About The Author
Tim Olsen 4 articles
Residence: GB Gallantry Bank, Cheshire
Head Of Projects/Portfolio/Programme

more about Tim

All Authors


The Unified Project Management Dictionary

more terms

Why do RPA projects fail?

Robotic Process Automation is the latest ‘Big Thing’ in IT, with many companies staking their futures on a highly automated operating model. This is often the right aspiration as there can be no doubt that RPA can significantly reduce operating costs, but whilst solution providers are seeing the logo count expand exponentially, few are seeing success run deep, with most companies plateauing at around 3-4 bots. What is preventing these companies from scaling?

Lack of early validation

New entrants to the RPA industry will quickly find that they will face the same problems over and over again unless they learn quickly. Opportunities need to be validated quickly in order to fail fast, or risk burning time and money with limited ROI. Experienced providers will be able to offer a high level business case very early in the discovery process with some confidence, if they have learned the lessons through bitter experience.

Need to see the big picture

RPA tools are great, they are highly effective, but on their own they are limited in their scope. It is only when they are partnered with complimentary technologies such as Optical Character Recognition, Natural Language Processing, Chatbots, AI etc that their scope becomes significantly wider and processes can manage e2e process automation. Your supplier should have these solutions as part of the overall offering, and the solution design should incorporate these applications.

Wrong processes targeted

It is critical that the right processes are targeted during the implementation – the ones that have structured data, high volumes, repeatable processes, rules based and low complexity will deliver high ROI. Get the wrong process, and it will sink.

Strong Support model

Things do go wrong. Generally the software and platforms are reliable, and whilst RPA solutions have become significantly more robust and adaptive, change to the client’s applications is always a weak point unless strong governance is in place. The RPA team need to be involved in Change Control/Release meetings and /or they need to have sufficient capacity and communication channels to respond quickly to rectify any unplanned changes to processes.

Build it and they will come – but only if the framework is in place.

Experience shows that if a bot is implemented, it will build a snowball of demand for more automation – but only if the support framework is strong, and effective. If a bot is allowed to flounder, the business will lose confidence and dismiss RPA as another passing fad.


The reader might conclude from the comments above that RPA cannot be started small scale, however with the right ingredients, an automation acorn can grow. The drive must come top-down in the organisation, and a CoE should be established. Fragmented approaches rarely work well, with cottage industries being implemented but with no central support model, economies of scale won’t be realised. It is possible to start small and grow, but the support model needs to grow in line, not left as an afterthought.

If this still seems daunting, one option to the aspiring (or failing) RPA implementer is to partner with a solution provider who already has the scale to bring all these ingredients to the table from the outset, and implement RPAaaS in your organisation. If you think that sounds expensive, ask yourself, can we afford to fail?

Published at pmmagazine.net with the consent of Tim Olsen
Source of the article: {Linkedin} on [2019-10-22]