He helps clients automate and improve their operational business decisions by applying decisioning modelling and artificial intelligence. Over the past five years, he has focussed on machine learning enhanced decisions, specializing in model development/deployment, data visualization, client training/mentoring and improving the integration of predictive analytics and machine learning within financial, operational decisions (e.g., Liquidity Risk Management, PEPs/Sanctions, Customer Lifetime Value and Legal Classification). With James Talyor, Jan is co-author of “Real World Decision Modeling” (the ‘big blue book’) on decision modelling, now in its second edition.

• While engaged by Deutsche Bank, Jan managed and delivered decision modelling training to over 70 liquidity risk professionals across six teams distributed worldwide using plans, courseware, and applications he developed.
• Jan developed shallow and deep learning machine learning models in Python and R, including data wrangling, visualization and feature engineering in both platforms for applications such as customer lifetime value, sentiment analysis and fact-checking.
• Recently, Jan consolidated his machine learning skills and was awarded an MSc in Machine Learning (Distinction) by Royal Holloway, University of London. He also received an outstanding project award working to improve the performance of generative AI models.
• Dr Purchase is retained to advise RapidGen Ltd’s board on the application of automated decision-making and machine learning in their products.
• Jan is the chairman of the DMN On-Ramp and a strong advocate for the use of the decision modelling standard in financial, legal and healthcare contexts.

Why DMN?

Jan is an ardent practioner of decision automation and has used Decision Model and Notation (DMN) on most of the financial and legal projects in which he has been engaged over the past eight years.

DMN provides a  , accessible framework for defining, analyzing, and managing operation business decisions. Its primary benefits are transparency, agility, and alignment with key business objectives, enabling decision-makers and analysts to capture and communicate business requirements effectively. DMN acts as a “prime record,” which ensures that business requirements are both understandable to stakeholders and deployable in real-time systems, minimizing translation errors between business and IT teams. This format empowers subject matter experts to define and adjust decision logic directly, aligning decisions with KPIs and reducing dependency on individual experts.

DMN improves operational transparency by documenting each decision’s logic and hierarchical structure, which supports accountability and traceability—both crucial in regulatory contexts and for post-hoc outcome explanations. It also allows for the integration of generative AI and machine learning into decision making in addition to decision tables and deterministic calculations.

Promoting adaptability, DMN simplifies change management by allowing subject matter experts to swiftly assess and implement adjustments without recoding. This fosters cost-effective automation, shortens delivery times, and enables quick adaptation to new data or regulatory shifts.

The standardization of DMN is widely supported across industries, with tools from over 20 vendors, ensuring model interoperability and vendor flexibility. Industries like finance and healthcare have adopted DMN, further supporting its utility in regulated sectors. Additionally, DMN’s compatibility with BPMN enhances process and decision integration, making it ideal for enterprises with complex decision dependencies.