How Can You Determine Your First Bot Use Case and Assess Its Value?

How Can You Determine Your First Bot Use Case and Assess Its Value?

In the high-tech world that is software automation, much of the focus goes to the enabling technology, but true value doesn't lie in the solution but rather the problem. The world is filled with buzzword applications and solutions for every problem you could think of... but identifying the true pain point is where the value chain begins.

Problem identification

These days, companies have a myriad of problems to deal with that are slowing down their throughput and increasing their operating expenses. Some of these problems include:

  • Using antiquated, legacy software that isn't keeping up with business demands
  • Integrations between systems
  • Lack of open APIs to move data in and out of systems easily
  • Using people, not technology to operate these systems
  • Data integrity issues

These issues have led to linear increases in operating expenses for companies as they scale revenue. For example, let's look at a simple invoicing process problem that one of our customers was facing (case study here).

It starts with repetitive clicks

Our first customer, a behavioral healthcare company, came to us with a repetitive-click problem. Their end-to-end invoicing process took a person 45 clicks to complete. After walking through the process with the process owner, on average, it was taking that person 30 minutes to complete the task. In addition, only one invoice could be processed at a time, leading to a max output of 15 invoices sent per day per person.

This company had two people responsible for this task every day, so the max output of invoices sent for this customer was 30 invoices per day.*

*Note: 30 invoices per day max output for this customer was leading to a growing backlog of invoices not being sent out.

Then the as-is cost must be assessed

Once we calculated the as-is max output, we were able to assess the current cost per invoice sent by using this equation:

Cost per invoice sent = (Fully-burdened hourly cost of person) * (Person time spent sending one invoice)

Our customer was spending on average $9.06 to send each invoice using two people.

Knowing the transactional cost (per invoice sent) is the first key metric that is needed in the value equation.

Next, the to-be cost must be assessed

Thoughtful analyzed the current invoicing process and determined that 90% of the process could be automated with robotic process automation technology, or bots (What is Robotic Process Automation?). In addition, we estimated that a bot could perform the task in 1 minute versus 30 minutes. We used this equation to determine the to-be cost:

Cost per invoice sent = ((Hourly labor rate of bot) * (Bot time spent sending one invoice)) + ((Fully-burdened hourly cost of person) * (Person time spent sending one invoice))

Our customer would spend an estimated $1.11 to send each invoice using bots.

This is an 87% cost reduction. Using a basic framework, we can bucket this cost savings into a simple value assessment.

  • 1-20% cost savings = Low Value
  • 21-40% cost savings = Medium Value
  • 41%+ cost savings = High Value

In this example, 87% cost reduction is a grand slam from a value perspective.

Last, the automation complexity must be determined

How complex an automation is determines its feasibility and likelihood of operating successfully. At Thoughtful, we use a complex analysis to determine this scoring, but to keep it simple, I've broken it down into a basic high-level framework.

Low Complexity

  • ≤ 75 clicks
  • 1 - 3 systems
  • No Machine Learning, Artificial Intelligence, and/or Object Character Recognition

Medium Complexity

  • 76-150 clicks
  • 3 - 4 systems
  • Some Machine Learning, Artificial Intelligence and/or Object Character Recognition

High Complexity

  • 150+ clicks
  • 5+ systems
  • Major Machine Learning, Artificial Intelligence and/or Object Character Recognition

In the invoice example for our customer, their automation scored as a low complexity automation.

Time to use a classic 2x2 to rate the opportunity

On the X-axis, we assess value, and on the Y-axis, we assess complexity. These two variables are inversely proportional to the quality of the opportunity.

For the invoice automation example, it rated super high on value and low on complexity, which put it on the top-right quadrant of this 2x2, or the 'Automate' section. For opportunities that fall in this category, there is high ROI and low payback periods for investment in automation.

For opportunities that fall in the yellow or 'Needs Assessment' portion of the 2x2, these use cases need to be evaluated at a deeper level to ensure optimal ROI will be ensured based on the companies prioritization of investments.

For opportunities that land in the red area or 'Low ROI', these opportunities are not suited for automation at this time.

Discovery made easy at Thoughtful

All of this information can seem overwhelming. That's why we help our customers in the first part of the automation journey... the discovery phase. Our automation experts know how to assess use cases for value and complexity to ensure you're prioritizing the right processes to automate and deliver 150%+ ROI in year one.

Discover your first automation use case here

Check out the original article


Sign Up for Thoughtful+

Get product updates, company news, and more.

Sign Up

Published On:

April 25, 2024

Related Articles:

Automation & AI

14 Ways to Stand Out in a Marketing Role in an AI Startup

To excel in marketing for an AI startup, you need more than just a surface-level understanding of the industry. A solid grasp of the technical aspects of AI and related technologies is crucial.


Automation & AI

The Impact of Al on Healthcare Supply Chain Efficiency

The seamless flow of medical supplies, equipment, and pharmaceuticals is crucial for maintaining high standards of patient care and operational efficiency. Read more about it here.


Automation & AI

Everyone is talking about Artificial Intelligence; what exactly is it?

AI is the ability of a computer system or machine to exhibit intelligent behavior, often using techniques from computer science such as machine learning, probabilistic reasoning, fuzzy logic, robotics, computer vision, natural language processing, and more.