Intelligent automation will play a significant role as enterprises look to lower operating costs and build greater business resilience, both to weather pandemic-driven disruption and improve operations. Significant value can be added back into the workforce to the tune of $134 billion in 2022 alone.
In this live webinar, presented by HelpSystems and featuring Forrester Research, we discuss trends in the automation space, using RPA bots to increase business resiliency, and where the bright spots are for automation. Forrester Vice President and Principal Analyst, Craig Le Clair, joins us as a guest speaker to lend his expertise and help clarify expectations for businesses adopting new intelligent automation solutions.
In this webinar, we’ll explore:
- Definition, scope, and progression of intelligent automation and how RPA underpins IA efforts
- Digital assistants and RPA bots in a post-pandemic world: resilience and cost efficiency automation
- Bright spots: where to find the best IA opportunities
A complete transcript of the webinar is below.
Pat Cameron: Welcome, everyone, to today's webinar on Intelligent Automation and RPA, Driving Business Resilience in the New Normal. My name is Pat Cameron and I will be one of the presenters for today's webinar. I have about 25 years of IT experience and various flavors of automation. And I've been with HelpSystems for a little over 20 years. I'm joined today by Craig Le Clair from Forrester Research.
Hello, Craig, and thank you for joining us today.
Craig Le Clair: My pleasure.
Pat Cameron: First, a little bit about HelpSystems and who we are. At HelpSystems, we understand that building a better IT is not a destination, it's an ongoing commitment. According to the [inaudible 00:00:51] IT, complexity is the top risk for managing an IT group. MuleSoft also notes that the average corporation, think about this, has over 900 allocations that are being used. We're building a product portfolio that can help IT organizations get the critical functionality that they need while reducing the complexity that they face.
We focus on mid market in particular for three reasons. They need pricing that makes their use cases attractive. They need software that's accessible to those in the front lines that need to deploy the automation. And they need software that's not dependent on having lots of additional software from different vendors in order to operate it. Here at HelpSystems, we focused on two broad categories, security and automation. To bring value to our customers on the security side, we provide a broad portfolio of solutions that include data security such as our GoAnywhere secure file transfer software, and our Clearswift data loss prevention product line.
Our core security identity governance and administration product offers an integrated identity and access management suite that includes tools to manage your bot profiles. So those profiles that need to execute for your automation can also be managed and must also be managed. And finally, we have our infrastructure protection products that secure networks and hardware, etc. On the automation side, we also have a broad suite including infrastructure automation such as Vityl capacity management, workload automation for back office and batch automation, and RPA where we're going to spend our time and focus today.
HelpSystems Automate currently has about 3,000 customers worldwide, and our customer base is pretty much vertical agnostic. We're in all kinds of industries. But recently we've really been focusing with our customers on banking, finance, and healthcare customers. They need fast RPA, and they need quick ROI on their investment. Automate provides powerful RPA bots with no restrictions. We're simplifying the bots available to encourage bot usage wherever it makes sense and for whomever it makes sense.
Bots can deploy as attended or unattended, and can run multiple task executions concurrently with no throughput restrictions. This allows users the flexibility to create bots and workflows that work best for them without being constrained on the type of bot execution that they need. Automate is built for the citizen developer, kind of like me. I don't have a development background. I have an operations background. So without jeopardizing power, our bots are simple to use and no coding is required.
Both technical developers and citizen developers can start creating bots immediately by leveraging our 600 plus prebuilt automation actions. Our drag and drop functionality is unique because we don't require those citizen developers to understand coding concepts. They just need to understand the business concept. If they understand the process, they can build a bot to automate it. We eliminate the number one challenge that organizations are facing when scaling with RPA.
We enable users to get started with RPA in a small scale, one bot, and to scale up and grow incrementally to an unlimited offering of bots as well. That unlocks the possibility of what and where those citizen developers can afford to automate. With subscription or perpetual licensing models, we offer further flexibility in where and when organizations are ready to grow and scale up.
So, in summary, HelpSystems is financially stable. We have a mature RPA product, and we help organizations maximize the value of their current applications. You don't have to change your current applications through RPA. Customers are looking for providers that can solve more than one problem. And through our broad security and automation portfolio, we can offer a complimentary suite of solutions, backed by our strong financial footprint and proven leadership through the good times and the bad.
A few of the use cases for HelpSystems Automate. Over on the left for shared services, traditional RPA solutions for manual and interactive processes. And IT also can automate a lot of those user provisioning and cloud management processes. Automate is used by the extended enterprise. Lots of data transfer, file transfers are all managed through Automate. Automate can also monitor inboxes and email, and then they can use the input from that email to trigger a complex process.
