RPA is the gateway drug. AI is the drug...

October 10, 2018 | Phil Fersht

Anyone failing to escape the swirl of intense hype threatening to destroy everything great about RPA is probably thinking that these cute products are going to solve all their artificial intelligence needs and deliver them with a "digital workforce" that will go way beyond scraping screens, producing scripts and running unattended recorded process loops.

Now, don't get me wrong - I LOVE RPA... jeez, I bloody helped create the space when I first wrote about it in 2012.  I don't want to toot my own horn, but this space probably never have would have got off the ground if we hadn't been curious enough to get deep into it and articulate its value to the world.  And no one's paid me a billion dollars (well not yet, anyway).

Click to Enlarge

RPA creates a genuine experience, where the underlying fabric of decades-old processes can finally be altered

When we released the first "Intelligent Automation Continuum" in 2015,  we made it very clear that RPA was clearly the first step in a much broader roadmap to achieve beautifully-automated intelligence across your enterprise.  And today, this gateway philosophy has never been closer to reality.  RPA, when executed well, delivers a digitally-transformative experience to business operations executives, where they can - for the first time - fundamentally change how a process is designed to process data much, much faster.  Suddenly, firms have the chance to make fundamental changes to how they design workflows, instead of persisting with doing things the same old way, but with lower cost people and more efficient delivery models. Isn't that enough for now?  Why does the hype take it to a place where it's only going to disappoint?  If IBM's leadership already thinks these firms are massively overpriced, are there really others out there which will take the plunge?

When I see executives who previously stared at excel sheets all day (while beating up BPO providers for overcharging for insurance clerks in Delhi) actually getting trained to redesign workflows using scripts and GUIs, it warms the soul.  We are actually trying to do thing better... not just cheaper!  So why can't we be content with making this actually work before we get too carried away?

Time for a reality check:  RPA is firmly on the radar, but let's see it become properly industrialized and scaled before we get too carried away

The vast majority of these initiatives are project-based, not scaled - only 13% of RPA adopters are currently scaled up and industrialized, according to new data from 590 enterprises worldwide.  Most RPA adopters are still tinkering with projects and not rushing towards enterprise scale adoption:

Click to Enlarge

Suddenly, the whole RPA value proposition, which has carefully matured from the "Oh my God, a robot's going to take my job" to "OK, I get it now, RPA actually frees up time and fixes process breakages and staves off costly investments" has been injected with some serious hype-steroids, where suddenly these firms are worth billions of dollars, some are actually declaring they are going to deliver their own consulting services (really) and quickly move up the continuum to offer real cognitive and AI capabilities.  I'm sorry, but when were the RPA firms going to compete with Google and Microsoft? Am I missing something here? 

The Bottom-Line: Enjoy that RPA high a bit longer before you graduate onto the harder stuff...

The real data shows just how not-ready we are to declare some kind of robo-victory - executives must evaluate how all intelligent automation technologies can work together to take us to the promised land. RPA provides a terrific first stop for executives to make real underlying changes to their processes.  Once processes are digitized, there is so much more we can do with the data being produced, which is where other automation and AI tech comes into play, such as Machine Learning and predictive analytics and sophisticated cognitive computing.

Now it's always critical to focus on the "what next", and in the case of RPA the possibilities are limitless, but only when you have mastered how to digitize your underlying mess that has plagued your organization since before the days COBOL was the next big thing.  Then it's about how you reel in the analytics and AI possibilities that truly take your business to a new level of data heaven.  But let's get past the gateway first... let's not get ahead of reality and mess this one up, folks.

Posted in: Robotic Process AutomationIntelligent Automation

Never Miss A Story

Sign up for the HfS Research newsletter and get the best research delivered to your inbox weekly.




  1. Anshuman
    Posted Oct 10, 2018 06:04 PM | Permalink Reply

    Great insight Phil Fersht. RPA is the equivalent of 'assembly line' in manufacturing and 'AI' will be the automated work that the assembly lines enabled.

  2. Sachin Kumar
    Posted Oct 11, 2018 04:02 AM | Permalink Reply

    Good analogy...I would not go as far as Drug though!

  3. Phil Fersht
    Posted Oct 11, 2018 04:04 AM | Permalink Reply

    @Sachine: well... it's got a lot of people hooked ????

  4. Vasiliy
    Posted Oct 11, 2018 10:02 AM | Permalink Reply

    "where they (operations executives) can - for the first time - fundamentally change how a process is designed to process data much, much faster." We can see that for some vendors process change (reengineering) is anti-pattern of delivery as it adds additional risk - facilitate behavioral change within big companies with complex ecosystems and politics.

    "firms have the chance to make fundamental changes to how they design workflows, instead of persisting with doing things the same old way" Again they have chance but can easily select path of quick benefits realization of processes automation AS IS without any changes. At the end of the day they will get infra where changes are even harder now as in addition to old infra they need to support automation layer.

  5. Michael Kotowski
    Posted Oct 12, 2018 11:26 AM | Permalink Reply

    Great article for those that know. That said, at this critical juncture in RPA and AI growth for real business value, we still need to win over those with fear and uncertainty.

    Headlines and sound bites are what catch the eyes and minds of those not yet bought in.

    I’m in it - deep - drank the koolaid - understand and appreciate the “gateway drug” analogy. Im suggesting that may not be the favorable light with which to paint what we know works as we try to convince others.

    Positively framed, real life achievements in colorful, relatable storytelling is what we need more of.

    “Build it and they will come” (in our own business Field of Dreams)


  6. Phil Fersht
    Posted Oct 12, 2018 11:31 AM | Permalink Reply

    @Michael: Thanks Michael... what I'm seeing is RPA being the first experience many business line managers have had of using real tech tools to fix certain processes. In many cases, it take them a year or two to learn to write scripts etc., but once that has been accomplished, many have the awareness and confidence to learn other technologies, especially those with low-code dependencies, such as data ingestion and orchestration engines. Hence, it's a gateway drug that turns them into digital process management addicts =) I'm not in the business of mincing words and spinning pretty pictures and analogies, so apologies if this offends anyone,


Post a Comment

Your email is never published nor shared.