Behind the Scenes of Pecan’s Predictive GenAI | Pecan AI

Behind the Scenes of Pecan’s Predictive GenAI

We’re changing how predictive analytics is done with Pecan's Predictive GenAI capabilities. Read more from Zohar Bronfman, co-founder and CEO.

When Noam Brezis and I cofounded Pecan in 2018, the term “generative AI” wasn’t yet widely used. We were laser-focused on making the benefits of machine learning available to all businesses through an accessible, automated software platform.

We couldn’t have known then that in 2023, the AI world would be turned upside down by the introduction of a variety of generative AI tools. They’re already transforming industries and occupations. And now, they’re transforming the Pecan AI platform.

With our launch of the new Predictive GenAI capabilities in Pecan, we’re changing how predictive analytics is done. These tools combine the creative, conversational strengths of generative AI and its ability to generate code with advanced, automated predictive AI. This is a significant leap forward in our mission to democratize predictive analytics. 

Here’s a look behind the scenes at why we’ve chosen to integrate generative AI into Pecan. 

A generative guide for predictive modeling

It might be tempting to think that we’ve simply tacked on some generative AI component to ride the coattails of the GenAI buzz. But in reality, generative AI has turned out to be vital to solving some of the fundamental challenges we’ve seen data analysts and business teams face in their efforts to adopt AI. 

To be clear, we’re not applying GenAI directly to predictive modeling. As we’ve discussed on our blog before, generative AI isn’t well suited to making predictions with business data, most of which is tabular in format. It’s not the right tool for solving predictive problems on its own. However, using it in partnership with classical data science techniques — plus Pecan’s proprietary data science automation — can unlock the true potential of both technologies.

But until now, using generative AI and predictive AI in tandem hasn’t been an option for data analysts and business teams. They’ve primarily been tasked with analyzing past data, looking for patterns and trends. Using predictive analytics has been uncharted territory for many teams.

For example, here’s a common scenario today: A customer experience team dives into their past data on customers who churned in the past year to determine the right actions to reduce churn. While they might find some valuable insights, those insights will speak only to past events. 

Instead, what if this team could look into their customers’ future by using that past data to anticipate which customers might churn? Then, they could concentrate on intervening before customers were on their way out.

Adopting that future perspective — what you might call a “predictive mindset” — is a more efficient, cost-effective solution. However, teams don’t always know how best to address their business challenges with this future-focused approach. They know there’s a problem and have the data they need. Yet predictive solutions are either unfamiliar or seem too far out of their technical comfort zone to attempt without calling in data science resources (which are often unavailable or overstretched).

A simple chat launches users toward predictive success

Adding generative AI to our predictive analytics platform has made it even easier to take the leap into a predictive approach. Pecan’s new Predictive GenAI tools make it straightforward for users to identify and define opportunities for solving problems through predictive modeling and expedite their modeling. 

A user starts by simply talking about the business problem in their own words. The Predictive Chat uses a natural-language conversation with what you might consider an AI-driven “virtual data scientist” to help users define exactly the predictive question they want to address. Whatever the user’s industry or department, the chat will help them turn their business challenge into a specific, answerable question, incorporating the user’s unique considerations and concerns.

Then, Pecan guides them through the predictive modeling process from start to finish, thanks to our fusion of generative AI, predictive AI, and automation. Using the chat input, Pecan generates a Predictive Notebook with SQL queries and sample data already created for the user. No more wondering about the right SQL functions to identify the correct data for training the model. The queries are already in place and easy to customize. 

You can read about the Predictive GenAI features and watch our intro video to learn more. 

With these new tools, Pecan’s users will fast-track AI adoption. They can go from an idea to a production-ready model more rapidly than ever before, making it possible to tackle multiple business challenges effectively and achieve even greater outcomes. They’ll use the united powers of predictive and generative AI to unlock the real opportunities within their vast data repositories.

Leveraging Pecan’s Predictive GenAI to harness value

These innovative Predictive GenAI features close the divide between traditional data analytics and state-of-the-art predictive modeling, providing a unique solution for data and business teams that’s never been seen before.

Pecan’s Predictive GenAI is a significant step in the joint evolution of generative AI and predictive analytics. Join us in this exciting opportunity to experience how today’s most intriguing technologies combine forces for maximum effect.

What would you like to predict? I invite you to take the leap now and try out our platform.