Robotic Process Automation (RPA) software is in vogue these days. Over the past 12 months, I’ve used three different software solutions for different clients and thought I’d share my experiences to help fill a gap in the RPA information available online.

Current RPA solutions tend to benefit human-involved processes with simple tasks and structured data at a high volume.  It is anticipated that RPA solutions will continue to evolve to being able to handle unstructured data and complex tasks.

There have been lots of articles written about why you should implement RPA software (improve consistency of repetitive processes, reduce cost, improve reportability of operations) and innumerable articles on how to implement an RPA project (start with a pilot, build capability and expand), but there are very few articles on which software to use. So, to help fill that gap, here’s what I‘ve come to realise:

It actually doesn’t matter what RPA software you use.

This is true, at least, from an output perspective. Whilst each of the different software packages have strengths and weaknesses, unsurprisingly, they are all pretty good at automating processes. The differences in the software relate to:

• how the vendors licence their software and
• the fit of the software with the skillset of your team.

RPA Licencing Models

It’s pretty clear from the different models being used by RPA providers that they are still trying to figure out how to make money from RPA software. The licencing models range from a Blue Prism you-can-only-use-our-software-if-you’re-a-big-consulting-firm-willing-to-put-a-shed-load-of-staff-through-our-expensive-training-program to a WorkFusion you-can-use-our-software-for-nothing-provided-you-don’t-need-to-connect-to-an-API.

I categorise the vendors into two main groups:

• Those who think the major consulting firms will keep a strangle-hold on RPA, and
• those who think that the RPA software winner will ride to victory on a ground-swell of support from end users.

The first group includes software vendors such as Blue Prism. I call this the Siebel model. The second group includes vendors such as UIPath, WorkFusion and Softomotive who are pursuing a number of different models ranging from cripple-ware (WorkFusion), to cripple-licencing (UIPath) to single-robot deployment licencing (Softomotive).

Below are my views on each of these approaches:

The Siebel model

Blue Prism and others in this group of vendors only permit consulting groups who have put a number of staff through an extensive training program to implement their software. I can’t see this being the winning business model in the RPA space. I see a future where almost every company can automate their own processes in the same way that today almost every company can create their own spreadsheets and presentations. Blue Prism’s Siebel-like model will inevitably be swept away by this wave.

The WorkFusion cripple-ware model

WorkFusion takes a different approach to licencing. They want as many people as possible to use their software so you can download RPA Express at no cost and go through their training exercises in your own time. It’s a capable piece of software with the ability to use Groovy (a Java-based language) to write your own functions. WorkFusion’s business model is based on users moving from RPA Express to their Enterprise-level Smart Process Automation (SPA) application. From a licencing perspective, you can use the Express application in your organisation but it is missing some key features available in the SPA application — most notably, from my perspective, is the ability to easily connect to APIs.

I like this model better than the Siebel model. I can see a technically-inclined user in an enterprise:

• picking up WorkFusion Express,
• doing something useful with it,
• realising that they need the SPA version to expand their usage, and
• getting budget and buying SPA.

The hardest part of crippleware is deciding which parts to cripple and managing your userbase who will, at times, be pissed off that they cannot do what they need to do without buying a licence.

If you and your team come from a Java background, WorkFusion is a good choice for you.

The UIPath cripple-licence model

UIPath has taken a different approach to WorkFusion. You can download the fully-functional version of the application at no cost but:

• you don’t get access to the Orchestrator component that allows you easily manage an army of bots, and
• you can only deploy it without a commercial licence if your company has under $1M in revenue.

Technically, this approach is easier to manage than cripple-ware, but it comes with some risks to their business model. A reasonably technical user can push robot usage data to an API which would allow them to somewhat mimic the reporting functionality of the Orchestrator component. If they do this, and they don’t care about licencing violations, then they can use the software for free.

UIPath is betting that this will be a rare occurence and, at this point in time, I believe it is. But I see it as almost inevitable that someone will release an open-source Orchestrator that will be easy to use with UIPath and UIPath will need to reconsider their model.

UIPath is an excellent piece of software and their training resources are exceptional. Building and integrating your own .Net functions is a trivial exercise. If you and your team come from a .Net background it would be hard to go past UIPath.

The Softomotive single-robot licence model

Softomotive is my go-to software for most of my smaller clients because of the single-robot licencing model available in the Winautomation edition .

Unlike WorkFusion and UIPath, you can’t extend the software by writing your own modules but the functions native to the application are sufficient for most use cases I encounter.

Runtime licences for a single robot can be purchased for a one-off fee of US$59 and a machine-licence that allows an unlimited number of bots to run on a single machine can be purchased for US$279

Like UIPath, Softomotive has an upgrade path to Process Robot which allows the orchestration and monitoring of robots.

Whilst the training material for Softomotive products falls short of UIPath and WorkFusion, the application is easier to learn and easier to use than either of its competitors. You can get up and running very quickly even if you don’t have a programming background.

The coming tsunami

So, which model will reign supreme? In my view, none of the above.

Siebel died not only because of the rise of Salesforce but because the software “Suite” providers started including sales and marketing functionality at no additional licencing cost. This, combined with the rise of solutions like Salesforce, ate away Siebel’s customer base from both ends — the enterprise market and the SME market.

RPA software is headed along a similar trajectory.

Most of the solutions that I deploy rely on cloud functionality. I use RPA software to handle the interaction with on-premise systems but, as soon as practicable, I push the data to an AWS Lambda function or, more frequently, an Azure Logic App to manipulate the information and manage the workflows. This is a robust development pattern because, as early in the process as possible, it moves data from systems that are not in my control onto systems that are, thus reducing the number of things that can go wrong.

This pattern is not escaping the attention of the cloud platform providers and I expect to see RPA offerings emerge from all of the major cloud platform providers — either through in-house development or acquisition. In fact, I’m surprised that Microsoft has not yet acquired UIPath or Softomotive and that Oracle or IBM has not yet picked up WorkFusion.

In any case, despite their performance over the past year, I won’t be buying shares in Blue Prism right now.

Doug Hudgeon  is an  experienced electronic invoicing and back office automation experts and Director and Founder of My Business Automated http://mybusinessautomated.com/. Contact him at doug@mybusinessautomated.com