Business Process Automation in Infrastructure Projects

In a broader sense, automation may be defined as a change in a process producing some output with an objective to reduce or eliminate the need for human involvement.

Business Process Automation (BPA) considers business a collection of low-level processes. Their content, sequence (or workflow), and ways of interaction are the subject of automation.

Instead of human involvement, BPA focuses on the business speed and flexibility to scale up or down.

Examples of a low-level process (LLP) are process engineering, mechanical engineering, control systems design, procurement, quality assurance, project management, and others.

An example of BPA is a preliminary design generator of the water treatment system developed by Transcend.

BPA starts with identifying the bottlenecks. Across many industries, the biggest one is the data exchange between LLPs. This bottleneck is fundamentally different from the siloed business (or ride-sharing) problem, as it moves the vantage point outside any LLP. It leads us to the data interoperability problem discussed in “Information Interoperability ”.

Why the 20-year-old interoperability problem has not been solved yet? The general answer lies in the business evolution. It is visibly captured by the business standards and good practices – the foundation of the business culture. In turn, the culture defines the complexity and depth of the automation and the process output quality.

The water industry (torn off by the decentralization idea) and the oil and gas (consolidated and governed by the API standards) illustrate the full cultural swing. BPA matching the former will not be compatible with the latter.

Business evolution is missing in the traditional BPA software development cycle as it uses the process snapshot from the past and normally takes over five years to complete.

BPA is not about digital imitation or replication of human processes. The post-automation software-driven business process may substantially depart from its manual version; some LLPs may be dropped or merged. BPA silently adds a new objective – making what is left of human involvement more productive. BPA is transformational as it changes the perspective on the business itself.

BPA is driven by business rules - stable patterns observed in business. Conversely, its evolution may be reflected through patterns' identification, application, and creation. While identification is rooted in personal experience and tacit knowledge, the application follows some rationale and adjustments in the Information Classifying and Coding System (ICCS).

The pattern creation includes adding new LLPs - often from industries with higher business cultures. A notorious example is FEED. Only recently it became a standard in bidding for desalination mega-projects. The above-mentioned points make BPA fundamental, traceable, and irreversible.?

Business rules and maxims equally apply to LLPs and BPA itself. One of the BPA top-list rules is an unconditional implementation of the international standards and industry recommendations. (They are the primary source of rules for LLPs.) Today the technological debt is the biggest in the P&ID development, design for safety, control systems design, and project management.

Next, BPA shall target high-culture cases which will certainly become the norm by the time the BPA development is completed. The mentioned above Transcend is still in doubt about which cultural niche is the best fit for its product – municipalities or engineering companies.

Another example is Augury. Having positioned itself as a world leader in the AI-driven predictive analytics of rotating equipment, this company failed to strengthen its presence in infrastructure projects executed in desalination, oil and gas, and pharma. Its website reveals the reason – indifference to existing standards and practices.

Finally, BPA shall be implemented from the bottom up. This fundamental principle is grossly disregarded in industry. It leads to a flood of the BPA software products claimed to be the best for procurement, or maintenance, or project management. They fail to work in tandem because of an extremely localized perception of the problem they intend to solve.

Can AI be a mainstream way of business rule inference? Given the time and cost of the AI agents' development, it cannot compete with the domain experts operating on a substantially bigger pool of tacit knowledge already sorted out and classified.

Moreover, the said development depends on the expert ability to identify the problem, “wrap” it into the AI algorithms, and perform data engineering. Surprisingly, the latter is in the scope of BPA. It means that AI starts where BPA ends.

Reprinted from crenger.com