December 10, 2024

Office Optima Pro

Learn to Be the Best

Business Process Management In A Post-GPT World

Business Process Management In A Post-GPT World

Ian Gotts is the founder and CEO at Elements.cloud.

I mapped my first business process diagram over 35 years ago when I joined Accenture. I used a hierarchical diagram format called IDEF0 to describe an ERP system we were implementing at a U.K. defense contractor. We used it to validate the customizations the client wanted us to make. The principles of IDEF0 inspired the creation of UPN (universal process notation).

Fast forward to today. The validation of requirements is still the most valuable use case for process mapping. It gets every stakeholder on the same page so that the business needs are truly understood. The spin-off benefits are that the process maps can be used as training and to support regulatory sign-off.

Many organizations use UPN process maps, but this is mostly concentrated in the highly regulated industries—food, pharma, financial services, oil and gas—where there are financial penalties if processes are not documented and followed.

Challenges In Process Mapping Adoption

There is no widespread adoption or process mapping despite the obvious benefits. The process community has not helped itself become more relevant to the business users. In fact, the evolution of process documentation has turned off businesses and made the process community even more insular. Here are some examples:

• Lengthy business process reengineering projects deliver a pile of incomprehensible flow charts that only the authors understand.

• Complex diagramming notations like UML and BPMN have 300-plus shapes so that BPM workflow engines can understand the diagrams and build applications.

• Six Sigma training requirements and different levels (belts) make it seem unattainable for most people and require huge levels of investment.

The UPN standard was deemed by parts of the process community as too simplistic. It didn’t appeal to the process specialists. However, on closer inspection, there is more to UPN than boxes and lines. But more importantly, it is designed to engage stakeholders because it is easily understood by people at every level and in any department.

This is particularly important now that many enterprise applications are configured using metadata (for example, Salesforce, ServiceNow and Workday). They also have easily configured integration capabilities or can work with third parties like Boomi or Zapier. This means that the implementation is very business-oriented “low-code” rather than IT-centric “pro-code” custom development.

In the future, business analysts will utilize process mapping techniques, such as UPN, to actively engage with business users. These process maps will help validate requirements and guide the creation of user stories. These user stories, in turn, will outline the necessary configuration changes, which will be implemented through metadata rather than custom code.

The Role Of AI/GPT In Process Mapping

It is clear that AI/GPT will be able to write code and configure applications in the next five-plus years. But what it can’t do (yet) is understand what business users need. That is the role of the business analyst. Business analysts can benefit from AI/GPT by getting massive productivity improvements to support their work. We are already seeing this with consultants and our customers.

AI/GPT can take text, which could be the transcript of a discussion with a business user, or an image as the prompt to draw a UPN diagram. AI/GPT combines the prompt with best process practices from its AI model. It saves time but gets people started when they are staring at a blank screen.

Business analyst skills are key: asking the questions to prompt AI/GPT, then validating the output and refining it. In our experience, when AI/GPT draws a diagram that seems wrong, it is normally because the prompt didn’t have enough context or was ambiguous.

Off the back of a process diagram, AI/GPT can write user stories with acceptance criteria. This not only reduces the effort taken but also increases the consistency and accuracy of the user stories. These can then be reviewed and refined based on the business analysis.

Based on our experience, if we find issues with the user stories, it’s likely due to inaccuracies or ambiguities in the process maps. It’s more effective to correct the process maps first and then use AI to rewrite the user stories. Previously, before AI/GPT, we would update the user stories but often neglect to revisit the process maps, which could become outdated.

Now, with AI/GPT, we’re forced to improve our core foundational documentation since everything is generated from it. This improvement in documentation is a key factor behind the significant productivity gains we’re seeing.

But there are further gains. Enterprise apps are now configured via metadata rather than custom code. AI/GPT can look at each acceptance criterion in each user story and then evaluate if existing metadata can support it. If no metadata will work, it suggests what needs to be created. It saves time and reduces technical debt. An analyst can’t look at thousands of metadata configurations and evaluate every one for each acceptance criterion. AI/GPT needs as much information about metadata as possible.

The Future Of BPM

The future of BPM is strong. However, it needs to evolve and focus on creating foundational documentation (process maps, ERD, metadata descriptions) that AI/GPT can use to accelerate future phases of application development. Business analysis skills will be in demand for creating and validating content.

With AI/GPT now capable of reading documentation written in English, it can suggest configuration solutions and, in the future, build and configure applications. This brings us closer to the vision that has driven BPM for the past 20 years: building apps from maps. What makes this possible is that computers are now able to interpret maps in a way that is accessible to everyone, rather than relying on pseudo-code diagrams that only the author and a computer could parse.

We’ve finally crossed the silicon-carbon divide.


Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?


link

Leave a Reply

Your email address will not be published. Required fields are marked *

Copyright © All rights reserved. | Newsphere by AF themes.