How to Optimize IT Project Management with Intelligent Automation

We have all been there. The project is in trouble. The system started to fall behind during system integration. The developers found that some information needed by downstream systems was not collected in the new system. In User Acceptance Testing, things really got bad. Clients kept finding the functionality they needed, but that was never captured in the requirements. Management is getting ready to throw in the towel.

The program manager (PM) is trying to understand what went wrong. The PM had been working with a team of trained and certified project managers with their Green and Black belts to capture the requirements for the new system. They worked with a group of in-house Subject Matter Experts (SMEs) to capture everything that is known about the old system. Using industry-standard techniques, such as Lean Six Sigma and Kaizen workshops, they have walked through all the functions of the old system.

These techniques allowed the team to extract knowledge from the people who know the system best. During days of facilitated sessions, the team walked through every part of the system. They documented what the experts said the system did now and their requirements for the future.

And not just the people. The program manager created a master plan to review all of the documentation about the system. The team went through page by page to extract the functionality described in the manuals.

The work was intensive. The team needed to clear dozens of schedules for days of meetings. They carefully documented everything they heard during the hours of meetings. They came up with a comprehensive set of requirements to make it easy for the developers to create a beautiful new system that does everything the old one did while meeting all of the new requirements.

Or at least they thought they did. The test results show instance after instance where it is now clear that they did not understand the old system. And even when they did understand it, the system kept changing while they went through months of planning. When the designer does not understand the system, it is easy to make fundamental mistakes in the design. These mistakes cannot be corrected by a minor change in one module. They require changes that thread through the entire system design. These are the types of changes that can add weeks or even months to the development effort.

Understanding old systems and what they do is one of the most difficult tasks in system development. Reviewing the system documentation is dangerous. In many cases, the documentation describes what the system was originally designed to do. But that is not what it is doing after years of enhancements and system updates.

And the SMEs; they are like the story of the blind men describing an elephant. They each know a part of the system and are adamant that their understanding is correct. But it turns out that the people that really understood the system retired years ago. So, the hours of facilitated sessions resulted in volumes of requirements that only partially map to reality. The diligent effort to integrate the divergent descriptions and wordsmiths of each requirement created a beautiful document with one major problem. The system they describe does not do what the organization needs it to do.

Despite getting the right resources and using all of the right techniques, the system wound up way over budget and months behind schedule.

Is it any wonder that 33% of IT projects fail?

But now there is a better way. With the modern tools that are now available, the program manager can really understand how the process works. Not a best guess or an approximation, but smart tools that dive into levels zero and one of the process models and allow the team to look at the alternate paths they can choose and recognize the impact of each decision.

Instead of focusing on what the people who use the system know, at Roboyo, we leverage emerging technology to improve the design and development phases of the automation process. We work with SMEs to take the guesswork out of process analysis in intelligent automation from UiPath, Appian, and/or Celonis. These services allow us to visualize the system with you in a way we never could before. When we can visualize the system and see the impact of changes, it becomes practical to focus on continuous process and service improvement. We reduce our risk of implanting changes and increase the probability of exceeding our project goals because we are finally bringing data-driven decisions into the decision-making process.

Until now, process discovery was really like medicine in the 1950s. It produced volumes of data in the hope of capturing the essence of the system, but much of that information was useless or even misleading. In the last century, if a patient came in with a broken leg because they fell off a ladder, most doctors would understand the need to set the broken bone. But only the best surgeons would understand if that break needed pins to support the bone while it healed or if there was the potential for complications like blood clots. In the same way, the best system designers could start to understand the system based on their years of experience with similar systems. However, for most designers and project teams, the volume of information was simply overwhelming.

Today, modern doctors have advanced diagnostic tools such as X-rays and CAT scans that allow the diagnostician to see the problem in great detail. Process discovery options today give process owners, process improvement specialists, and IT software versions of X-rays and CAT scans for noninvasive looks at processes. With Artificial Intelligence (AI), even a new intern can identify hidden problems.

In the same way, Roboyo uses modern process mining tools like UiPath or Celonis to mine the existing systems data logs and extract the precise information that process improvement specialists, infrastructure architects, process owners, and agency decision-makers need to understand the system processes. And the tools use AI to separate the dead-end paths that are no longer used from the work the system is actually doing.

The traditional techniques provide a “best guess” at what the system looked like at the time it was studied. However, because they are so labor-intensive and cumbersome, it is not possible to keep them current. The tools Roboyo uses can provide a real-time understanding of the current system as it changes.

With these tools, it is possible to completely overhaul Continuous Process and Service Improvement efforts. When you can see the impact of a change on the process, you can start to make more precise changes to an existing system and focus on the new requirements rather than replicating the old ones.

Roboyo teams digital labor with the PM, creating a project manager multiplier. Using AI and Machine Learning (ML) to dig deeper than previously possible. These tools collect data relatively quickly, automatically updating the reports and databases the PM needs for tracking the project. This liberates the PM to spend more time with stakeholders and provide them with more accurate and timely information. The project’s success or failure is not solely on the PM’s ability to perform repetitive and mundane administrative tasks.

We leverage a variety of tools. The market is still separating the good from the better and the best. Some vendors are playing catchup, adding process discovery to their offerings. Others have proven business platforms that seamlessly integrate process discovery and design to enhance the build, test, and deploy phases. Some of the offerings include:

a) Automation Hub (UiPath) or Converge (Roboyo)

These are the nerve center of the IT process improvement function. Like the nurse’s station in a hospital, they provide a central location for all of the information about the status of an automation.

