RPA 2.0: What’s Next for Automation?

Most large companies will by now have some form of RPA embedded, and if they don’t, they will be looking at it (if you’re not, where have you been the past 5 years?). The market is still expanding quickly but some analysts predict saturation within the next 2-5 years. Most of us now know what RPA is, and what its capable of, so what’s next for Automation?

The term Hyperautomation was touted, meaning the integration of RPA with complementary technologies such as Optical Character Recognition, Chatbots, Natural Language Processing, and other solutions which, together, grew the cognitive capabilities to broaden the scope of RPA beyond strictly structured data. Depending on your definition of AI, it could be argued that these were the first ‘intelligent’ elements brought into the ecosystem, but it is definitely the case that AI and Machine Learning are being tentatively and selectively applied in automation albeit not yet living up to the tremendous hype surrounding them. A new all encompassing term, ‘Smart Process Automation’, seems to be emerging, meaning much the same. This direction of travel is set to continue, and we will see AI embedded more and more in the native vendor offerings – such as Automation Anywhere’s OCR offering – IQ Bot and Antwork’s ANTstein.

RPA is seeing a strong shift towards Cloud, with all the main vendors now offering some form of Cloud implementation. This is key to scalability, once the achilles heel of RPA. It also enables RPAaaS, whereby the supplier will host the platform and deliver bots to the client as part of a maintained service. This gives rise to new and flexible delivery methods which can take advantage of shared economies of scale to optimise value for money for the end client and avoid high infrastructure and maintenance costs. Expect to see this market expand considerably in the coming years.

The emergence of Microsoft’s Power Automate in the top Gartner quadrant this year belies a forthcoming shift in the balance of power in the market. The top three vendors had been quite complacent to milk the cash cow dry but the new upstart will rock the boat by offering significantly cheaper licencing models. At the moment it is limited in its reach due to a lack of skilled resources and the absence of an integrated hyperautomation ecosystem, but this will change very quickly as demand for the product grows. It will inevitably force prices down in the market and build pressure to innovate.

The focus on new Discovery tools, such as Process Mining, is the latest hot potato, with skills in high demand. The complexity of these tools gives them a niche value, but where used correctly they can add an element of transparency which would otherwise be lacking and inform end to end automation strategies to maximise RoI. These tools are likely to get easier to deploy as they develop more standard interfaces and reduce in cost.

Vendors will continue to drive the ‘citizen developer’ mantra and push the ‘bot for every worker’. At this stage this just isn’t realistic. The development UIs still demand a high level of technological understanding (despite the vendors’ assertions) and the cost is too high. Over time we may see a gradual shift in this direction, but it won’t happen overnight, unless Power Automate proves to be the necessary fulcrum to democratise automation.

All of the above really extrapolate trendlines which are quite established. I don’t need to dust off my crystal ball to make these predictions.

There is one longer term prediction which I think is happening under the radar which will bring about fundamental change to the market. As organisations expand the use of RPA throughout more and more processes, we start to suffer from ‘automation spaghetti’, and the sweet spot of RPA is being stretched to manage integrations and become a proxy for digital transformation. The perennial question of whether RPA is, or even should be, strategic, is brought to the fore.

The fundamental issue is that user journeys are constrained by applications, and more importantly, the gaps between the applications. The handovers are where the majority of dysfunctionality lies. RPA is ultimately touted as the band aid to fix the problem but this is way outside of its original raison d’etre.

We need to move to a more strategic user centric view, whereby we are not restricted by the applications themselves, but one whereby the journey is orchestrated by a workflow, which calls the applications and hyperautomation tools as required – using the right tool for the job, at the right time.

The market will experience a gradual convergence of Business Process Management tools such as Appian, ServiceNow and Pega, and the traditional RPA vendors. The BPM vendors are already building or buying niche RPA toolsets but these don’t compare with the full functionality sets of the top three. The first large vendor to offer a high quality BPM product alongside a full hyperautomation ecosystem will make great strides in the market and leave the others clamouring to recover their positions. Watch for some interesting mergers and acquisitions as this starts to emerge.