If we look at the CEO mandate in most annual reports, there is a goal to reduce costs. Clearly, the first thing that comes to mind is to automate what is being done within the company.
With the new normal due to the pandemic, this imperative has only been further emphasized. While the need to automate is nothing new, what has changed over the years is what types of tasks can be automated and how much can be automated.
Logically, it makes sense that many activities that are being performed by a person can be accomplished by a machine more cheaply and with more accuracy.
Machines historically have been good at reading what is structured or what is called machine readable form where there are defined columns and data appears exactly in a pre-defined area. The machine picks up the data from the correct place every time and mimics what the human being does. Many companies have already taken care of these automation requirements or at least it is on their radar. Having said that, we still discover in some cases that there are some of these processes running in some parts of the company where no one has bothered to check. This has nothing to do with the technology team’s lack of will or knowledge but a larger question having to do with who identifies use cases for automation.
While many companies talk about an automation mandate, the question is whose mandate is it to identify those processes?
While many companies talk about an automation mandate, the question is whose mandate is it to identify these processes? Unless someone calls it out as a problem, these processes continue to be run manually.
Structured vs. unstructured data
There is a whole lot of work done today where data from one company needs to be captured in another company’s system (i.e., invoice data, claims etc.) These are data in different formats in different companies and the solution required is to understand the formats and extract the relevant information from these templates. While the templates are not structured, companies are able to solve these problems based on what is required in the document.
These are problems that automation can solve.
In some cases. automation may take longer to decipher what a person can do quickly intuitively. For instance, a human can look at a document and understand it is a prescription, invoice or a phone bill while the machine needs to scan through the content and then figure it out. Both as technology evolves as well as keeping in mind criticality, it still makes sense in some cases to automate. Over a period of time, machines can also learn to do these tasks quickly.
For instance, consider the example of address verification. A company receives documents and it needs to verify if the address is identical or changed. It’s a simple task for a person to compare but not necessarily that easy for a machine when, for example, the address line 1 has moved to address line 2, or where there is additional information such as 2nd floor and that relevant additional information needs to be captured. This requires automation to understand the context and only update all relevant additional information. While it was tough to automate this a few years back, now technology has evolved to solve these problems.
Another example involves unstructured data like prescriptions handwritten by a doctor in which documents include some words like “urgent” explicitly written. These terms can be easily identified by a person as urgent but a machine will need to be trained to understand that. While people talk about solutions that are now 80-90% accurate, the question is whether that solves the problem.
The cost of an error can be much more than the benefit that can be accrued. You need an expert to be able to help you make the automation decision or how to gain the maximum benefit from automation by including elements of human judgement. In spite of this, a couple of years back in one of our automation engagements, there could have been a billion dollar bot error but we had designed some checks and balances which ensured that there was no business impact.
Role of change management in automation
There could be a few hundred processes running in each department in a company and there will also be quite a few which have been running the same way for decades. In some cases, people would have considered automation as an opportunity a few years back but perhaps would have given up or were unsuccessful in their attempts. However, as technology evolves, there is merit in revisiting the feasibility of automation. There are a number of instances where what we could not automate last year because of technology constraints has been automated this year and running successfully.
As technology evolves, there is merit in revisiting the feasibility of automation. There are a number of instances where what we could not automate last year because of technology constraints has been automated this year and running successfully.
Consider the example of a finance or a HR professional who has been doing the accounting activity or hiring data capture for a number of years. Most would like to believe the work we do is complex. Many times when interviewing people processing any work they end up talking about the most complex scenario . While they need to be proficient to handle those exceptions, the fact is that in 80%-90% of scenarios, humans don’t need to handle those tasks.
However, if the HR or finance professional doesn’t communicate they need support, the IT team or whoever is responsible for automation will not know there is an opportunity for automation in that area. Even if they do reach out for support, how do we know if there is a good assessment done and is comprehensive?
As part of our HGS transformation team, there are consultants whose job it is to identify opportunities to automate. In one instance, an external consultant provided a report assessing the automation benefit at 10% while another reported it at 50%. The whole business case and priorities change with these two different recommendations and consider that these insights are from professionals who have some experience in this area. The difference in this case was that some amount of reengineering makes the process amenable to significant automation. That’s why it is important to look for solutions from various angles.
Organizations are challenged with employees who are working on their daily tasks to step back and be proficient to identify automation opportunities. There will be many who will want their repetitive tasks to be automated and some who may be worried about what will happen to their jobs.
Automation Center of Excellence vs business-led
The question of who owns automation in the enterprise is a critical one for organizations to address, There are multiple stakeholders in the company: IT, business and support functions. Each organization is different and in some, the business takes a lead in all decisions while in others, IT is tasked with championing all automation. Here are some benefits depending on who drives the automation agenda:
IT team championing automation:
- Awareness of “big picture” on company-wide technology strategy
- Knowledge of applications, constraints, system utilization and enterprise impact
- Policies around application life cycle management, data integration
- Partner network
Business team championing automation:
- Identification of use cases
- User-level understanding of challenges and workarounds
- Ownership of achieving savings or improvement in quality
While the right structure depends on how the organization is set up, most often the solution is a hybrid approach. For instance, IT can own the automation strategy, define the enterprise road map and application integration while the business function can play a critical role in use case identification, owning the project implementation and ultimately the return on investment. They also need to take care of communication, human resource management, training and a back-up plan. In some successful organizations, the IT and business functions jointly own ROI. With this model, there is a clearly defined measure of success and the real benefits are calculated on actual reduction of human resources, processing more volumes, impact of reduction in cost of error, etc.
While in theory it is nice to talk about collaboration between IT and business to drive automation, in reality to make it work it is important to understand the business dynamics of the organization.. Incentives can play a critical role in the success of automation implementations.
Many companies that experience success have identified automation brand ambassadors from their functions who worked with the IT to educate teams on the art of the possible and potential impact of automation.
HR has a key role in incentivizing people to be open to automation. Unless this happens, there could be impediments in knowledge sharing, taking ownership in implementation and measuring success.
Additionally, it is important that organizations revisit the scope for automation as technology evolves.What may have been constraints to automation yesterday may no longer be the case. Similarly, the business case may also be a function of the scale of operations, availability of skilled resources, cost of errors and other factors, all of which may vary over a period of time.
Many companies that experience success have identified automation brand ambassadors from their functions who worked with the IT team and ran roadshows on educating the teams on the art of the possible and potential impact of automation. This can be given a more formal structure and be recurring to prioritize and revisit opportunities.
Since there are multiple business functions in an organization, some successful implementations allow the IT team to take the lead on deciding the strategy, providing updates on emerging technologies, and selecting the platform and partner. Some organizations have given a % target to the business and the central IT team partners with the business to figure out how that works.
The key to success is how to secure buy-in from all stakeholders in the organization. As machines keep learning, it is important to use the technology to make it easy for a person to do their jobs. After all, who wouldn’t want their very own machine personal assistant?
While not everything can be machine solvable, it is important to have a culture where people embrace the change.
Some organizations decide to partner with process and automation consultants to help them on their entire change journey from defining the roadmap to identifying processes for automation to creating the business case and delivering success.
Regardless of what approach your enterprise takes, make plans now to invest in automation. It will derive critical benefits that will allow your organization to thrive and grow.
by Ram Mohan, Senior Vice President, Business Transformation , HGS