Automate is a great integrator of other applications as well. Its workflows can orchestrate in the BPA process as well as RPA. It can get the data from wherever it's sourced, either a webscraping maybe or a SQL database, and then send it to whatever the target is, maybe data entry into an application or use our API to update your cloud applications such as Salesforce or your other ERP system. At the end of today's webinar, you can download our RPA roadmap guide and that will help you to plan for a successful RPA strategy.
An introduction for Craig. Craig Le Clair serves enterprise architecture and business process professionals. He's an internationally recognized expert in automation, AI, and the future of work. His technology coverage areas include robotic process automation, AI solutions, and financial services, and the potential workforce distribution, excuse me, disruption due to these technologies. In his 2019 book, Invisible Robots in the Quiet of the Night, has been met with wide acclamation. A prolific writer and speaker, Craig is frequently quoted in the Wall Street Journal, the New York Times, USA Today, Forbes, and many other applications and media outlets.
Dynamic case management, electronic signature, and customer communications management round out his coverage. Craig brings a wealth of experience and knowledge to Forrester through his 25 years of experience in IT and business process transformation. Prior to joining Forrester, he was a vice president at ADP focusing on next generation strategies for investor communications. He holds a patent for electronic delivery of compliance information. Craig is also a co-founder of docHarbor, an early SaaS provider. He has both a strong business and technology background, including time at Mitre and BBN.
Craig Le Clair: Thank you.
Pat Cameron: Thank you for joining us today.
Craig Le Clair: If we do this again, I'll have to shorten that intro.
Pat Cameron: Lots of good stuff in it, though.
Craig Le Clair: Yeah, it does. Well, thank you. So am I the presenter now or I have to...
Pat Cameron: You should be able to click on the-
Craig Le Clair:... do a formal application here?
Pat Cameron: I handed over the mouse.
Craig Le Clair: See, how do we do that? Yeah. So thanks for joining. I know everyone's probably a little Zoomed out and virtualed out, and here we are, on my first slide. I appreciate the virtual time that you're allotting to this, and I'll try to be brief, and crisp, and impactful. And here's what we'll talk about, really a lot of discussion about intelligent automation. It's variously called intelligent process automation and other things. But the Forrester's term, which we think is generally accepted, is intelligent automation. I'll talk about what that is and the progression.
And then talk specifically about the role of RPA, and the bots, and the digital assistants, and all of this sort of emerging technology in a post-pandemic world. And then take a look at where we can find some real opportunities going forward. So I am not able to change the slide. Do I need to do something on my side? Here it is. No question, our automation psychology and our roadmap, roadmap meaning as a company, as an enterprise, as a government entity, you had a project roadmap before all this happened. And that's gotten jolted. It's gotten changed. If not right away, it will be by the impending aftermath of the pandemic, which we expect to be a significant recession which will shift gears, and I'll talk about how.
Importantly, our psychology has changed as well. And I'll talk about what that means. Discretionary spending for all projects. This is typical of the past recessions. We've tracked the last three, in particular. For a multiple quarters, as we climb out from the negative gross national product on a quarterly basis, that might be multiple quarters. Might be two to four quarters. As we crawl out of that, investment, non-discretionary investments spending by companies will start to increase. But for a period of time, look for any discretionary spending to be on hold.
That has implications for companies in terms of their maturity and progression, in terms of the automation we're talking about. The AI and more transformational projects, they'll be pushed off the table a bit. Now, it's an opportunity for companies to double down and really focus and accelerate those. But the reality is when a company is faced with terminating people, putting them on furlough, when it's faced with a really tough economic climate, the tendency is to push those out. But we think that after this period of pause, that there will be rapid acceleration in the technologies we're talking about, which is why we're talking about them.
The priority will shift to pragmatic automations. And in fact, intelligent automation, as we defined it, the core is RPA task automation. But it's really the combination of RPA with more intelligent automation such as chatbots, and machine learning, and text analytics, or natural language processing applied to documents and unstructured content. These are the immediate intelligent components that are being linked with RPA. When you do that and also combine it with existing automations that have been providing value for years, and I'll talk about those a little bit, what you get is this notion of intelligent automation. So pragmatic is at the core of it, and there'll be a shift to that.
