Automation Approach for SAP Estate

      In my earlier articles, we explored the need for DevOps in an 21st Century Enterprise IT Estate and DevOps Value Chain for SAP Estate. This article can also be read in isolation. We will discuss Automation a key link in the Run To Optimize (R2O) value chain of DevOps.

      Automation is the use of control systems for operating processes that minimize or reduce or eliminate human intervention. The quick win benefit of automation is labor reduction; however, it is also used to improve quality, improve accuracy. Although automation is not new, it is now being looked at in IT Estate like never before due to changing demand, development, and evolution patterns.

Like all automation control systems, the IT automation approach also looks at two automation control frameworks. Namely, Open Automation and Closed Loop (Feedback) Automation.

      Open Automation is where the action is independent of the process output. It is automation ‘execute’ an activity, and not dependent on the output. For example, we can do a system build using a tool or script or do a code quality check automatically.

      Closed Loop (Feedback) Automation has a feedback loop that ensures the action to give a process is dependent on the output. It uses a reference point and can reset itself. For this reason, closed-loop (feedback) automation is used for self-healing processes, Machine learning, and artificial intelligence. For example, we can have a system monitor a business process and take autonomous calls to update the underlying technical KPIs to meet a target business KPI goals without human intervention, impacting environmental behaviors.

     Automation is an Incremental evolution built on a robust process foundation and is an evolutionary process. In an Enterprise environment where standardization is desired, trying to apply futuristic automation without having the right levels of foundations setup will not sustainable. Thus, the right automation goal is to remove the repeatable/highly repetitive/high touch point tasks and free the Highly Skilled Human Resource to focus on Higher Value add work which we call as Up-Placement. Vision for 21st Century Enterprise is to operate in a Multi-Modal IT Delivery model with Focus on Shift Left on Operations and Evolve the Shift Right on Development to increase pivot-ability of new business models.

      The above figure of typecasting sieve for automation is a simple metric to identify a roadmap for automation. It measures the readiness or available automation solutions for a process, and the level of human intervention. This metric should not be used for leaner adaptation but to be used purely to classify processes within your IT Estate. Type A and B are the low-hanging benefits that should be adapted on an immediate basis as they have ready solutions, off-the-shelf solutions with configuration. Type C and Type D are the ones that will need significant due diligence and an adoption roadmap and customized solutions. Do note that this metrics is true as of ‘today’, meaning, with the availability of more automation options and solutions, the same process can move from D to B, or from C to A. This is the inevitable shift. 

Type	Describe	Example
A	Task that are best performed by Machines.	Recognizing patterns of events and executing the relevant action System monitoring, self-services, run book automation, event-triggered activities, infrastructure provisioning, database & backup management, integration
B	Task that are best performed by Machines augmented by Human.	Resolution of off-pattern or complex or unstructured events where the final decision of appropriate action lies with IT worker.) Transport release, batch automation, system administration, system copy, process monitoring, data volume management, archiving, Run book automation
C	Task which are best performed by Humans augmented by Machines.	Non-patterned higher complexity tasks which involve a human relationship, human-centric process, or self-help scenarios where the action is taken by a human-based on analytical input from a machine. Process Automation, custom code assessment, upgrade impact assessments, security, business rules, mobility, analytics, workflows, data governance, testing
D	Tasks which are best performed by Humans alone.	Tasks that require an understanding of business priorities, organizational inter-relationships and human-centric processes Process engineering, design thinking, personalization, business & social collaborations, landscape transformations, analytics modeling, business planning

TYPE A automation are the lowest hanging fruits and there are standard solutions that should be adopted for automation. There is the maximum value in this space for Operations and IT stakeholders as mature off-the-shelf offerings are available. This should be the easiest to adapt, and has immediate benefit to operations. A process study of the customer landscape and tickets can quickly identify the cases where we can eliminate repeaters and runners. The approach will be to adapt to self-service and autonomous machine-driven processes.

TYPE B automation is in areas where there is mature potential to adapt a solution that has proven mature products but key decisions are taken by a human. Process mining needs to be done to simplify or eliminate human dependencies. Here the automation goes hand in hand with the Best Source of help (BSoH) that either replace human intervention or shifts the work to low-skilled resource and relieve high-skilled resources to do more value-added work. The approach will be to build automation strategies for manual intervention.

TYPE C automation is in areas where you may not have off-the-shelf easy to deploy solutions, nor are there continuous human interventions. Process mining will show them in critical business processes, scheduled occurrences, with a high impact on value. Process Automation, Design Thinking for User Experience and Business Process Re-engineering is to be used. There is maximum value in this space for Business and Functional stakeholders. The approach here would be to build a predictive model and adopt robotic process automation. 

