The GSA’s Intelligent Automation Special Interests Group (IA SIG) is a steering group for automation and artificial intelligence featuring some of the leading minds in the field. We got together with IA SIG members Terry Walby, CEO of Thoughtonomy, and prominent advisor and strategist Andrew Burgess, to discuss the work being done by the IA SIG and why this is a critical time for organisations looking to capitalise on these game-changing technologies…
GSA: So, let’s start with a look at the IA SIG: what’s the purpose behind this body?
Terry Walby: Well, firstly it’s about recognising that intelligent automation as a technology is having, and will continue to have, a significant impact on the world of work, both internally for organisations and in terms of the way they work with their suppliers and outsourced service providers. We must also recognise that the landscape is going to change quite significantly as the adoption of this technology matures, both for end users directly adopting and for providers changing the way they deliver services. For me the intent of the SIG is to help inform and educate and share experiences and knowledge across the community, to help them make better decisions and learn from each other as the market matures. That will include buyers of outsourcing services and technology; it will include providers of services; and it will include the ecosystem of consultancies and advisories that form part of the GSA membership.
Andrew Burgess: Hopefully we’ll also be a central point of contact to other sources of information, other communities within GSA – for example we’ve been talking about the APPG, the CPO Council – and externally as well; we’re looking at how these distinct groups can be brought together with a focus on intelligent automation.
GSA: Shall we outline a definition of “intelligent automation”? Because there’s a broad spectrum of opinion about where individual terms – RPA, AI, cognitive – blur into one another...
AB: There are indeed a few definitions around and it often seems as though it’s “whatever you want to make it”. One definition looks at the unification of two technological concepts that have been around for a long time – AI and RPA – and that’s probably the simplest definition.
TW: I’d say that what we’re really focussed on is the use or application or procurement of technology to deliver an alternative to people-based services. That’s obviously quite broad – but I think that’s why it’s of interest to the GSA’s community, wheras other software spending – such as buying a new ERP system, or CRM tool, wouldn’t necessarily be something the GSA got involved with, because that would be bought as a technology product to deliver an outcome for a technology user. Intelligent automation, as in the definition Andrew has just given, is about replacing people-based work with technology. It’s procured and delivered and implemented, generally, not as a technology for technology’s sake but as an alternative or an improvement to people-based work. That’s why it’s applicable to this community.
AB: I think you can almost imagine a Venn diagram that has RPA on one side and AI on the other, and then IA sits in the middle as the overlap between those two technologies. So it’s not just RPA, and it’s not just AI – because AI does a lot of other things as well as replacing people-based services – but it’s the combination of the two, using AI to enhance and augment the RPA.
GSA: Andrew you’ve been looking at this space for a very long time. Where do you see the biggest transformation coming in terms of individual sectors?
AB: One of the things I might be guilty of overgeneralising on, perhaps, but I think it’s valid, is: when you go into an organisation looking for RPA opportunities, you look where the people are; and when you’re looking for AI opportunities you look where the data is. So, if you’re looking for the overlap, you’d look at organisations which have lots of people and lots of data. That would be the sweet spot. So, for example, utilities: there are lots of people processing meter readings and exceptions, but there are also a huge amount of data from those meter readings.
TW: That’s a really interesting point you’ve just raised, Andrew, about “look for the people; look for the data.” To add to that, what we’ve seen – and this is around that point of the convergence of the two technologies – is that while clearly that people-and-data confluence is the sweet spot, the people who’ve implemented RPA have actually created the ability to generate data through the automation of a process, and then, therefore, the ability to use that data to make intelligent decisions using AI. Previously, that data may not have existed because the process was being carried out by humans and therefore the data was not being created. Once you automate a process it’s very easy to create data out of the automation of that process, which you can feed to an intelligent algorithm to make AI-based decisions. That’s actually how the orchestration of our platform works; it’s learning from what it’s doing, generating data it can then mine to make decisions on how to be better.
GSA: So to what extent is RPA a logical precursor to AI? It can’t be in every instance because AI has so many different applications.
TW: Yes, numerous appplications, and Andrew’s absolutely right: AI is much broader than just robotics or replacing people-based work. AI exists in your phone, on your PC, on the Alexa in the corner of your office.