The goals will be lowering costs, and supporting remote execution of business. Now, that has two aspects. One is customers having to complete business remotely, which given the still high number of analog and paper-based processes that are out there, that's not there for many, many companies. So if you can't do your business remotely, you can't do a mobile deposit in banking, that will be accelerated because you need to do that. And then resilience. We've depended on this highly structured global supply chain. And I think what we saw in the pandemic was that that doesn't allow the agility and the diversity and business continuity that's required.
So that's one aspect of resilience, is to create more agility and diversity in the supply chain. Another one is just being able to convert from an in-office type of work or in-plant type of work to remote work, another form of resilience. And then general business continuity. This whole notion of systemic risk is really going to be one of the remaining legacies of this time period that when you look at risk management in most companies, they're focused on harassment. If you're a college or university. You're focused on binge drinking in the dorm.
You're focused on risk from competitors. You're not focused on these broader systemic risks. And I think that's what this pandemic is really waking companies up to, that this is a pandemic, this is a global risk, this is a systemic risk. And there might be another one. It might be climate change. It might be something else. Essentially, risk management needs digital transformation. It needs to look more broadly at the potential risks and to prepare for them. And a lot of that is various forms of resilience, as we'll talk about.
What you end up with is, of course, two by twos are very popular to kind of position things in an understandable way. And this one shows an acceleration zone. What this means is that these are the areas, and these are general areas, that companies will be repositioning their roadmap for. So if you have a project and it doesn't do one of the four things in the upper right, it's probably not going to get funded. We think that RPA task automation is going to be very, very significant because there's a lot of very rapid automations that can be done.
I talked to an airline recently and they just were flooded with cancellation requests. What do they do? They built a bot. They had a partner build a RPA bot that was handling the six or seven-minute tasks required to check on the codes. And it was a human task to really see whether it should be a travel bank, whether it should be a refund, what are all the rules that they have for changing a ticket, cancellation ticket, and so forth. The bots are doing that task now for 80% of the volume. They had to do that because their volume had gone from 2,000 a day on average to 18,000 a day, 20,000 a day.
Panic's too strong a word, but this sort of very urgent type of automation really lends itself to the kind of deployment that RPA can do because it doesn't require that API integration, doesn't require any change or modernization of the core applications. It doesn't require a UI to be built because it's essentially just mimicking the keystrokes and data entry elements, the keyboard elements of a human. It's sort of its great strength, it's ability to do that.
Now, there are vulnerabilities that come with that, but that's really why it becomes in the acceleration zone. Text analytics is just getting much, much stronger. We've been working in the capture world for decades with OCR and being able to extract fields from forms, and paper, and documents, and emails, and so forth. But now machine learning has advanced to the point where you can raise the quality of the extraction and create very interesting patterns of data XML files that now you can analyze for fraud, you can analyze for sentiment, for plausibility, as it relates maybe to an insurance claim.
It really unleashes the modern and emerging text analytics. It really opens up a lot of value extraction from data that we haven't had before. So companies are looking at that and saying, "You know what? That's something I can really support. Strong ROI, let's move there." The other aspect is remote business. And again, I mentioned that being able to remotely conduct business if you can't do it today becomes a critical item. Other areas, again, pushed to the side, losing momentum, the more transformational projects.
The resilience zone is worth mentioning. You have to have the right cloud capacity. You have to be able to support remote workers that will be more mobile than before. And that's one thing for knowledge workers, as we'll see different personas in the workplace. But it's another thing for workers that are really used to coming into a location, that had different levels of management oversight. It's a big deal. You need new technology to be able to do that. It's not just, "Here's your VPN connection." They may not have a laptop. They're used to face-to-face interaction with other humans to manage their production, which isn't going to happen.
Employee customer facing chatbots, these are all elements that allow you business continuity. If you can't have your workers come into a contact center, you want to have at least a virtual agent to be able to interact with them. This is something that really surfaced during the initial part of the pandemic where companies were trying to cut over to these virtual agents that were going only two questions deep and basically not able to complete a transaction. They really were more like a frequently asked questions.
And so, the level of investment and improvement in those is going to be an accelerated area for resilience. Not for moving to more customer self service and reducing head count, as much as this is something we need for business continuity. Yeah. So anyway, that's kind of our redrawn roadmap with all of this. We're going to go to a polling question now.
Pat Cameron: Let's do that. All right.
Craig Le Clair: Yeah. Pat, do you want me to ask the question or do you want to just have people look at it?
Pat Cameron: Sure, you can the question.
Craig Le Clair: No, why don't you do it. Why don't you do it?
