We live in the era of automation. There is no doubt about it.
From simple Excel macros to supercomplex algorithms used Tesla GigaFactory or Google X lab: different types of scripts, software bots, and actual machines are being created every day to help people work faster, better, and more efficiently.
On the other hand, automation is often nothing more than just a buzzword. It is being put in every pitch deck by startups seeking funding and it’s part of every second strategy plan built by operation executives. In many instances, there is nothing behind these proposals – just empty words without any deeper exploration of the topic.
But you can achieve so much with a properly planned and executed automation! Increasing ROI, shortening payment cycles, freeing your employees from mundane tasks, reducing operational errors… The list is endless.
How to go about automating your business? What should you consider, how to start, how to select the right processes for automation, and what pitfalls should you avoid ensuring the success of your implementation?
You will find the answers in this article. Read on!
What Types of Automation are Out There for your Business?
First, you need to decide on what type of automation will be the most suitable for your business.
Depending on the complexity of the task that is selected for automation there are a few options you have:
Tool-dedicated scripting can be used for automating simple tasks inside a specified application. Deployment usually takes a few days and is limited to a single application. Examples? Excel macros, Outlook rules, SAP scripting.
These are scripts written with the use of a specific programming language, i.e., PowerShell or Python. They can be used to automate a bit more complex task. These automations will still have a limited application interaction. The development and deployment of these usually take a few weeks.
Robotic Process Automation (RPA)
Robotic Process Automation allows building more complex automations. It allows the connection of various tools (CRMs, SAPs, databases, etc.) and automates the processing of data and documents between them.
Thanks to the existence of automation platforms, such as UiPath or Blue prism the development and deployment usually take a couple of weeks and are relatively low-cost. These platforms are also built with the low-code principle, so the developers can focus more on tasks themselves than the tool they are using.
There is also a matter of user interface – RPA can automate any application by interacting with it using UI „just like a human”.
100% custom-made applications, allowing for automating multiple processes and tasks of the highest complexity, all in one box, written from scratch. There is no limit to what can be achieved here, but the same goes for the potential cost. The full deployment time can be counted in months, or even years.
It’s always a great idea to automate your business processes with Intelligent Automation.
How to Identify Processes That Can Be Automated?
Next, you need to select processes for automation. But how to pick the right ones?
From my experience, a vast majority of processes in a given organization can be automated. Although the most important question at this point should be: is it worth it? For example, if manual execution of a process takes a few hours a month, while development would take half a year, can this process be automated? Yes. But should it be automated? No.
The following part of the article will help you in the initial process selection, after which a detailed assessment should be conducted. You can also read our article about RPA metrics to give you more insights on what should you focus on.
Here are the most important factors you should consider before selecting the right process candidates for automation:
- Process Execution Cost: cost and time needed for the manual execution of the process
- Process Errors Cost: how many errors are made during the manual processing, how important is it to limit errors (i.e., compliance violations), what is your company’s business loss due to these errors?
- Execution Time Cost: Does your company lose opportunities due to delays in process execution, is the process a bottleneck wasting the time of employees and other processes?
- Other value factors: limited workforce availability, dynamic process growth, SLAs, specific times of execution (at night), other business KPIs
Ease of implementation
- Complexity: the number of different paths of execution and logic rules
- Stability: how long will the process be executed in its current shape. How big can the modifications be? (For example, we know about tax changes incoming next year, so there is no sense in automating existing business logic as it will have to be redesigned in January)
- Standardization: is the process executed according to defined rules or based on human judgment?
- Data accessibility: whether required data is:
- structured (data tables vs pure text)
- digital (digital invoice vs invoice scan vs paper invoice)
Based on the above factors we can calculate two dimensions:
Automation Potential = Process Execution Cost + Process Errors Cost + Execution Time Cost + Other value factors
Ease of implementation = Stability + Standardization + Data accessibility – Complexity
With these two dimensions we can create a simple matrix that will help you assess the automation potential of your processes:
QUART 1: High value, low (development) effort. The first choice – automate now!
QUART 2: High value, high effort. Processes that should be automated in a second order, after the organization gets familiar with the RPA technology
QUART 3: Low value, low effort – consider for building Proof of Concepts, for utilizing empty licenses,
QUART 4: Low value, high effort. These processes do not necessarily have to be automated and each case should be analyzed separately.
Processes from the 2nd and 3rd quart can be automated simultaneously depending on organization needs and available assets.
Additional factors to consider:
- Packages/bundles: in other words: the economy of scale. With multiple processes that are using the same application/logic, specific parts of automated processes can be reused, making other automations easier to develop (the dots on the chart move to the right)
- Enablers: automating one process might generate data that is required by another one. i.e., uploading scanned invoices data to the database (the dots on the chart move to the right)
- Ice breakers: first few processes automated in the organization shouldn’t be overly complicated and optimally should be inspired by employees, not by management. Process simplicity lets us avoid deadline violations, conflicts, and other problems, thus allowing easier change management.