AB: Going back to my Venn diagram, this is where AI compliments RPA; that’s the intelligent automation piece, I think. You can for example use AI to do predictive analytics, but I wouldn’t put that in the IA category. If you’re using AI to make unstructured data structured, or to help make decisions in a process – all the things that robots can’t do – that definitely does come under the banner of IA. This may be too strong a statement but, bearing that in mind, it’s almost like IA has to have RPA as the foundation for it to be defined that way.
GSA: Looking at the general perception of this technology: we’ve seen some research published recently saying that around 85% of enterprises have now implemented AI to one extent or another. Are we at a more mature stage in the market than many people think we are?
AB: I think it’s exactly the opposite, to be honest. I think 85% is a wild overestimation. Many people don’t even know what AI stands for, let alone implementing it.
TW: I was going to say exactly the same thing. I think that’s a complete misrepresentation of the market. I guess if you go to a customer base that’s bought your intelligent automation product from you, and ask if they know about it, the answer’s probably going to be “yes”…. But if you actually go and talk to the real world and you’ll find that the great majority of the market is very uninformed, about even the fact that this technology exists, let alone what it’s about and how it can be applied.
AB: Most people’s understanding of AI at the moment is centred upon consumer devices - Siri, Alexa etc – and very few have much of a concept about how it can be practically used in business. So in terms of awareness and usage I really do think the reality is probably the inverse of what that research says. There are, of course, great examples of where AI is being used and where people don’t realise how intelligent the software is: there’s an old joke about how AI is always 20 years in the future because as soon as it becomes usable on people’s desks or in their pockets it loses the “magic” and it’s no longer AI…
GSA: With that in mind, when you’re both talking to people who are at the stage of, at least, contemplating an IA solution, what are the key questions that potential buyers are asking right now?
TW: If you get into a situation where people have contemplated and perhaps tested or evaluated, it’s really about the application of that as a concept in their business environment, specific to their domain and expertise. One of the things we have observed is that this is a general-purpose technology, as an augmentation or a replacement for people-based work, and that’s pretty broad in scope: potential buyers need to have that contextualised so they can understand what that means for them specifically and the operations they run. What are the things they are doing which could be augmented or replaced? What are their service providers doing which could be done better through the application of this technology? Who else in their space is doing it? That contextual relevance is hugely important – and I think that’s true of any general-purpose product, isn’t it? Make it real for them by showing them, or telling them, or putting them in touch with someone who’s done something similar in their world.
AB: The uniqueness of AI, because it’s so hyped, is that I get people coming to me saying, “I’ve heard all about this AI: how can I put it into my business?” – and clearly that’s the wrong question to ask. They should be asking, “What are my business objectives, what are my challenges, and can AI – or IA – meet those challenges?” That’s the direction from which you should be approaching things. Talking about capabilities: particularly with IA, it’s around end-to-end processes, and if you’re just looking at RPA you are limiting the scope and benefits you’re going to get from your program – because you can’t handle unstructured data, complex decisions and things like that. Using IA means you’re opening up much greater possibilities for end-to-end automation across the whole business.
GSA: One of the elements of our forthcoming event agenda for the IA SIG is a workshop looking at how to build a business case for intelligent automation. We’re talking about technology that’s being used in a huge number of different ways, for different organisations in different sectors. What are the commonalities that can be connected in a workshop like that?
TW: Businesses spend money on things because either they have a net reduction in their operating costs or they allow those businesses to grow in ways, or offer solutions and services, that they couldn’t otherwise. Or, they reduce operating risk and exposure to something that might damage the business. Those three reasons are pretty much what drives any investment decision for any business of any size, in any sector. If that’s the case, then in order to support someone evaluating and making an investment you need to be able to illustrate either where they’re going to reduce cost, where and how they’re going to remove risk, or where and how they’re going to create growth without a linear increase in operating costs. Ultimately at the end of the process someone needs to sign a cheque for something and in order to do that they need a justification – and that’s where building a business case comes in. That’s why we think it’s an important thing for the community to get a better understanding of how to do that.
AB: The key thing goes back to what I just said about aligning with their individual strategies: what are they trying to achieve? To ensure that what IA can deliver is aligned with that. If their strategy is to reduce headcount, then what can IA do for that? If it’s around increasing C-Sat scores, what can IA do for that? Start with the strategy, and then try to break it down into the different elements that IA can impact within those broader objectives.