These web-based platforms democratize an agency’s intelligent automation program by empowering every employee to submit automation ideas and report all the symptoms of the ailing process. They automate the automation program, reducing the workload in the Center of Excellence (COE). They allow the automation program to be smarter, cheaper, and faster. PMs can mine the Automation Hub or Converge for process improvement opportunities. The voting and comment features of the Automation Hub allow process improvement specialists to easily identify the community of employees that agree with the need to fix the process.

Both Automation Hub and Converge manage the lifecycle of the automation tasks so PMs can avoid unnecessary manual data entry.

b) Process Mining

The process mining function performs a detailed analysis of the current process. This function quietly monitors and collects process data for days, weeks and months using logs from both Government and Common-Off-the-Shelf (GOTS and COTS) systems. It uses transactional data in systems and applications to reveal the true processes, giving agencies the ability to see opportunities for automation and process improvements in their end-to-end business process.

The output of process mining provides a visualization of the process with a drill-down capability that makes it easier to understand exceptions to the norm and why they occur. This allows data-driven decisions to be made on the best way to proceed. The PM benefits in the requirements generation phase because the visualization of the process comes complete with execution times and variants.

The PM and the Process Improvement Manager (PIM) can use the same process mining logs weeks and months after deployment to determine the performance of the modified system. It is now easy to show if the original ROI was met. Incremental process improvements can be measured on an ongoing basis so that any negative impacts can be quickly identified and remedied.

Agencies can stop reacting and go on the offense with respect to continuous process improvement.

c) Task Mining

This is where the process gets its detailed scan. Where process mining takes a macro view of the server logs, task mining provides a micro view at the desktop level. It can be deployed on a single desktop or on multiple ones to get a broad view of a single process performed by multiple people. Where Process Mining is an X-ray, Task Mining is a CAT scan for process diagnostics.

Because the team can visually compare actual employee variants and understand the process times for each subprocess, they can now see both the normal case situation as well as the exceptions that require an inordinate amount of time. In the past, the PM needed to find the perceived SME for a process and then capture what the SME thought was happening. Now the workflow is captured in real-time. Using this information, the AI component creates sample code to rectify inefficiencies. This code guides the developer, showing them what needs to be done, saving time, and ensuring accuracy.

d) Task Capture

UiPath went one step beyond the X-ray and CAT scans. Their platform offers an MRI. Task Capture ensures we keep humans in the loop and leverage their expertise by allowing them to document the “as-is” process. They can annotate exceptions that might not have been present in the discovery phase and incorporate this information into the Automation Hub or Converge to consolidate the manual and automated results.

Again, the PM is spared the effort of manually entering data into the work package or the Process Design Document (PDD). No more yellow sticky notes on a board. No more manually writing it down or taking a picture and then converting the sticky notes to a draft requirement as both text and a flow diagram.

The Process and Task Mining functions team with Task Capture and the Automation Hub to standardize automation data and provide a repository for each candidate and deployed automation.

e) Deployed Robots

PMs and PIMs both benefit from deployed robots. These tools can monitor the system and respond to changes automatically. Standards are enforced instantly, and exceptions can be captured for future improvements. These tools monitor process mining and task mining in real-time so that the impact of a change can be measured and can be compared within hours of implementation.

f) Insights (UiPath)

The program Insights function serves both the PM and PIMs well by finding trends and system impacts. It not only tracks the performance of your robots; it calculates Return on Investment (ROI.) This means you can see the current ROI calculation (updated by automatic recording and reporting) compared to the ROI agreed upon early in the design lifecycle. PMs no longer have to spend days collecting data and hours analyzing it to find trends. They now have time to evaluate why changes did or did not happen. Both the PM and continuous PIM are able to determine if the changes implemented are having the positive impact anticipated when the project starts.

We cannot blame COVID or the Great Resignation for project failure. Nor do we need to. We don’t even need to toss out PMP standards, nor should we. We must realize that work is continuous, and process improvement should be continuous as well. Using our Roboyo methodology, we show you the inner workings of not just one but all the processes you want to improve.

No agency can afford enough staff to run new projects, spending lots of time and money while continuing to see failure after failure. Nor should they. There is no reason today that every process should not go through what the former CFO at the U.S. General Services Administration called the EOA test. He said every process should be evaluated to determine if the process should be Eliminated, Optimized, and or Automated.

The process can be truly understood using process mining, task mining, and task capture products, so the debate is on what to do next, not what is happening today. It is time to stop accepting a 33% failure rate and to take your projects to the #NextLevelNow.

Roboyo thinks our Human+ approach to overhauling project management is more than just an incremental improvement. We suggest that the process loops associated with Continuous Process or Continuous Service Improvement will need to be renamed. These loops will go from continuous process loops to Continuous Confidence Loops. Confidence because we are always able to track a process and evaluate if it is improving. And Confidence because we can match it against the citizen/customer experience to see if the desired outcome was achieved. In Human+ project management, the story can have a happy ending.

No modern doctor would try to diagnose an internal injury without taking advantage of the incredible detail provided by modern medical imaging. These tools prevent misdiagnoses and allow the doctor to quickly and efficiently treat the patient. In the same way, Roboyo’s Human+ approach gives agency leadership the information they need to make their organization more efficient and effective. Wasteful motions are eliminated, and taxpayer money does not continue to be wasted on projects that are doomed to fail because we do not have all information. Like all good stories, this one has a happy ending – better government through intelligent automation.

At Roboyo, we can help you leverage Intelligent Automation to improve your project management and workforce productivity – and take your enterprise to the next level. Please contact us for more info on how Intelligent Automation can help you automate your government.


By:
Jim Walker
SVP Public Sector, Roboyo

Jim works across all levels of government to help show the tremendous value to agencies as they move to the fully automated enterprise.