reborn55 - Fotolia

Fix bottlenecks before tackling business process automation

Business processes evolve over time, and intelligent automation requires a thorough assessment of what is not working optimally

If there is one mantra of business operations, it is that nothing stands still. This means that business processes, which may have been valid on day one when they were first documented, evolve as soon as people start following them.

Getting started on any intelligent business process initiatives requires an understanding of how business processes are currently run.

Nelson Petracek, CTO at Tibco Software, says: “People have a tendency to find shortcuts if given the opportunity, or they may find creative ways of circumventing steps or tasks that are seen as unnecessary or as bottlenecks. Fraud and other undesirable activities may also influence process execution within or across organisations.”

Petracek recommends that, when looking at how to automate business processes,  the starting point of any optimisation strategy is to understand such behaviours. Without this view, Petracek says it is very difficult to not only identify the right techniques for optimisation, but also to measure and benchmark any improvements against a baseline.

Process mining, in which system and applications logs are analysed, enables organisations to discover real processes and resource utilisation. By using associated analytical models against this data, says Petracek, organisations can start to understand the variants, patterns and bottlenecks in their business processes. This analysis informs the decisions they then make on which business process optimisations should be prioritised. 

“Given the number of data sources that may need to be included in the analyses, it can also be useful to employ integration or data virtualisation mechanisms as a way to access the data,” he says. “This eliminates the need for a large number of point-to-point connections, and creates an abstraction layer between the process mining capability and underlying source systems.”

Why automating processes fails

Akshay Anand, an IT service management product ambassador and evangelist at Axelos, says it is not good enough simply to automate a business process. Businesses need to identify the root cause of inefficiencies that prevent the process from running optimally. 

As an example of an automation that may not fix underlying problems, Anand describes a line of reasoning that people may take when attempting to optimise a typical payment process. “A simple description of accounts payable work might be to receive the invoice, match the invoice to a purchase order, check that goods or services were received, check payment terms, get the invoice payment approved by a finance controller, schedule payment, update the accounting system, and finally to complete any other paperwork,” he says.

“In this example, getting payment approval takes two weeks, but the rest of the steps take less than a day. The whole process needs two weeks to complete, which means that investing in automation to reduce time and effort outside of the payment approval step is highly unlikely to have any significant impact on the total time needed.”

Another technique for understanding how business processes work is borrowed from car maker Toyota – value stream mapping (VSM). Akshay Anand, an IT service management product ambassador and evangelist at Axelos, says that in a manufacturing context, the VSM is used to depict the flow of materials and information from one part of the factory to the next.

In recent years, VSM has gained traction in the IT community, he says. IT frameworks and approaches, from ITIL 4 and IT4IT to DevOps, talk about the application of VSM. Anand says it is commonly used to identify the flow of work and value such as support tickets, requests, code under development, and other knowledge artefacts. These artefacts are equivalent to what Toyota describes  as “materials”, he adds.

“The power of the VSM lies in its ability to track work from when it is identified to the point when it is completed,” says Anand. “By running collaborative workshops, stakeholders can gain a better appreciation of the work that goes on in other teams or departments.”

This, he says, enables the organisation to identify how much time a piece of work such as a support ticket, a request or some other task waits before someone starts working on it ,versus how much time it takes for someone to complete it. Anand points out that such analysis reminds people that the rate at which work can be completed is the rate at which the slowest link can work.

“While automation can bring significant gains in efficiency, quality and speed, it is unlikely to do so if the bottlenecks and constraints haven’t been addressed,” he says.

Removing bottlenecks and modernising

Family granite-mining business Cosentino is an example of a business that is going through the process of assessing and working around the bottlenecks that prevent it running more efficiently.

The company has been on a digital transformation drive over the past three years, and mapping out business processes and optimising them is a core part of the digital transformation initiative to support the company’s expansion.

In just 40 years, and while remaining family-owned, the company has transformed from a local marble quarry with 17 employees to a business operating in 32 countries. Evolving to a truly global business has required improvements to the way the business runs and its people work. This has resulted in IT working on business process optimisation projects, says Rui Novais, director of IT at Cosentino.

“There are lots of processes that business owners ask for to improve ways of working,” he says.

Novais says the company is often looking at developing more structured approaches to existing business processes. IT is also finding opportunities to use tech to improve processes by identifying pain points and ways to work around them.

For example, he says the company optimised its employee onboarding process, which was previously unstructured. It involved several people in the process and used to take up to seven days for a new hire to be fully onboarded.

Describing the approach the company took to optimise the process, Novais says: “We started with pen and paper to define the process, then modelled it using Tibco, to identify gaps in how it was working and to describe what we wanted to achieve.”

The overall objective of the employee onboarding process was to ensure new employees get all the applications they need for their job at Cosentino.

Putting in place new and improved business processes is most successful if someone from the business can champion the change. Novais adds: “It is not easy to show someone they are not efficient.”

Cosentino identified key users who could help others to understand how the business process improves the way they work. Novais says dashboards are used to help the company assess business processes to understand bottlenecks. “We can review processes on a regular basis,” he adds.

The company has a cloud strategy based on Microsoft Azure and the Tibco cloud and is actively building applications that extend its legacy SAP enterprise resource planning (ERP) system. For instance, Novais says Cosentino is extracting data from the ERP for a new purchase-to-pay business process that is being run outside the ERP. This then feeds back into SAP to keep the ERP up to date.

Adding intelligence

While modernising applications is a key stage in business process optimisation to reduce manual processes, the next stage is to add intelligence. This is what Citizens Bank has started doing, says Robert J Bush, senior vice-president for home mortgages at the bank.

“We partnered with Infosys to draw on its applied AI [artificial intelligence] capabilities and intelligently automate our mortgage information extraction and audit process,” he says. “Having significantly reduced manual efforts and rework, we are able to rapidly onboard new loan portfolios and enhance customer experience.”

Read more about intelligent business process automation

  • AI is used to find hidden meaning in large datasets. What if an AI system could truly understand an end-to-end business process. What if an AI system could optimise a business process autonomously?
  • Business process automation might find new territory to expand into as businesses have become bloated with excess processes since the 2008 crash. But has IT drifted out of business alignment again?

French domestic insurance firm Covéa has fully digitised its procurement process using an Ivalua procurement tool. Although the company has been an Ivalua customer since 2012, Sylvie Noël, chief procurement officer at Covéa, says the firm recently rolled out electronic signatures, which removes another manual process.

The company has also begun trialling projects in cognitive sourcing, in which chatbots and smart AI-driven contract summaries automatically extract the relevant data and terms from contracts and present them in a summary sheet for review or sign-off.

AI can indeed be deployed as part of a business process optimisation strategy. But as Axelos’ Anand points out, intelligent automation does not necessarily fix underlying problems.

In a paper published in 2019, McKinsey senior partners Alex Edlich and Vik Sohoni discussed the merits of addressing a pain point in a business process, rather than using automation as a quick fix. “Taking an end-to-end view of the outcome needed and measuring that delivered output is better than applying a robotic Band-Aid to a particular pain point,” they wrote.

“That’s not to say that there aren’t some pain points that should be immediately addressed, but that, at scale, deploying thousands of bots isn’t always the best answer. Better to figure out what the desired goal is and then figure out how bots can (or cannot) help. This often means collaboration and coordination with other silos, or creating a corporate business-process management group.”

Read more on Artificial intelligence, automation and robotics

Data Center
Data Management