RPA is a fantastic tool but it can be complex and expensive. Whatever your vendor tells you, implementing RPA correctly takes considerable time and money.
An automation program within an enterprise of reasonable scale may cost $250k in the first 6 months alone (e.g. software, consultancy, robotics lead, cloud architecture, process mapping software, coders, analysts, etc.). Therefore, selecting your RPA tool should not be a quick decision. Building the right foundations for your digital transformation program is essential to delivering multi-million pounds of long-term business success (i.e. significant investment returns from the automation of processes). As we discussed in our last article there are many questions you should ask before deciding to invest in RPA in the first place but done right, this is a very worthwhile journey.
Now that your company has decided to invest in RPA you need to select a product that best suits your bold business strategy. Below are the first 8 key criteria that should be considered as you select and score your RPA product or platform be it propriety or off the shelf.
1. Cultural readiness of your organisation – if your organisation is not ready to transform (e.g. it is siloed); is resistant to change; or has IT and business teams that simply won’t ever talk to each other then don’t read any further; stop now and don’t invest in RPA until your solve your people issues.
Alternatively, if you truly believe that the establishment of a co-dependent IT / Operations automation program will bring your organisation together as they deliver on a common goal, then proceed forward at pace.
2. Strong and active user community – to learn and grow you need an active community from which you can source new learning. RPA has a growing community that is welcoming and open to sharing information but not every RPA product has lots of experienced users.
3. Demonstrated capability in delivering for an organisation of your size, industry and complexity – does your system provider have experience of your line of industry with a product that has proven to work in that space or will they be learning with you? Does your vendor software work well with your systems and software?
4. Solution durability (low cost of maintenance with a less buggy, more durable system) – how durable is your vendor’s product? Is it robust, non-buggy and has a 99.999% uptime or does it fall over at the slightest opportunity? References and live practice with the product during a POC / POV period will help you understand this.
5. Automation process discovery and mapping tooling – being able to find suitable processes for redesign and automation is one of the most time-consuming tasks in RPA land. Therefore, using a tool to map processes can prove invaluable IF the processes are end-to-end computerised and you are not simply going to replicate the process using RPA without deciding if the process can be stopped or if it can be vastly improved using intelligent automation technology first.
6. Access to skilled talent– there are a variety of ways to get access to skilled RPA talent (e.g. recruitment partner like us at Edge Tech, direct-hire, system integrator, contract, vendor, etc.). Skilled developers who have real-world practical experience can be hard to find if you try and go at it alone. Working with Edge Tech can take that pain completely out of the process for you. Ensure your RPA product has a well-stocked supply of affordable market talent otherwise your program may grind to a halt. You can also train and develop your own talent internally too, so ensure your vendor has clear learning and certification paths available.
Low-code application development software gives non-programming, tech-savvy business user, the ability to visually build out applications. An in-house, citizen-developer lead approach can help ensure that end-users rapidly build out the functionality they need to succeed.
7. Coding language and product extensibility – an RPA product should include a lot of ‘out of the box’ functionality, but you will still need to amend the vendor-supplied code objects to better allow you to meet your specific business needs.
Different vendors allow more or less access to their code and operate using different code languages e.g. c#, VBA, python etc. Look behind the scenes to ensure you can amend the code and that the coders required are both plentiful and affordable in your market otherwise this is a hidden cost you will discover at some point in time that can make your business case fall over.
8. Attended, unattended, hybrid automation capabilities – RPA products offer come in a variety of flavours from which to pick. You need to decide whether attended, unattended or hybrid is right for your organisation and score accordingly. Choose carefully and resist the temptation to ignore an unattended automation product that might be better suited to your needs ‘just in case’ things change later. Unattended automation products can be extended to become more like attended products with 3rd party tools or clever coding and robot monitoring.
9. POV / POC experience – a Proof of Concept (POC) or Proof of Value (POV) exercise is key to your trying before buying a tool. If your vendor is truly committed to your success, they will provide significant help during this phase. Do expect to pay a systems integrator up to £50k for a proper process selection/proof of concept build but do expect your vendor to help.
Automating a simple task in 3 days is NOT a valid test of a product. You need to select a process that will stretch your vendor’s product but not be so big as to prove impossible to complete in 8-12 weeks. Test several RPA products, build your business case, get fully involved in the process including coding; design thinking and lean for digital sessions, documentation, Agile ceremonies and planning sessions to fully appreciate the journey ahead.
As your organisation completes a POV and learns more about the art of the possible and the impossible, you can start to further refine your intelligent automation business case. If you have no automation experience then work with a company who does.
10. Complete a back-end database compatibility check with your IT road-map and affordability. Ensure the vendor’s database back-end is on your organisations IT road-map (e.g. numbers of firms are deliberately moving away from SQL due to the cost of licences but SQL is the database back-end of many RPA products).
11. System security standards (e.g. encryption standards) – key for banks, financial service companies or indeed most companies is security. Ensure your vendor’s products are certified as secure (ask); that data is encrypted in transit and the excellent security surrounds the product in general (e.g. has the code been scrutinised for security leaks and product been pen tested by a reputable 3rd party).
