Back to Blog

Generative AI for Audience Building: Introducing Marve

Create Audiences from natural language using AI, right on top of your data warehouse.

Anthony Rotio

Anthony Rotio

Flywheel’s mission is to enable all teams to access customer data to drive better communications, and transform their data warehouses into enterprise growth engines.

In pursuit of that mission, we initially created Audience Builder. With Audience Builder, any team can create customer audiences from their data warehouse (like BigQuery, Snowflake, or Redshift) without SQL or data expertise. This democratizes the power of data, enabling business users on Marketing, Sales, Customer Support, and Product teams to access and utilize customer data easily for their campaigns. Now, we’re taken things a step further with Marve.

The World's First Natural Language Audience Builder

Today, we’re launching Marve – short for “marketing velocity” – to make it even easier for Marketers to create audiences for their campaigns with the full combined power of cutting-edge AI, Audience Builder, and their own data warehouse. Simply type a description of an audience in natural language and generate an audience. It’s that easy:

Marve in action. Note: this video is sped up slightly for your viewing pleasure. Actual end-to-end is approximately 45s at the time of this writing.

Once the initial audience is built via Marve, you can then edit it (just like you can with Recipes today). Since Marve is natively integrated with Audience Builder, you get all the benefits of Audience Builder out of the box: cross-channel audience syncing, central holdout groups, statistically sound incrementality measurement for performance metrics, insights panel, audience overlap and exclusions, governance, Enterprise security by design, and more. 

Marve represents a rare opportunity where the availability of a solution outpaces expectations. This technology has found practical applications that are powerful and fast, allowing enterprise businesses to leverage the power of LLMs immediately for marketing and sales. Marve plus Audience Builder flips the usual process of imagining and waiting for a solution on its head, giving teams the ability to bring their ideas to life at a pace that's unprecedented in our experience. This is a truly amazing and empowering innovation.” - Brian Himstedt, Chief Information Officer, Kansas City Royals

Why Natural Language Makes Sense for Marketers

Marketers typically organize and initialize campaigns in a document called the Marketing Brief. As a former Marketing Director at Anheuser-Busch InBev, I’ve personally written these marketing briefs for everything from national integrated marketing campaigns to one-off pieces of content. Briefs can vary quite a bit in the types of information they provide about a campaign, but they almost universally include a field that describes the “Target Audience” in natural language. 

Working with marketers at enterprises who use Flywheel today, we want to make it easier for them to go from brief to launch as quickly as possible. Faster velocity means more opportunities to launch audiences. When you pair that velocity with the ability to measure the effectiveness of those campaigns longitudinally, while they’re running, that means faster optimization and more value for our customers. 

A simplified example of a marketing brief. Notice the “Target Audience” section in the upper left corner where marketers describe their audience in natural language.

In short, natural language input for our users means meeting them where they work, so they can get campaigns launched faster, get more shots on goal, and optimize their efforts more quickly. 

When marketers can launch campaigns faster, it creates an opportunity to optimize faster towards outcomes that matter. Marve represents a new paradigm in connecting marketers directly to the customers they’re hoping to target, using AI to take full advantage of the data available in the data warehouse, to drive better communications.” - Victoria Vaynberg, Chief Marketing Officer, Zola

How Marve Works - LLMs for Audience Building and Customer Segmentation

Marve leverages the best in class large language models–including PaLM2 from Google–which have been trained on the entirety of the internet and refined using feedback from human experts. We’ve taken those models and integrated them into our platform in a way that makes it easy and effective for a marketer to input text in a way that they would naturally describe an audience, without necessarily knowing the names of columns in data tables or how the underlying data is structured. 

For example, since the underlying Transformer models have effectively read the entire internet they can infer that “tri-state area” in our example above means we’re talking about the states “New Jersey,” “New York,” and “Pennsylvania” and that "bucks" probably means dollars in this context. We’ve then done further work to nail some of the translation into the underlying data structure and a format that Audience Builder can understand. 

While LLMs are not perfect 100% of the time, this is where the precision of Audience Builder comes in. A user gets the benefit of interfacing in natural language with Marve, and in the rare cases where the audience doesn’t come out exactly right, they can review and refine in Audience Builder prior to export, if desired. This architecture creates a benefit to the user where the combination is better than the sum of its parts. Our users get the speed and intuition of natural language input with the precision and measurement capabilities of Audience Builder.

Here is a full demo of Marve, including a voiceover for more context:

Today’s launch of Marve extends Flywheel’s mission to democratize access to organizational data. No longer do users need to be intimately familiar with the structure or available fields in their data warehouse (which could number in the thousands), which opens the door for collaboration and creativity. We see Marve being very useful for anyone involved in the marketing process from the creative agencies to higher level executives. 

Marve is available in Limited Beta today. Current customers can contact your Solutions Architect to learn more and prospects can fill in the form below to be added to the waitlist:

Share on social media: 

More from the Blog

How to maximize ROAS without third-party cookies

How to maximize ROAS without third-party cookies

Digital advertising is an essential way to reach customers across their many online channels — but evolving privacy regulations have made tracking and targeting audiences harder than ever. This guide provides best practices and resources for improving ad performance in our privacy-focused era.

Press Releases
GrowthLoop Launches Computed Attributes generative AI feature powered by AWS to increase productivity and deliver hyper-personalized campaigns

GrowthLoop Launches Computed Attributes generative AI feature powered by AWS to increase productivity and deliver hyper-personalized campaigns

The solution allows marketers to execute highly-personalized, data-driven marketing campaigns in a fraction of the time with significantly less overhead. 

Cloud Data Warehouse
Guide: Using the GrowthLoop composable customer data platform on Amazon Redshift

Guide: Using the GrowthLoop composable customer data platform on Amazon Redshift

Discover how an AWS customer data platform (CDP) will help your marketing team create targeted campaigns faster, improving revenue and ROI.

Looking for guidance on your Data Warehouse?

Supercharge your favorite marketing and sales tools with intelligent customer audiences built in BigQuery, Snowflake, or Redshift.

Get Demo