14 rules for Robotic Process Automation (RPA) and Intelligent Automation (AI) success

Back to Blog
Phillip Glickman - Unsplash

14 rules for Robotic Process Automation (RPA) and Intelligent Automation (AI) success

Robotic process automation (RPA) is a very exciting, but simple software tool; yet 50% or more RPA programs are failing. RPA is not a panacea for digital transformation but it is a highly effective tool in your intelligent automation armoury. This article highlights 14 rules for automation success and follows on from a similar article released as part of a Reboot RPA series.

1. Begin with the end in mind. Always understand where you are going before you begin your digital transformation journey. Businesses need to clearly articulate the business problem they are trying to solve, and only then purchase the intelligent automation tools they need to deliver the outcomes they want to deliver. Real transformation involves re-configuring how your organisation delivers value in this digital age (e.g. think Uber and Amazon, not Blockbuster sending DVDs in first class as opposed to second class post). Set a game changing aspiration of 35 percent plus impact. Then use RPA and Intelligent Automation (AI) to fundamentally restructure your cost base as you re-image how you deliver value to your customers to gain competitive market advantage. 

“Companies need to define the ideal situation that they want to reach and then prepare a plan towards to that. If that plan requires RPA, then that is fine. But not every company needs RPA to solve their big challenges. If RPA is required, then it must be part of the wider business strategy and officially accepted by the both the Business and IT.”

Konstantinos VogiatzakisGlobal RPA Lead at Babylon Health

2. Buy the RPA platform that best fits your organisations delivery needs. First understand the problem that RPA will solve and then pick the right RPA platform for your organisation. For example, is RPA needed to solve integration challenges between external / internal legacy systems? Is RPA needed to automate back-office processes (e.g. finance, HR, supply chain) or is it needed in the front-office (e.g. call centre, sales teams) or both? Do your business users need to be able to configure their own robots (i.e. citizen developers)? These are questions your organisation needs to answer before you pick your RPA platform. Take as much time as you need to fully understand the RPA platform your organisation needs, understand what RPA can do and what it can’t do. Otherwise, you run the real risk of your RPA implementation falling very short of your initial expectations. Remember to focus on business outcomes not on the number of robots purchased.

“Automation is complex and multi-faceted; the chances that one tool fits all are slim to none. Therefore, it’s imperative that you aren’t naive, negligent, and don’t buy into the vendor hype and marketing BS – do your research and ensure proper due diligence of the available and chosen tech stack against your needs.”

Paul Arnold, Head of Product and Development at Cortex Intelligent Automation

3. An intelligent automation toolbox is required to fix your digital transformation light bulb. RPA alone will never drive end to end business transformation. RPA is a powerful tool when business use it correctly but just as Rory McIllroy probably can’t win a major with a 7 iron (though he might) a business can’t digitally transform with one intelligent automation tool. Limiting yourself to one tool will limit your scope for automation. Businesses must adopt a practical approach to intelligent automation using a sequence of intelligent automation tools to solve well defined, end-to-end business problems that deliver tangible business outcomes (e.g. RPA, AI, ML, NLP, Chatbots, OCR, Python, Excel, .NET, APIs, VBA, process mining, digital workforce and smart workflow software, etc.). Businesses must use the right tool for the job because it is the best tool, not because it is the tool you just happen to have available.

4. Decide where RPA sits within your organisation. One of your first decisions will be to determine whether the RPA delivery team will sit under IT or Operations. Regardless of where your organisation decides RPA should sit, experience suggest it best fits within Operations, IT buy-in is needed from the very start of you automation journey. If your program does not have IT support, it will not be successful. 

5. Build in-house capabilities, then coordinate the right teams to deliver your intelligent automation program. 

Business and technical skills are required to succeed. Business process automation requires both technical skills (e.g. IT, design thinking, lean for digital, RPA, etc.) and contextual knowledge of a business and business processes to prosper. Engage IT from the beginning as you will need software developers (RPA, VBA, Python, C#, etc.), cloud specialists, RPA architects, IT architects, business analysts, security specialists, data scientists, and more besides to run RPA at scale.

“Building a cross functional and cross skilled team is an extremely important aspect of any digital transformation. You need to ensure you have the right business knowledge and technical knowledge for your company to exceed. As well as this you need to ensure the team has the right attitude and is not afraid to get their hands dirty and step outside their comfort zone. You want team members who are not afraid to try new things and help out wherever possible, even if it is not their immediate responsibility”

Matthew Coffey, RPA Delivery Lead at Pearson

IT resources often lack business context. Therefore, an RPA transformation program also needs subject matter experts with end to end business process knowledge who can clearly articulate what is happening in the business and what business outcomes are required from the process too. You will need a business architect to redesign your organisation for the digital world before you simply automate your current processes. In addition, common sense; a willingness to try-try-try and try again when things don’t go right the first time; a great dollop of realism; an accountant to help build your business case; an executive sponsor who is genuinely dedicated to the program; an Agile scrum master and a sense of humour are also often required. This mix of skills will help you get the most out of your intelligent automation program. 