12. Role-based access and/or Active Directory Integrated – does your product integrate with your active directly allowing you to enjoy the same secure single sign standards your other applications offer or do you have to manually create and delete user accounts and roles? If it is the later then please cost for additional time, and security audit checks, to manage user accounts outside of your normal process. Your IT team need to be involved in this process to ensure your RPA platform follows your organisation’s security standards.
13. Logging for security, compliance and auditing – your risk and compliance team should insist your software logs all user (and robot) actions and that this log is easily auditable for security tractability purposes.
14. Solution cost – initial and approximate ongoing cost per year (inc. staffing requirement) and Cost of maintaining the product e.g. updates, staff time, etc. – buying an RPA product is often the cheapest part of the process; the upkeep, team and maintenance of your RPA product over multiple years can be significant.
15. Disaster recovery capabilities – how easy is it to recover from complete or partial system failure. Always remember to hope for the best but plan for the worst. As you scale your RPA program your organisation’s dependence on robots will grow exponentially. Do remember to implement a robust business continuity/disaster recovery program. Note: Firms need to ensure robot processes are well documented and fully considered in business continuity planning and testing programs. Firms often ‘forget’ a process when people no longer complete it.
16. RPA process analytics and management information capabilities – one of the most significant benefits of your design for the digital program is the insights you may glean from newly visible process analytics e.g. process volumetrics, wait times, cycle times, exceptions, etc. These are vital statistics that can help you re-engineer and optimise your process outcomes if used correctly.
17. Vendor road-map – bind yourself to a large or small vendor that has a solid, well though thought plan for the future that maps to your intelligent automation enabled digital strategy.
18. A number of existing clients and industry examples – existing volumes of clients can indicate success but the top vendors (and the firms themselves) are struggling to get those clients to move beyond a small number of robots to scale. Whilst a useful guide to product-market fit, this number is not always an indicator of success.
19. Exception handling capabilities – exception handling is the management of process exceptions in ‘unassisted automations’. You may not be able to plan for every scenario a robot might encounter. Dirty data can cause a robot to stop working. Were humans come across something new they have a manager or colleague to query, robots can’t do that. Robots rely on robot controllers / coders to help answer new queries. How quickly you identify and resolve exceptions will determine how quickly your robots begin to work again. As you become more and more dependent on robots fast exceptional handling is key.
20. Cloud is proven & scalable – to scale RPA you need to build on a cloud platform and not all products are cloud-ready. Check. Seek reference sites and have a conversation with a vendor cloud architect. Expert RPA cloud architects are an elusive golden goose of RPA specifies who are few and far between. Do mark the strength of sample clients, client satisfaction, and industry references of RPA product users in your business area.
21. Web-based interfaces or PC based – how easy is it to control your robots from anywhere at any time? Secure, web-based interfaces offer operable from any device, anywhere offer you a great deal of flexibility as you program grows.
22. Scheduling functionality – can you set in place times or logic to get your robots to trigger actions? If you robots need to remain always on for a process to run, you can find that you need many robot licences (and much cost) to run many processes. Having total control over when, which and how many robots run will not only help your cost management, but this will also provide layers of control over process completion 24/7, 365 days per year.
23. Version control and code repositories (e.g. git hub integrated) – does the product have code and system version control? if you delete a sizeable chunk of code in error can you recover fast or is the code gone forever? if you are auditing what happened, and when, by whom have you the code that was executed then (but since changed) easily to hand for your audit team or is this lost forever?
24. The flexibility of a licence model – are you locked into a robot per process model or can your robots, just like human workers, run many processes every day? The first model can be very expensive if you have a lot of different processes to robotise.
25. Ease with which your developers & your citizen developers can code – is key to your truly expanding RPA across your organisation. Key also is the ‘controllability’ of that code (i.e. code governance). Coding or scripting should be reasonably simple, after a short period of training on code standards, for both techies and non-techies.
Code needs to follow organisational standards (these need to be developed and rigorously monitored by your Robotic Center Of Expertise | Intelligent Automation Centre of Expertise) and be reviewed by centrally by an expert coder before distribution. The easier it is to build process automation, both simple and complex processes, the more likely RPA is to scale and scale affordably. If you need specialist RPA coder talent to develop your program, then expect to scale developer numbers, as well as developer recruitment and training cost in tandem with the growth of your digital transformation program. Your RPA platform should allow you to build reusable components. Unless businesses think in terms of reusable components then they run the risk of waking having learned none of the lesson’s learned from the past. The pace of change from traditional to digital businesses requires businesses to think forward and design reusable code, business processes, data and infrastructure components to compete now and in the future.
26. Reference industry feedback on RPA product and vendor – how satisfied were other customers with the product? Don’t just ask your vendor for a reference site as none are going to pick one which was a failure but look around to get a range of opinions from folks who are using the software in live environments.