GSA: A few years ago when cloud really came to the fore, a lot of buyers began to think of it as a magic bullet which could solve all their problems, without really knowing perhaps that much about the technology. Are you encountering that kind of attitude now in the market for IA?
AB: Yes, I think so. Not naming names but certainly I think some of the software vendors are saying that this stuff is really easy to do, and the benefits are huge – and when it comes down to actually doing it the buyers are finding that, “hang on, we thought this would be much easier than this; we’re finding a lot of niggles and we’re having to deal with a lot of IT issues…”. If expectations are managed better in the beginning, then maybe those problems wouldn’t come about and users could plan programs properly.
TW: It’s actually a good analogy to the cloud. In any hot market where there’s a buzz and a lot of growth and a lot of interest, then the vendors fuel that growth by either reimagining what they do under the banner of whatever the buzz technology might be – cloud, historically, and maybe RPA now – and that creates a lot of complexity around how different solutions relate to each other and what they can achieve, and with each vendor claiming their solution is the most straightforward and effective. Then, on the other side they have the emergence of advisors and experts from organisations who last week were advising expertly on something else… Those tend to give the opposite view, which is that the new technology is really tricky and problematic and what you need is to pay the advisors consulting fees to advise you what to do otherwise everything’s going to be terrible… Everybody is trying to serve their own needs by making the technology sound one way or another.
AB: And I think this technology lies somewhere between those two extremes: it’s not horrendously difficult to the extent that you need a team of advisors for months, but nor is it the case that you can just knock up a pilot scheme in a day.
TW: Well, perhaps you can, but if you do you’re not addressing a strategic objective. I’d summarise it like this: because conceptually what we’re talking about here is a very easy-to-implement technology that can deliver rapid business value, and because you can deploy it frictionlessly, it’s easy to do it tactically. But that’s the wrong thing to do if you want strategic benefits. If you really want to adopt it as a strategic asset to your business – which is where it has real value – that’s not something to do tactically. What we’re seeing is that people who’ve taken that first approach – “let’s just download some stuff and get going” – have driven into cul-de-sacs they’re having to reverse out of. They’ve created a poor and negative impression of the value of the technology, and it’s now a case of having to rework all the stuff they’ve done in order to do it properly.
GSA: So assuming they’ve not gone down one of those cul-de-sacs, what are the next-stage challenges for people who’ve gone along the IA journey?
TW: People tend to come from this saying, “I think we can use this technology; I’ve been introduced to the concept and we’d like to work out how we can get value from it” – and they approach it as a journey. Start small; deliver some value; and then scale quickly into something that’s strategic – but do so in a way that the “start small” is not a tactical play, but something that’s enterprise-grade that can scale. Or, they’re coming from the direction that says, “my outsourcing provider is inefficient and I don’t want to be paying millions of pounds for a people-based service: I want a strategically different service that’s delivered via automation.” For the first example, the importance is making sure the start point has the right strategic drivers and involves the right architecture, the right implementation, the right landscape, the right stakeholder management across the business; and that the scale-up happens, because as it’s easy to deploy tactically and as a quick fix it’s easy to forget about it and lose the real value.
AB: People need to understand their ambition before they start. It’s fine to say they just want to tweak some processes and get some savings, or at the other end of the scale to say that they want to fundamentally transform the way they do business through automation. If you don’t understand where you are at the start, though, it’s easy to make the wrong move. Your first step needs to be different depending on the level of ambition you have, and it’s crucial to understand that from the beginning. Understanding the journey, the roadmap, is really important. So once people have done a pilot, or proof of concept, they should really know already what their particular next steps are because they’ll already have mapped them out.
GSA: Is it actually possible to have a firm understanding of the end point of that journey when we don’t really yet know what the end point of the technology is?
AB: You have to base it all on today’s technology – extrapolate a bit, sure, but base it on what’s available today. For instance there’s new technology coming in which uses AI to automatically map processes, which didn’t exist before – and that’s going to have a big impact on consultancy fees, incidentally! – and this will of course fundamentally benefit buyers, so from that kind of perspective things can only get better. Unless and until we get to a singularity, of course…
Andrew Burgess is running an exclusive GSA Masterclass on April 23rd in London; for more information see the event listing here.