“It is generally understood that all parties want the same things in a healthy relationship. This question needs to be RE-answered at regular intervals. A typical project will includes phases like Assessment, PoC, BETA, ALPHA, and Production. Structure your agreement with Go-NoGo reviews between each phase to limit risk for all parties.”

JD Wilson Jr. RPA Hyper Innovator

6. Be willing to invest considerable time, money and energy over many years in your digital transformation program. Tool selection can take 4 months; a pilot should take 8-12 weeks and cost around £50k; the first 6 months of an RPA program will most likely cost you £250k (think licences; system integrator and consulting help; recruitment cost of an RPA lead/team; cloud installation costs; RPA process analyst; purchase cost of process mining software; etc.) and an end to end and a fully-fledged RPA program with a business case built out over 3 to 5 business years will cost your multiples of that first 6 month cost. You need to be confident you have counted all of the costs of a multi-year scale RPA / IA program and have built a business case that returns a multiple upside way in excess of your program cost. Vendors often quote programs paying back after 6-9 months but this is often not the case in larger enterprises where it can take months to simply get a program up and running. Therefore, any vendor promises that suggest everyone gets their money back in 3-6 months is as questionable as it is indefensible. RPA, done right, is neither cheap nor fast. It can be inexpensive and faster than other methods, but RPA isn’t free!

‘I’ve seen far too many instances where Operational Leaders are duped by the sales pitch, believe RPA to be the ‘silver bullet’ to reduce cost to serve and dive head long into running a pilot on the most obvious manually intensive processes, involving IT infra set up and purchasing tens of licenses, without completing ‘due diligence’ on the remainder of the organisation. Result is a struggle to identify further opportunities with tangible benefits reducing the ROI’

Allan Surtees, Head of Intelligent Process Automation at Lowell Financial

7. Have your systems integrator put skin in the game. Ensure that your systems integrator is willing to put skin in the game. That way they share both the considerable risks and rewards of your transformation program. Don’t take all of the risk onto yourself, nor be greedy and unwilling to share any financial returns that you never had before. Vendors will complete gain share agreements if you ask. That way they, and you, have skin in your RPA game. Don’t short change your system integrator by not putting 100% into working with them side by side so that they and you succeed together. Set meaningful SLAs and measurable KPIs and hold yourself accountable for delivering real business returns. Count returns as a measure of your success not the number of robots in your business.

8. Executive sponsorship for your intelligent automation program is a must. Your intelligent automation digital transformation program needs to be one of your organisations top 5 focuses and an executive sponsor with real money, time and political weigh needs to be 100% behind your program. As a minimum, an RPA program should have the blessing of the COO and CFO and ideally the CEO themselves. Anything less than 100% executive (and senior management) support will almost certainly result in your program failing. If the c-suite are not pushing for transformative change then your program probably wont succeed.

“RPA requires buy-in from all levels of the business, but for it to be successful firmwide, it needs to start at the top. I’ve seen many examples of where the wrong stakeholder have been engaged from the start and this has led to a failure in getting traction on the automation journey“

Amyn Jaffer – Head of Intelligent Automation at Ultima

9. Start again if your program does not deliver the value you said it would. There is no shame in starting again if your current RPA program is not working. RPA can fail the first time if you don’t have the right tool, advice or people running the program.  Alternatively, if there are systemic reasons why your program failed then walk away. Don’t throw good money after bad if there are genuine reasons why RPA / IA wont currently work (e.g. sudden merger with another company is taking everyone’s focus).  

10. Be biased towards speed and impact but work toward the long term. Build a small centre of expertise (COE) team with the technical, resiliency and organisational capabilities needed to make automation stick. Then create an intelligent automation digital factory that rapidly delivers sequences of automation modules into production.  Combine quick wins within a larger, longer-term transformation road-map. Quickly test what does and does not work and make changes accordingly. Optimise and improve your code over time to deliver sustainable value and move towards your end to end target operations model systematically. Implement design thinking, lean for digital and Agile ways of working to re-imagine the way your organisation captures and delivers values.

“Create a methodology and set strict reviews around it. The art is keeping it simple; any idiot can make this difficult and leave a solution that is impossible to maintain.”