Pat Cameron: With the current economic climate and pandemic, where does your organization stand on RPA? Nice to get a little bit of feedback back. RPA is still a priority, are you shifting to a more pragmatic approach, is it still a focus but delaying plans, and have you canceled plans to implement RPA? It sounds like from what you just said, Craig, I don't think a lot of people will be canceling plans to implement RPA. It sounds like it would be more of a priority.
Craig Le Clair: Yeah, I wouldn't think so. I wouldn't think so.
Pat Cameron: Exactly. All right. We'll give people just another second to answer, and share those results. So 69%. All right. RPA still remains a priority. 15% are shifting to a more pragmatic approach, and 15% is still a focus, but some delays in planning, and that's certainly understandable. Well, thank you for that feedback.
Craig Le Clair: Yeah. No, that's great. Thank you very much for that. And now we can go back to the regular scheduled program.
Pat Cameron: There you go.
Craig Le Clair: Which I'm not seeing.
Pat Cameron: All right. Hang on just a second.
Craig Le Clair: Yeah. Okay. I've seen the screen now, but I'm not seeing the... This is a build that's actually not building, but it's-
Pat Cameron: It'll build?
Craig Le Clair: There's the build. So, basically what is the path or progression with intelligent automation? Question being, where did it start and where's it going to end up? And that's what this build is representing. In stage one, we're in this rule of five world. That's just rule of thumb that Forrester came up with a couple of years ago that just says, "Here's where you should apply RPA." RPA doesn't do everything. It doesn't do end-to-end orchestration. It really focuses on very deterministic, very limited rules, structured data only. It doesn't have any learning capability.
This is the early phase of the market, and this is where there's a lot of opportunity. I mean, companies have these 18-minute tasks that are being done routinely, maybe hundreds of hundreds of times a day that you can build a bot for and you can pay for the bot very easily. And that's really what's driven the first part of the market. And I'll explain a little bit more about rule five in a second. But basically, there are two flavors of that, attended automation, which tends to be in a customer service context where the humans interacting with the bot, maybe triggering the bot to start work. Maybe the bot is asking the human for a account number and going off and getting data and presenting it to the human. So there's a little bit of interaction there.
Unattended tends to be batch work for back office applications. So it's really an automation that's going off and attacking a virtual desktop and doing a very repetitive task. But this is sort of where we've been with stage one rule of five. The issue is if you go to the next phase is that there's a limited number of those types of 18-minute, or 12-minute, or four-minute task automations that are out there. And then, so companies will look for, where can I go with this RPA platform that I've invested in? How do I make it more transformative? How do I connect it more to line of business applications.
That's where the stage two comes in. Here you're putting a layer of analytics on top of that task automation, and you're saying, "You know what? I'm going to go into unstructured content like invoice management, email management, document handling. And I'm going to take natural language processing. Maybe I'm going to take some multimodal AI to understand the relationship of objects on a particular document to get a better level of extraction accuracy. I'm going to combine all those, but I'm going to use these analytics to get me a tagged file that a RPA bot can then do and tell things with."
All right, so this is a level of analytics that's applied on top of the RPA bot, that really provides a different set of instructions. Maybe there's some analysis of the more rich set of data that's now being extracted that allows you to understand if there's a fraud. Is this plausible claim in insurance? Is this an inappropriate invoice to pay? These kinds of things can be done. So that's stage two. It's intelligent extraction, we're calling it. Intelligent instruction plus RPA.
Stage three is sort of the ultimate intelligence of the intelligent automation. And this is where you're using the central management capability. All of these RPA platforms have a central repository where the automations hang out. And they have an event management layer that allows you to enlist other automation building blocks so you can apply them collectively to a use case. And this is one of the talents of the platforms in that, because the automations that had to be built, which started out being essentially just smart macros on desktops, but because they were a lot of them, you needed some central way to manage them. You didn't want to house them on your desktop, so you moved them to a central repository.
Well, that just evolves to have more of an orchestration capability, to build APIs into that central environment that would allow a machine learning decision to be made in an algorithm. It might be in Azure or might be in Amazon, to have that trigger the dispatch of an automation from the central repository. So this sort of RPA as automation orchestrator is the stage three, and there you're linking with chatbots for conversational intelligence. So the chatbot's talking to the customer, virtual agents talking to the customer, and the customer wants something done. Well, the virtual agent sends a event request to the orchestrator for the RPA tool, and that will enlist RPA bots to get that done on the customer's behalf.
