Predictive data analytics startup Pecan AI Ltd. today announced a new capability called Predictive GenAI — and as the name suggests, the new feature combines the company’s artificial intelligence-powered models for business predictions with the generative AI features found in models such as ChatGPT.

The company claims the new feature is one of the first genuine enterprise use cases for generative AI, combining it with predictive AI to unlock greater value from businesses’ troves of customer data.

The startup is best known for its user-friendly data science platform, which enables companies to easily build AI models for tasks such as predicting customer demand and finding opportunities to improve business efficiencies. Its platform automates the development of these models, while reducing the need for data science skills.

Using Pecan AI’s models, companies can generate individualized predictions for every customer and bring these insights to the customer relationship management platform of their choice, such as Salesforce, Dynamics 365, Freshworks or something else. The platform automates both the creation and deployment of these predictive AI models, and can also provide insights into their performance, monitoring them for signs of degradation. This is important because such models can become less accurate over time from changes in customer behavior, so they need regular updating.

With its Predictive GenAI feature launched today, Pecan AI says it’s making predictive modeling more accessible than ever before, allowing any business user to harness its capabilities using simple, natural language commands.

There are two elements to the new features, namely Predictive Chat and Predictive Notebook. The company said users can kick things off with Predictive Chat, simply telling what is essentially just a chatbot about the business problem or challenge they are attempting to solve. Predictive Chat will then help them to define the predictive model required to solve this problem.

Once the model has been roughly outlined, Pecan AI will generate a Structured Query Language-based Predictive Notebook that contains all of the necessary queries and sample data for creating the business-relevant predictions. Users will then have an opportunity to refine this modeling approach before letting the platform build and deploy the model automatically.

Once the new model is up and running, Pecan AI says it will generate accurate predictions that enable users to take action to try and solve their business challenge and improve outcomes. The startup says it can translate user’s business needs into accurate, predictive models regardless of the type of business challenge. As an example, it said it can predict the risk of customer churn, or the chance that a critical piece of manufacturing machinery might malfunction.

Holger Mueller of Constellation Research Inc. said it was always only a matter of time before someone attempted to use generative AI to create other kinds of AI. “This is what Pecan AI is trying to do now, and the success of this effort will most likely depend on the quality of the data being made available to its generative AI assistant,” he said.

The startup reckons this is one of the most viable enterprise use cases for generative AI available thus far, providing a significant opportunity to scale the business value that resides within companies’ datasets.

According to Pecan AI Senior Vice President of Product Limor Segev, many businesses struggle to solve problems even when they have predictive modeling capabilities at their disposal, because of the lack of know-how on how to leverage these tools. “We have eliminated the need to learn complex code, invest in expensive hires or enlist the help of a third party,” he promised. “Predictive GenAI unlocks transformative business capabilities through an entirely new path to building data science models.”

Image: Freepik