 TYPE D automation is in areas of complex process areas that require human interventions with advanced cognitive skills and knowledge. The maturity of machines in this area today is very low and it has high involvement in human intervention. Although there is a high investment in this area, leading to the concept of singularity where decisions made by machine will not be any disadvantageous than that of human or could be better.  This area is not the immediate goal for any enterprise looking for   The approach here would be to evaluate artificial intelligence and advanced machine learnings.

Let us do a use case of TYPE A and TYPE B automation through the approach of Best Source of Help. This is a stepping stone for successful automation in an environment where you have existing users that consume the IT services through help-desk or an ITSM tool or through an ad-hoc process like email or walk up the team. Most process mining in existing landscapes shows that there is always something to improve. This use case is not going to be the pill for all ills, but a guideline around how to address it. IT Estate big or small, simple or complex, centralized or distributed will all benefit from this.

Best Source of Help for Run To Optimize (R2O)

Best Source of Help (BSoH) is the methodology where a front end staff, say the person manning the helpdesk or the operations command center, can follow Standard Operating Process (SOP), execute Guided Procedure (GP), Self-service Portals, and scripts to resolve problems, queries, and requests. This will also give guidance to what requests can be closed upfront, and what needs to go to support teams. This will enhance the user experience around how IT resolves a situation. The BSoH is not just another Known Error Data Base, but a methodology that brings all the essence of ITIL under one umbrella that includes autonomous self-healing, to proactive response, faster time to resolve, integrated root cause, and creation of virtual collaborative team rooms.

The assessment for service management can starts with process-mining the existing tickets and process from Helpdesk to its resolution. A good process mining will optimize in the following areas:

1. Wrong ticket forwarding – SOP will likely result in a ticket assigned to the right team, the first time. BSoH will help facilitate correct and timely resolution by identifying rightful owners for assignment.
2. Ticket redirections (Multi hop) - tickets are redirected between many departments, leading to an unnecessarily long solution duration, low efficiency, and the loss of money and time. BSoH will facilitate collaborative team rooms and war rooms to get all dependent parties together for common resolution.
3. First Time close -  Through clear guidelines and strict monitoring of these tickets, these cases could be reduced. Use BSoH components of Guided Procedure (GP) and Scripts.
4. Increase Time To Resolve – Shift Left by putting the tickets in the right queue the first time using BSoH
5. Ticket Prevention – Implement Self Service workflows which will enable users to own primary resolutions.
6. Ticket Elimination – BSoH with RCA  will help enforce Root Cause and Corrective Actions (RCCA) through mining the data. So Runners and Repeaters will get addressed faster by quicker corrective actions

Automation is one of the leavers in DevOps for SAP. While we understand the need for automation, we see some deterrents in the realization of full value. Deterrents to DevOps and continuous delivery are Legacy IT systems, old technologies, complex system, heterogeneous integration, poor project-management processes, unclean data, manual processes, and uncoordinated actions by disconnected teams. Thus it is important we do treat our IT automation with the type casting sieve for automation for us to build a roadmap and do the required treatments.

In my next article, we will discuss more on DevOps adaptation for SAP Estate: Extending DevSecOps to your SAP Landscape and DevSecOps - Embedded Security With Omni-Speed DevOps

About the author: Trijoy Saikia is an Enterprise Architect having 16+ years of experience in IT across the chemical, natural resources vertical domain and working experience in mills and process industry. He has worked on a complex SAP landscape in various capacities in outsourcing, system integration, and transformation projects. He has experience working on large multi-geography landscape transformation related to acquisition and spin-off, including managing integration, transition, rebadging, and handling sensitivities. He has been in roles with accountability in Delivery, Financials, Customer Satisfaction for Service Management and Service Delivery across SAP and non-SAP applications. He is involved in roles across technology – design, integration, IT transformation, technology architecture, project management, mobilizations, and change management. He has contributed to pre-sales, solution architecting, service management, mobilization, transformation, and delivery roles. He is involved in building capabilities in technology architecture. Have co-authored white papers, leads innovations and has built capabilities in SAP Architecture, Administration, Security, Integration and Shop-Floor Technologies. He is a Computer Engineer by education and hails from Assam in India. He has worked with Accenture and Patni Computers Systems prior to joining HCL in 2014. He is is a Director in SAP Service line. He can be connected at trijoy.s@hcl.com