First Look: Pegasystems Update 2013

April 8, 2013

in Analytics, BPM, Business Rules, Data Mining, Decision Management, Product News

Share

Pegasystems had good years in 2011 and 2012 they say, with Q4 2012 being particularly strong and good results from their decision management products. They have over 2,000 employees as they go into their 30th year and are targeting $500M this year. Pegasystems maintains its focus on a platform business with a number of specific verticals although with an increasing focus on channels. They continue to leverage widespread frustration with IT and legacy software to focus customers on a new platform for technology-led innovation. Pegasystems is expanding and improving its channel partners – certified partner are up and a number of partners are beginning to sell directly with a number of large SIs taking the lead in growing a much larger partner pipeline. Pegasystems’ focus remains its integrated platform (BPM, Case Management, CRM, application development and Decision Management) across marketing, sales and service with a particular focus on companies that are moving to next best action. They continue to add industry applications outside their traditional industries (like manufacturing and CPG for instance) and deliver both on premise and in the cloud.

For 2013 they (again) want to make the product twice as easy to use and twice as fast to deploy – a constant challenge they set themselves. Social, mobile and case management are focus areas, combined with next best action everywhere. The mobile capabilities include a full app development platform that takes advantage of the rules in Pegasystems to better manage the user interface experience (location and device-aware). Solution-wise, Next Best Action Marketing and SFA were released in 2012 (adding to existing Customer Service options) as new solutions based on the stack taking advantage of next best action engine.

As far as Decision Management goes Pegasystems has spent the last couple of years truly integrating the old Chordiant decision management capabilities into the process/case management platform based on Pegarules. The decision strategy editor is now browser based and integrated with the rest of the Pega editors. Decision tables and decision trees from Pega have been retained and elements from the Chordiant product, like champion/challenger rules, added. The integrated environment supports 5 main decision rule types are:

  • Predictive Model rules (from Predictive Analytics Director or built externally)
  • Scorecard rules
  • Decision Strategies defined using the editor
  • Adaptive Models that learn while predicting
  • Interactions.

The Decision Strategy editor has been revised and allows a complete decision flow to be laid out with predictive models or adaptive models, decision tables or trees, champion/challenger etc. Strategies can be chained or drilled down into more detail. Every element has version control, security and access controls.

Decisions in process flows can be directly linked to a decision design in the Decision Strategy editor. Processes can also be flagged to capture responses to decisions made later in the process, potentially after a long time, and automatically feed these responses back into adaptive models.

Recent updates in Decision Strategy Manager since last year’s first release of a unified decisioning platform include support for large scale batch decisioning, increased PMML support, forecasting in Visual Business director and more advanced segmentation.

As noted Pegasystems has had success selling the new integrated Decision Management capability with a number of new wins in retail banking and other direct to consumer businesses such as TV and Telecommunications. One example customer (a telco) identified several benefits from the use of the Pegasystems Decision Management product including more relevance in offers (real-time context that includes customer intent, optimized logic across all potential actions including built in champion-challenger testing), more intelligence (handling business constraints such as service problems or busy call centers) and more control (real-time reporting and dashboard-based control of logic using Visual Business Director). This particular customer identified hundreds of millions in revenue from using the Pegasystems Decision Management product as well as a significant improvement in net promoter score each time new decision management capabilities were rolled out – more sales and an improvement in net promoter score thanks to more targeted, more personalized micro decisions. This use case, like most use cases for decision management, is focused on customer decisions but Pegasystems reports an increasing number of customers focused elsewhere such as incident severity and suitable responses to problems.

The unified Pegasystems platform:

  • The unified platform is based on a single enterprise repository (all cases being processed are managed in a unified case database also).
  • The base of the stack is a set of persistent object services that handle the object-relational mapping for the repository and cases.
  • Business rule services sit on top of these objects for handling user interface, forward and backward chaining etc.
  • Decision Management Services layer this with capabilities around predictive scorecards, decision trees and support for predictive models (both those built off-line and others using adaptive analytic technology).
  • Business Process Management Services can consume these decision management services and can be called by them.
  • Dynamic Case Management extends this set of capabilities to manage unstructured or dynamic processes and can also consume the decision management capabilities to guide case workers, for instance.
  • A layer of presentation capabilities supports various channels from portals to social and mobile.
  • Continuous improvement capabilities in terms of business activity monitoring etc wrap it up.
  • The system is available on the Cloud and can be licensed both on premise, on the cloud or in a hybrid deployments

The design environment remains browser based and a set of APIs for integration and event monitoring are available. Pegasystems’ domain-specific solutions use the stack and use cases include fraud detection, routing, escalation and case management tuning, automated approval across various industries.

One of the key elements of the Decision Management stack is Visual Business Director. This is now available as part of the Pegasystems platform and anything running on the Pegasystems platform can be instrumented for use in VBD. This allows process volumes and details to be presented alongside decision information for example. Sections of processes can be identified for simulation, allowing a change in decision making to be simulated in terms of its impact on a defined subset of the process than is executed as a consequence of the decision. What-if analysis can now include process implications of a change in decision-making.

Predictions and forecasts have also been extended. VBD could previously support back-testing, comparing a new strategy with what was seen in the past, and can now be used to simulate and forecast likely consequences of, for instance, outbound campaigns in terms of business results as well as call center loading and volumes. This uses projections from historical data.

Pegasystems’ Predictive Analytics Director is a capability for building predictive analytic models that supports the entire model creation process (though it is targeted more at business analysts rather than expert modelers) while Adaptive Decision Manager is for model tuning (and can automatically tune models built in Predictive Analytics Director as well as those brought in from a third party tool). ADM can also, and typically does, build models from scratch.

Since I last wrote about Predictive Analytics Director, they have improved the validation process by allowing a user to compare models based on various data sets at any point in model building and deployment. The final validation can be done against data never used in the modeling process and detailed reports generated. The process has been templatized so that projects can decide where they want to be on the default-settings to custom analytic project scale, supporting a wide range of potential users. Usability and visualization has been improved too but the major focus has been on integrating fully with the Pegasystems stack so that, for instance, all the model meta data is carrier forward to both document the model and support its automated monitoring and tuning.

Future plans reflect their focus on multi-channel solutions and include more sophistication in attribution (to allow adaptive models to correctly attribute what convinced a customer to the right channels) as well as support for managing the interaction history of a customer more centrally.

Pegasystems is one of the vendors in our Decision Management Systems Platform Technologies report and you can get more information on them here.

Share

Previous post:

Next post: