As Digitate’s Chief Customer Officer, traveling the world to talk with our customers, I see that often they struggle to realize the value of an ignio implementation. After many meetings and discussions, I concluded that the major cause of ignio falling short on delivering the value it promised is a lack of focus on input.
When I analyzed the less successful customers, I found they were singly focused on the benefit of becoming an Autonomous Enterprise. This means they concentrated mainly on analyzing the output (for example, service tickets) and not what activities were involved in managing these tickets. For them, success came from stepping back from this kind of single-minded focus (even though it’s what conventional wisdom often advises) and thinking critically about what functions were yielding the target outputs and how to document them.
Our most successful customers all had a common approach: They were very careful to identify the right activities to automate (i.e., the inputs for ignio). Then they focused on translating IT operations’ Standard Operating Procedures (SOPs) into clear documents “understandable by machines.”
This means that to have a successful ignio transformation it is fundamental to focus on the quality of your SOPs. Despite all the current talk about apparently human-level AI tools such as ChatGPT, Intelligent Automation (IA) solutions are not actually that intelligent when understanding how to perform a task. Detailed specifications need to be provided, as I mentioned in my November 2022 blog post, “How to make an elephant fly.” The linchpin of any Autonomous Enterprise transformation is to understand your elephant, i.e., have a clear and well-defined set of SOPs.
Focus on the dirty secret: There is no glory in managing SOPs
A clear and well-defined set of SOPs is the key to any successful transformation – as I learned as a rookie in IBM’s reengineering department, where I wrote quite a few of them. Over time I realized that focusing on process input is fundamental to achieving a desired output.
This is particularly true for Intelligent Automation transformation. Machines achieve a specific output using a totally different logic and process than humans. After all, machines work in a binary way; we do not. For example, machines don’t use decimals to calculate, but they are very good at math!
ignio, like most IA solutions, has a set of Machine Learning algorithms that suggest which action will prevent, eliminate, or resolve a problem and offer a set of predefined actions to take. Each of these actions needs to match specific SOPs. Machine Learning will tell what action to take and when to take it but not how to do it. The “how” needs to be coded to reflect the SOP’s requirements.
For example, an IA Machine Learning algorithm can forecast when a server will run out of CPU resources and crash. To act, the IA solution needs the right instructions on how to reboot the server. The sequence to reboot the server is the How. Bottom line: The success of enterprise solutions is detailed knowledge of the SOPs necessary to run operations.
This is easier said than done. SOPs are the “dirty secret” of any IT operations: SOPs are usually outdated, incomplete, and poorly written. There is no glory in efficiently and effectively managing IT SOPs. But today, it is the only path to success in any IA transformation. Our Digitate studies prove that more than 60% of the effort to deliver a use case is wasted on managing inefficient and effective SOPs.
Three steps to expedite transformation
My experience suggests that these three steps will help to address the SOP conundrum:
- Have an effective and efficient template to “translate” SOPs from human language to machine language.
- Create a good SOP inventory. Not all SOPs will be translated in use cases; a few (approx. 30%) will be dropped. The major reasons SOPs are no longer relevant are:
- Currency (IT production environment is changing, which makes these SOPs lose their validity)
- Security concerns
- Not enough ROI on the use cases (cost of implementation is higher than benefit achieved)
- Establish a quality control review for each SOP.
Of these steps, the first is the fundamental one: It drives both how to create an SOP inventory and how to exercise proper quality control.
Back to basics: Good requirements drive good solutions
In the last few years, I have been using an Automation Requirement Document (ARD) template that I find very efficient and effective. This template, created by Narayanan Thekkepat (a.k.a. Unni), head of Intelligent Automation at our parent company, TCS, is a well-structured document. It has two main sections:
- Thesis definition: What needs to be automated
- Body: All necessary details to translate an SOP into “ignio” – What, How, Who, and Where. (“When” is not part of an ARD. This type of document is not a project plan, and therefore does not contain dates.)
- The ARD starts with the problem statement. It is good practice to define what each use case needs to solve (what is done to do the job) and what value it creates. Such an approach is of great help to later define the ROI for creating an Autonomous Enterprise.
Body: How, Who, and Where
- The ARD body contains all the important details to code an SOP into ignio. These details include input and output, prerequisites (for example, system access), architecture (description of the IT environment where use cases need to operate), nonfunctional requirements (such as expecting the use case to complete in 15 seconds), and a description of the team involved in the specific SOP
Bottom line: You’ll find that with an effective ARD, your journey will be shorter and involve fewer miscommunications and frustrations. In fact, our analysis shows that 60% of the effort to deliver IA use cases is due to defining and clarifying requirements and resolving prerequisites. Efficient and effective management of ARDs reduces this effort by at least half.
The template I’ve shared provides the framework for a good ARD. It drives users to provide the necessary details to fully understand how an SOP works, and creates the conditions required for any IA process. I have been using it for a few years and it has been yielding excellent results. I strongly suggest you use it when embarking on an ignio journey.