There are several approaches that you can consider while planning RPA implementation in your company. Each of them has its pros and cons, and weighing in on these can be crucial to making the right decision here. So let’s look at the four most common RPA implementation solutions and explain the difference between them. To give you a quick introduction, Robotic Process Automation (RPA) is a technology that revolutionizes the field of business process automation. Smart virtual software robots can take care of repetitive tasks normally done by humans, which leads to increased efficiency, minimizes errors, and reduces costs. But how to choose the right RPA solution and make sure it will be cost-effective? First, let’s see what options you have.
4 most common RPA Implementation Solutions for the enterprise
When it comes to the question of how to implement RPA in the enterprise and how to maintain and keep your software robots up to date, there are 4 different RPA implementation solutions:
RPA Consulting + RPA Center of Excellence – hiring external consultants to work on your RPA processes and establishing a Center of Excellence (CoE) especially for RPA implementation within your company.
Custom in-house RPA development – your internal team of developers working on basic RPA functionality through custom software development and open source tools.
Outsourced RPA development – hiring an external consulting team to cover all RPA-related tasks.
Robot as a Service (RaaS) – the ‘RPA as a service’ provider offers a complete end-to-end service, including the RPA software.
Now you know the options, let’s see what costs you have to bear in mind and what the advantages and disadvantages are of each RPA solution model.
RPA Consulting + RPA Centre of Excellence (the traditional way)
A Center of Excellence (CoE) is a special team of qualified experts, whose main goal is to provide the best practices around a particular area of interest in the company. In the case of RPA, a CoE focuses on proper RPA technology implementation and automation.
With traditional RPA approach, you have to bear a lot of additional costs to support the robots in-house on top of the license costs:
Process consulting fees
Consulting fees for setting up a Centre of Excellence (CoE)
Hiring and teaching employees for CoE
Continuous running costs of a CoE
On the plus side, you can benefit from possibly faster reaction times, provided your CoE is well staffed, well trained, and knows RPA implementation best practices.
However, this approach leads to a sudden decrease in ROI – in many cases even negative ROI in the long run, as the CoE running costs in many cases surpass the FTE savings.
As a rule of thumb, you will need to recruit or assign a new RPA developer for every 5-10 additional robots, just to keep them running.
Custom in-house RPA development
Let’s suppose your company has a capable software developer team. It may be tempting to implement basic RPA functionality through custom software development or even using open source tools such as web scrapers, test automation tools and headless browsers (like Selenium).
Be wary of this approach though, as reinventing the wheel can cost a lot. RPA platforms have matured over the years and have already ironed out all of the teething troubles your team will experience for the first time. A major setback of developing robots in-house is that your developers will have to wear many hats: consulting with the end user, developing the platform features, and programming the robots will all eat away precious time.
If you talk to companies who developed robots in-house, you’ll realize that in the long run, it’s hard to maintain them. Various developers might use different RPA development methodology, might be haphazard, and with each new robot they implement, the support is getting harder and harder. The technical debt is piling up and soon the robots will have to be treated like a traditional legacy system.
On the plus side, (provided you have a well-staffed RPA developer team) you can expect quick reaction times and license costs are relatively insignificant if the team uses open source technology. So if you have surplus developers and can hardly justify the operating expense of RPA licenses, this approach may work in the short term.
Outsourced RPA development
If you don’t want to bear the overhead and costs of operating a CoE, the outsourced model is a viable option, although it is the most costly of them all.
In this model, you will have to pay the consultancy fees during the entire operation. Both the technology vendor and the consulting firm will take their cut in the value chain.
On the plus side, you don’t have to bear the cost, overhead, or the organizational burden of setting up and maintaining a Centre of Excellence in-house.
Robot-as-a-Service (also known as RPA as a service)
RaaS (Robot as a Service) is a shared software model in which a client can lease the robotic devices as a cloud-based service (all data collected by robots is stored in the cloud). By only paying a subscription fee, Robot as a Service (RaaS) gives you a complete end-to-end solution for process automation from one single provider.
A RaaS subscription usually includes the following:
RPA consulting to identify the processes & workflows to automate
Development of software robots
Installation and configuration of the workflow automation by professional RPA programmers
Automatic upgrades of the robot infrastructure
Support and maintenance, including changing the robot to reflect changes in business or changes in systems the robots use.
Some RaaS providers let you decide if you want to host the robots in the cloud or on your own internal enterprise servers. Having them in the cloud gives you the advantage of being able to scale process automation almost infinitely.
By having a monthly subscription fee, costs are more predictable than in the case of other solutions. The monthly price normally is dependent on the number of processes that are automated. Some also charge by the number of robots. But when on a RaaS agreement, you’re normally not paying for the number of consultants involved. The experts from the RaaS provider are included in the price.
On the other hand, this model gives you less control than RPA implementation in-house among your own employees. You have to consider if it is worth it to spend their precious time on complex automation processes, which require a lot of time and expert-level knowledge or if you prefer to invest in a shared software model.
Summary – RPA Implementation Solutions for the enterprise
Now we’ve covered possible RPA solutions and how they may affect the total cost of automation, let’s summarize:
Here you’ll see that Robot-as-a-Service is the most cost-effective solution. With traditional RPA solution, a customer has to pay for a computing unit, a “robot”, and is responsible for filling up the daily computing time of that device with as many processes as possible. In the case of RaaS, for practically the license cost of a competitor’s robot, you can receive a complete end-to-end service.n/a: not applicable
This has several setbacks:
The overhead of scheduling and orchestrating the different processes the robots need to carry out;
The slow speed of sequential execution;
Any time the customer wants to run a process in parallel to speed up the transaction, they have to pay for extra licenses.
Eg: creating 100 invoices with X company takes 100 units of time.
If the customer wants to speed it up and run four robots in parallel to get it done in under 25 units of time, it is four the price. Cloudstorm, on the other hand, offers process-based pricing. We have a client where literally 100 robots run in parallel to do the same for 1 license.
Choose RPA solution that fits your preferences
However, we all know that every business has its own character; you should make a decision based on your own preferences. We hope this post has made it easier for you to choose the right model and face all RPA implementation challenges.
P.S. In case you need help and want to talk to our experts, you can request a free Process Automation call here.