By automating model deployment, Pecan takes critical next step in bringing data science to BI analysts, marketers, and other business users

Pecan AI, the leader in AI-based predictive analytics for BI analysts and business teams, announced the addition of one-click model deployment and integration with common CRMs, marketing automation, and other core business systems. Pecan’s customers can now take immediate actions based on the highly accurate predictions for future churn, lifetime value, demand and other customer-conversion metrics generated by Pecan.

In addition, Pecan added live model monitoring to its automated predictive analytics platform for non-data scientists. The platform now continuously monitors live models for signs of degradation related to factors such as internal changes in consumer behavior or external changes in data integrity. This process ensures predictions maintain a high level of accuracy and generate an improved uplift over the rules that were previously used. Models can now be deployed quicker and easier without requiring any support from data engineers, in addition to becoming more useful over time – enabling BI and marketing analysts to continuously monitor them for external changes in the data (such as drift and leakage) and empowering analysts to correct course as needed.

Marketing Technology News: Rarify and BitGo Partner to Provide Investors With NFT Analytics and Help Them Gain an Edge

Pecan is a first-of-its-kind data science platform for business teams and their SQL-skilled data analysts that automates the creation of highly accurate, ready-to-use predictive models focused on key customer journey KPIs without data scientists on staff. Once these models are deployed, they generate individualized predictions for every customer and send the information directly to the client’s system of choice, such as a CRM or a Mobile Measurement Partner (MMP), to orchestrate precise, future-informed actions such as serving an ad with a special offer only to customers who will reach VIP status in the next 45 days. This enables organizations to replace a simplistic BI-based logic with sophisticated and accurate AI-powered predictive scores. Pecan also automates what typically is a very manual and lengthy data engineering and model creation and evaluation process. This is accomplished through an easy-to-use drag-and-drop interface, SQL queries, and state-of-the-art machine learning algorithms and techniques.

“Predictions generated with Pecan have a direct and ongoing impact on revenue-generating activities with companies representing the full gamut of products and services,” said Noam Brezis, co-founder and CTO of Pecan AI. “Data science models typically take many months and quarters to build, train and test – and that’s not counting the additional months it takes multiple data engineers and data scientists to connect, clean and prep the data for AI. Even when the original model works seemingly great with test data, many deployments simply fail. By automating the process and ensuring models are easy to deploy and monitor for value delivery against business goals, Pecan’s latest platform enhancements are helping customers not just build and successfully deploy more production-grade models, but are also ensuring they yield better quality predictions, saving those organizations time and money.”

Marketing Technology News: MarTech Interview with Debjani Deb, CEO at ZineOne

VentureBeat found that 87% of data science projects never make it into production, and earlier Gartner research reported that 85% of big data projects fail. This is due to a variety of factors including the complexities of data cleansing, misalignment on business objectives, a lack of data science resources, and improperly scoping the support needed from data engineering. The Pecan Predictive Analytics Platform bypasses all of this, enabling business teams to build and deploy working predictive models within a few weeks without requiring any coding nor data scientists’ or engineering support.

Live, accurate model monitoring and the ability to automate deployment is possible due to the platform’s automated label engineering capabilities. With most other data science platforms, data scientists are required to create, monitor and update labels on an ongoing basis. With Pecan, labels are created based on the specific use case. By acting as the label engineer, Pecan provides deeper insights into what is happening in their customers’ business and how predictions are actually impacting their business decisions.

With the addition of these latest capabilities, the Pecan platform now automates all aspects of model creation and deployment, including cleansing the data, building the models, ensuring continual data input and output to client systems, and MLOps. With thousands of models deployed in production, Pecan is now generating over 30 million daily predictions impacting billions of dollars in revenue for its customers.