DiPietro says.

And with new capabilities on the horizon, it’s only just getting started.