Gavin Price, RPA Jedi

11. Plan for worst and hope for the best. Always begin any transformation program with a well thought through risk and mitigation plan in advance. List the top 20 or 30 risks to your IA transformation program and fully work through the mitigation’s that need accounted for before you begin your journey (e.g. RPA skills gap, end user resistance, lack of suitable process pipeline, changing business context, business merger mid program, etc.). Organisations always plan for success but when things go wrong (and they will) they then scramble to recover. It’s best to think of all the things that can go wrong before you begin and plan for how you might solve these in advance of them going wrong (i.e. plan for worst and hope for the best). Most organisations lack RPA and IA experience. Therefore, you will find that you will learn many lessons as you roll out your program, so regularly update your risks document as you progress. 

“We often see companies investing an automation capability, but not securing the business engagement necessary to provide a steady stream of sizeable processes suitable for automation. That does simply does not work. In fact, these companies often make a net loss on their investments in automation”

Joe Wheatley, Principal Consultant.

12. Deploy in the cloud. If you want to digital transform at pace and scale then you will have to deploy your program in the cloud (e.g. on AWS, Azure, or even internally on VMWare). Cloud offers the speed and flexibility you need to transform at pace. Cloud services will take to get up and running (e.g. understand, hire cloud talent, secure holes in the firewall, get CISO / risk / compliance /onside; etc.) but will pay for itself many times over as you benefit from greater returns faster than any other method. Cloud companies (e.g. Amazon Web Services, Microsoft Azure, Google Cloud Platform) also offer cognitive cloud services (e.g. AI, translation services, NLP, ML, etc.) which can be used to make your digital worker cleverer (aka Cognitive or Automation or Intelligent Automation).

13. Prioritise the processes you go after. Process selection is key. If you select the wrong process you are setting yourself up to fail before development begins. You need to ensure that you apply clearly defined RPA selection criteria to every process before agreeing to automate. Consider volumes, rules, complexity, change schedule and run time. For your first process keep it simple yet ensure it delivers enough value so you can shout about it internally. This will help you to get buy in and understanding from the business. Redesign your subsequent processes for a digital world to ensure every process delivers real value against your objectives.

“One of the big mistakes we see companies make over and over is trying to take on this challenge of IT transformation in one fell swoop. They start looking at the tools first and try to work backward from that to prioritise the processes they go after, often getting overwhelmed and stuck in the process. This is all about driving towards better business outcomes, and it can and should happen in phases. Spend some defining – even in broad strokes – what business outcomes you want to achieve, then create a prioritised list. Stay focused, get some wins on the board, evangelise the success, then expand the scope to the next set of wins. You’ll be amazed at what’s possible.”

John Grancarich, Vice President Of Product Strategy

14. Focus on people. Culture eats strategy every day. Folks are worried about automation due to the never-ending negative press around job losses due to robots. Experience has shown that very few, if any jobs, have been lost due to RPA. Bring digital workers closer to your people by giving your bots real names and introduce them as junior co-workers hire to do the boring work, so that people can work on more cognitive, complex and creative tasks.

If you are going to shrink head count then focus on cutting future hiring costs (i.e. an attrition model) rather than replacing or removing existing staff. Have a well-crafted media campaign constructed that will address all the concerns and opportunities identified from your stakeholder mapping exercise and continue to really listen and communicate with those impacted (positively and negatively) by your organisation’s digital transformation program on an ongoing basis. Build business user confidence in advance and as you progress your program.

“It’s very important to have Exec level sponsorship but it is equally as important to make sure the people on the ground are communicated to from the outset….very often it’s the people on the ground who play a huge part in teaching the BOTs”

Andrew Hartley – RPA Consultant

Useful links:

  1. RPA Reboot: RPA 101
  2. If you are not willing to go all in, then don’t put on your RPA swimsuit.
  3. The biggest lie told to RPA customers – 50 robots equals success
  4. 40 essential selection criteria to choose an RPA platform part 1 of 5
  5. 40 essential selection criteria to choose an RPA platform part 2 of 5
  6. 40 essential selection criteria to choose an RPA platform part 3 of 5
  7. 40 essential selection criteria to choose an RPA platform part 4 of 5
  8. 40 essential selection criteria to choose an RPA platform part 5 of 5

The bottom line is that RPA is not a silver bullet for digital transformation. RPA is part of the digital transformation of the companies but it can be an exceptionally powerful part when employed in combination with other tools in your intelligent automation toolkit. Leverage RPA as a change agent and always employ a balance of technical and business skills to deliver tangible business outcomes. Do remember that intelligent automation is a marathon not a sprint, so strap yourself in for the long term.

What rules do you consider to be the most important for an RPA | IA program?

Free to reuse: If you think this article would be beneficial to others; please do feel free to share it. If you want to post this article on your LinkedIn page then please feel free to do so. The more information we share within the RPA community the more likely businesses are to succeed with this excellent technology.

Further Help: If I can help you in any way please do reach out.

Note: The views expressed above are my views and not those of my employer. 

Share this post

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to Blog