27. Financial strength of your vendor – don’t jump into bed with a vendor who may disappear. There are a great number of vendors appearing in the market chasing private equity cash and whilst the opportunity for vendor organisations is huge not all will survive financially. Consolidation is likely as is the exit of companies who cannot make money.
28. Scalability of product – Can your RPA platform scale as your scale both your business and your ambition? Does it work on the cloud to allow you to flexibly ramp up?
29. Code integration options – If you want to extend your RPA platform then common application integrations such as Windows .net framework; API integration capabilities; ability to embed code; ability to access external scripts; etc. are all key. The more integration options available to you, the more you can do with your chosen RPA platform.
30. Depth and breadth of the actual product – does the RPA tool have a lot of features and functionality that will serve you well this year, next year and the year after? Is there an ecosystem of partner add-on products that will enable you to deliver an end to end organisational digital transformation program by downloading assets or seamlessly integrating with other applications that can be incorporated into your digital program (e.g. RPA, AI, ChatBots, NLP, ML, etc.)?
31. Change management and governance capabilities – governance and change control are two massive considerations needed to manage your processes as you scale forwards. Your process selection, change control, governance and business case tracking process over time are key in ensuring you deliver real value from your RPA platform investment. We live in a digital age where IT-driven change had never been faster, nor will it ever be slower again. Having excellent change control and governance programs in place to support deliver this change is going to be key for RPA and intelligent automation success ahead.
32. Post-implementation partnering/support model – finding your first processes to automate is the easiest part of your journey. However, organisations often struggle to get past a POC or past their first 5 or 10 robots. Help is required to grow yet so far vendors have only been truly been focused on their first licence sales. Check what support your vendor offers on an enduring basis and score them low if none is available. A website with FAQs, training materials and articles is not a post-implementation service!
33. System integrator – if your vendor or a 3rd party is required to get you up and running then remember to score that capability too (e.g. number of deployments to date; number of successful clients delivered for; RPA resources employed with real experience of your sector – ask to see CVs of those installing, developing and supporting your program; number of automation architects with how many successful installs they have under their belt – ask for reference sites; cloud experience and credentials – ask for reference sites; etc.).
34. Strength of ‘out of the box’ process library – does the vendor’s RPA platform come with lots of pre-built routines for the many typical market software out there or do you have to build from scratch? Can you access a repository of other customers code blocks or again do you need to begin from zero? If existing code exists this can save you considerable development expense.
35. Ease and features of product configuration – how easy is the product to set up, configure and maintain? How malleable is the RPA platform to your business?
36. Debugging features – how easy is it to find and fix the code your developers or your system integrator’s developers or / contractors developers developed?
37. Embedded or accessible intelligence – RPA will get you so far and is excellent for automating rules-based processes that follow a reasonably simple, but well-defined logic. For more complex processes (and you will quickly discover how limited RPA can be) AI capabilities and/or integrations with 3rd party AI / ML / NLP tools is key. There are huge business benefits for organisations that can successfully deploy AI, ML and RPA. RPA and AI/ML together, known as cognitive RPA, enables you to automate more complex, value add process than either technology alone.
38. Cost of maintaining the solution – Buying and configuring an RPA platform represents only a small proportion of the costs your organisation will bear. There are many upfront costs (e.g. licences, staff hiring and training, consultancy, cloud costs, etc.) and many ongoing costs that need to be built into your business case (e.g. annual maintenance and upgrade costs, staff wages and emoluments, code maintenance costs, etc.). RPA can support deliver a great deal of monetary and non-monetary business returns (e.g. improved CSAT or staff NPS scores, reduction in process variability risk, etc) but make sure you account for all costs in your business case before you begin your RPA journey.
39. Does your RPA product work successfully with Citrix (ok, so prove it)? – if you run on Citrix (screen images) then you need to really test the vendor’s product to determine if it will work for your organisation. This can be a deal-breaker for many organisations.
40. Customers support options – what support services does your vendor offer, in what time zone, over how many hours per day (per geography) and what are the SLAs associated with that service? How much does each service offering cost? Does your vendor provide a comprehensive suite of documents to help you successfully build and run your RPA product e.g. architecture; security; integrations; comprehensive RPA run-book or playbook available; typical process suited to RPA; code libraries; cloud install methods; training and support materials; etc. As you grow your robot numbers and spread RPA into multiple, follow the sun, geographies your need for vendor query or bug fix support will grow (vendors note these are called ‘bug fixes’ not ‘feature updates’ in the real world!).
Selecting the right RPA tool or indeed intelligent automation tools is key to the success of your RPA journey. Get your tool selection right and you give yourself every chance of success. Get your tool of choice wrong and your costly business error may result in a loss of money and indeed your job.
What do you consider to be the most important criteria for selecting an RPA platform?
This article is the final part of our Reboot RPA series. Reboot RPA is a series of blogs written by real-world experts to help you overcome the challenges of selecting, implementing and scaling an RPA platform. We would like to thank Kieran Gilmurray and everyone else who contributed in putting this together.