Emerging Technologies in BPM

Transcription

1 Emerging Technologies in BPM Keynote: Emerging BPM Techniques & Technology Summit Building Business Capability 2012 Sandy Kemsley

2 Emerging BPM Techniques & Technologies Summit The Hurricane Sandy edition Thinking on the Job: Adaptive Case Management in Practice [cancelled] Modeling and Analytics for Process Excellence [speaker replaced] Process Mining: BPM Upside-Down [speaker arriving from Europe 9pm tonight] Copyright Kemsley Design Ltd.,

3 Technology: Social BPM How Social Changes Everything Copyright Kemsley Design Ltd.,

4 Consumer Tools Set Expectations Consumption Participation Creation User experience Access anywhere Copyright Kemsley Design Ltd.,

5 Social BPM Business Benefits Weak ties/tacit knowledge exploitation Knowledge sharing Social feedback Transparency Participation Activity and decision distribution (crowdsourcing) Copyright Kemsley Design Ltd., Source: Brambilla et al, A Notation for Social BPM

6 Collaborative Process Modeling Multiple people participate in process discovery, modeling and documentation Internal and external participants Technical and non-technical participants Preserves institutional memory Facilitates cross-silo collaboration and innovation Copyright Kemsley Design Ltd.,

7 Process Event Streams Timeline of activity for social monitoring Process models during creation Process instances during execution Publish/subscribe model to watch certain processes or event types Direct link to underlying process model or instance for unsolicited participation Usually mobile-enabled Copyright Kemsley Design Ltd.,

8 Technology: Dynamic/Adaptive Case Management The Changing Nature of Work Copyright Kemsley Design Ltd.,

9 The Extremes Of Work Routine Work Knowledge Work Copyright Kemsley Design Ltd.,

10 Goals Of Work Types Routine Work Efficiency Accuracy Process improvement Automation Knowledge Work Flexibility Assist human knowledge work Collect artifacts Classic BPM Adaptive Case Management (ACM) / Production CM / Dynamic CM Copyright Kemsley Design Ltd.,

11 Characterizing The Extremes Routine Work A priori process model Controlled participation Automatable, especially with service integration, rules and events Knowledge Work No a priori model Collaboration on demand Little automation, but guided by rules and events Copyright Kemsley Design Ltd.,

12 The Structured/Unstructured Debate If you can t model it up front, you just don t understand the process Exceptions are the new normal: every process is different Copyright Kemsley Design Ltd.,

13 But It s Not That Simple Structured Work Some process are that repeatable, especially automated processes Ad hoc process exceptions already exist, they re just off the grid Unstructured Work Some processes have sufficient variability that modelling is inefficient Instrumentation of unstructured processes provides value Copyright Kemsley Design Ltd.,

14 Structure Spectrum Structured e.g., automated regulatory process Structured with ad hoc exceptions Unstructured with pre-defined fragments Unstructured e.g., investigations e.g., financial backoffice transactions e.g., insurance claims Copyright Kemsley Design Ltd.,

15 Dynamic Process Runtime User can add participants from own network or recommended expert Non-participant can opt-in to process Audit trail captured within BPMS Eliminates uncontrolled processes Captures patterns for process improvement Copyright Kemsley Design Ltd.,

16 Technology: Process Mining Discovering Hidden Process Gems Copyright Kemsley Design Ltd.,

17 Process Mining Sources Copyright Kemsley Design Ltd.,

18 BPMS Event Log Format Trans. ID Activity Start Time End Time Resource 8287 Enter customer data 08:34:15 08:37:44 User jsmith 8287 Check credit 08:37:52 08:38:05 Equifax service call 1399 Enter customer data 08:37:59 08:44:40 User sjones 8287 Enter order 08:38:09 08:38:39 ERP system call 1399 Check credit 08:44:58 08:45:06 Equifax service call 4283 Enter order 08:45:01 08:45:35 ERP system call 1399 Enter order 08:45:18 08:45:38 ERP system call Copyright Kemsley Design Ltd.,

19 Combining All Event Logs Trans. ID Activity 8287 Enter customer data 8287 Create customer record 8287 Create address record Start Time End Time Resource 08:34:15 08:37:44 User jsmith 08:34:25 08:35:55 User jsmith 08:36:12 08:37:39 User jsmith 8287 Check credit 08:37:52 08:38:05 Equifax service call 8287 Enter order 08:38:09 08:38:39 ERP system call 8287 Check PO 08:38:10 08:38:15 System 8287 Create order 08:38:18 08:38:31 System Copyright Kemsley Design Ltd.,

20 Generating A Process Model Copyright Kemsley Design Ltd.,

21 Generated Model Data Source: Fluxicon 21

22 Working With Process Mining Results Actual flows, not idealized models Frequency and duration of each path Optimization: Detect main flows and common variations Detect loopbacks and other inefficiencies Detect wait times Analyze variations over time Copyright Kemsley Design Ltd.,

23 More On Process Mining Process Mining: BPM Upside-Down Thursday, 11:30am, Diplomat 5 Anne Rozinat Fluxicon Copyright Kemsley Design Ltd.,

24 Technology: Process Simulation Charting A Course In Uncertain Conditions Copyright Kemsley Design Ltd.,

25 Model-Simulate-Analyze-Optimize Copyright Kemsley Design Ltd.,

26 Simulation Goals Test and validate process models Establish path patterns Estimate end-to-end times Optimize resource utilization and SLA performance across peak/slack periods During runtime, predict performance based on realtime analytics Copyright Kemsley Design Ltd.,

27 Simulation in the BPM Lifecycle Source: Lanner 27

28 More On Analytics And Simulation Modeling and Analytics for Process Excellence Thursday, 10:10am, Diplomat 5 Denis Gagné Workflow Management Coalition (replacing Robert Shapiro) Copyright Kemsley Design Ltd.,

29 Technology: Predictive Analytics/Process Intelligence Smarter Processes for Smarter Outcomes Copyright Kemsley Design Ltd.,

30 Why Predictive Processes? Predictive analytics is not just about forecasting what s coming down the pike. It s also about keeping the bad alternative futures from happening. James Kobielus, Forrester Copyright Kemsley Design Ltd.,

31 Process + Analytics + Decisions = Intelligent Processes Business Process Business Rules Business Intelligence Copyright Kemsley Design Ltd.,

32 Process Analytics in a BPMS Executing process Realtime process dashboard Copyright Kemsley Design Ltd.,

33 What You Can Do With Process Analytics Information to support manual decisions E.g., display queue sizes to help manager to reallocate work Data to trigger automated actions E.g., spawn fraud detection process when series of events occur for same customer Predict missed SLAs E.g., compare history of activity timeline to estimate overall time to completion Copyright Kemsley Design Ltd.,

34 Focus On The Goal, Not The Task Compare: Current to baseline model Current to historical Analyze: Process dependencies and critical path Simulate to identify future problems Act: Self-adjust through feedback to decisioning In-process user guidance Copyright Kemsley Design Ltd.,

35 Slides at Sandy Kemsley Kemsley Design Ltd. blog: Copyright Kemsley Design Ltd.,