- Optimizers: developing automation is a great opportunity to standardize the existing process, create process documentation and even optimize it.
What’s important: as more processes are automated, we might find out that processes initially located in the 4th quart have moved to 2nd (by enabling another process) or 3rd (by common application/logic synergies or being enabled by another process).
Shortly speaking: order matters!
Automation Processes: Mistakes You Should Avoid
Here are three pitfalls I would suggest looking out for while selecting a process and planning your automation:
Starting with a high value, high complexity process
New, wonderful technology is announced, great effects come in a short time. Meanwhile, due to lack of experience, we discover that process is even more complex than expected (employees rarely are fully aware of all rules and throw in new ones as development progresses), some functionalities can’t be added, the project takes twice as much time as planned. Where there’s a possibility always start with a simpler task
Ignoring synergies and reusability between automated processes
Some processes use the same applications or same logic. Therefore, parts developed for the first process can be used in the second process, effectively easing and shortening its development. So, we may have 4 processes, each not worth automating, but all are using the same application and similar logic, so developing them together reduces the costs by 40% and now the business case is valid.
Automating processes for the sake of automation, ignoring business case calculation
Sometimes automations are developed because someone is annoyed by a task or a manager has to achieve his KPI “automate 10 processes in the following year”, even though the automated processes do not bring enough business value compared to costs of development and upkeep. Following this path will effectively hurt your RPA-related KPIs and will create a negative perception of the automation within your organization.
What Processes Should You Automate With RPA?
The answer to this question is always tricky, as plenty depends on your unique business needs and the specifics of your industry. Nevertheless, here are some actionable ideas of which processes can be automated using RPA.
Overdue Payments Monitoring and Reminders
Description: Bot monitors sales’ invoices payment status. If the payment deadline is exceeded a reminder is sent to the contractor. Multiple escalation levels are defined.
Why? Besides limiting employee engagement, the automation enables higher process execution frequency leading to a higher collection rate. At the same time, all required data is digitally accessible enabling entirely rule-based business logic, with limited paths (escalation levels) allowing for easy implementation. Medium value and low development effort make it a perfect “ice breaker” process.
Incoming Invoices Upload
Description: Contractor invoices are coming to the dedicated mailbox, bot extracts client Tax identification number, based on which the Contractor ID is queried from the database. Then the invoice is uploaded to specified contractor files in the ERP system.
Why? Thanks to higher execution frequency sales invoices are being uploaded immediately after receiving them shortening total receive-to-pay time. Reduced mistakes limit the risk of unpaid or delayed invoices. Semi-structured, digital, input data (tax number has defined format allowing its extraction even from scan OCR) and single execution path make it easy to implement the process.
Description: Daily bank statements are downloaded directly from the bank website. Scheduled payments are extracted from the ERP system. Data from both sources are merged to create a cash flow report. Expected account balance limits violations are reported to the accounting department.
Why? Cashflow reports have to be ready by the workday beginning, making the processing time crucial, which in pair with a high risk of miscalculation leading to insufficient funds for defined payment puts high pressure on employees. A large number of bank accounts (in multiple banks) and payments generate a significant process execution cost. Automation development for this process is not straightforward due to multiple input sources (multiple banks) and a few process paths, however business logic is rule-based making it a good process in the second automation wave – high value and medium development effort.
Products delivery automation
Description: Based on registered products orders and stock, delivery orders are created in cooperating delivery companies.
Why? The process is time crucial as order handling time directly affects both costs and customer satisfaction, high process volume generates high execution costs with several employees engaged on a daily basis. Time pressure generates mistake related risk of invalid delivery parameters. Process automation is feasible thanks to fully digital source data and target systems. Multiple target systems and execution paths increase development cost; however, it’s compensated by process volume.
Insurance claims handling
Description: Insurance claims are registered in a web application by clients. The web application is not integrated with the claims handling system, so all claims must be copied to the target system.
Why? Large volume in connection with high error vulnerability generates hefty process execution costs. Although business rules are clearly defined there are multiple execution paths (i.e., claims types) and defined exceptions, making it a very high value, high development effort process.
Want to read more examples of processes that can be automated by using RPA? Here is our list of 40 real-world use cases of RPA across industries.
Summary – what’s next?
As you can see, there are plenty of factors to be considered before proceeding with your first RPA implementation. I hope this guide sheds some light on how to go about it!
But if you are still unsure where to move next – simply get in touch with Flobotics. Hire RPA consultants and we will audit your existing processes and help you select and automate the ones with the highest potential of increasing your ROI.