This type of conversational intelligence linking machine learning for decision management, the next best thing to do, these are all logical progressions of RPA and RPAs development. It's really what we mean when we talk about intelligent automation. That's what we mean. So I'll go to the next slide, and I want to be respectful of your time. This is a complete definition of intelligent automation. So what I gave you was more the progression and kind of the spirit of it in terms of its evolution.
On the previous slide, this is what we call a Forrester Tech Tide, where we list all the different software markets that comprise a domain, in this case, the intelligent automation domain. So you see, on the left, the invest category. These are ones that today provide high business value. And you see the text analytics is there, that phase two. You see that both versions of RPA are there, attended mode and unattended mode. You see the process, mining process discovery are there because they can be applied to date just to help standardize the process, to help understand the process better.
The experiment category are just more nascent technologies that have a low maturity and low business value as shown in the two by two. Now these are the types of projects that I'm saying are getting moved off the table, unfortunately. But you were experimenting with them anyway, and these are domain-specific robots. Automation orchestration, I talked about that a little bit in the phase three. AI-based exception management. This is applying machine learning type dynamic decision making to things that a bot could handle. So they go into a more sophisticated environment that's absorbing data on the context, and being able to apply predictive analytics to determine what to do next. So these are the experiment categories.
I won't go through all of them, but it's interesting that, at least in our model, we wanted to just not include the new emerging technologies. But there's a lot of automation we've been doing for a long time. If you look at HelpSystems' total portfolio, they've been doing things like workload automation in some of these maintain areas for quite a while. So these are all providing value. They should be maintained. They need to be maintained. This is a guide to how you should be thinking about different technologies.
Yeah. Looking at the next slide, we're going to move into digital assistance and RPA bots in a post pandemic world, to sort of drill down into that somewhat. As we saw in the survey and the poll that we just did, RPA remains popular. It just seems like people want to buy bots. And so, this just gives you some data based on a pretty broad survey of 3,000 data and analytic decision makers that Forrester conducted. We'll go to the second... Pam, you want to fire off this polling question?
Pat Cameron: All right. Yep. What is your organization's position on adopting RPA? Interested but not yet started, implemented RPA or currently implementing, expanding or scaling RPA, or not interested? So from the previous question, sounds like our group is rather interested in RPA, and we'll see what that level of interest is. (silence) All right, well, let's get people checking the boxes here. Give them just a second. (silence) All right. So we have 28% that are interested but not yet started. This is a good place to be then. 39% that have implemented or are currently implementing, and 33% expanding. So we've got lots of RPA users out there.
Craig Le Clair: Great.
Pat Cameron: Thank you very much.
Craig Le Clair: Yeah, yeah. Let's get to the next slide.
Pat Cameron: All righty.
Craig Le Clair: Yeah. So I kind of went over this a little bit, that this is what the rule of five is for task automation, where you can apply RPA. Because we were seeing a lot of companies, enterprises, governments trying to apply RPA really where it wasn't a good fit. And so, you have to find repeating patterns, and they tend to be in relatively small number of steps, or actions, or clicks. So less than 500 clicks. This is just to put this in real terms. You might have 100 people that are working in the company. They're downloading emails 30 or 40 times a day, and they're opening up the emails, and then they're looking at a particular set of fields in the email and they're loading those fields into another application. That might occur 1100 times a week.
So basically you want to find these relatively small tasks that a bot can do. They have to be pretty structured, pretty standardized within your office. The great talent of RPA as I talked about, and the reason it's going to be accelerated during the pandemic, is that you don't need to change the core applications. The UI stays the same. You're just replacing human activity with a bot against those applications. But it's also a vulnerability of the technology in that applications change all the time. I think we heard Pam was saying 900 applications was kind of average for a company to have.
Well, those applications all are going through their own SDLC, and their own maintenance, and their own upgrades, and they change. And many of them break bots. In fact, the maintenance of bots is an issue in the industry. So there's lots of investment in R and D going on by the primary platforms to address that with surface automation, and machine learning, and other types of analytics to be able to have bots repair themselves when applications change. We're not there yet. So you want to keep the number of applications that the bots operating against to a relatively small number, and five is just a...
These are not cast in stone. I mean, you can have nine applications, and particularly for scraping websites, you can have... It's a way to think about the technology and what it was designed for. It wasn't designed for a lot of decisions. There's no decision management built into these design environments that you get with an RPA product, like you would have in a business rules engine, like you would have in some of the more advanced business process management products. So if you keep the number of decisions relatively modest, the number of applications modest, and look for those repeating clicks, you'll have success in RPA. I've seen it over and over again where companies don't, then they ran into maintenance problems and so forth.
This is really where multi-skilled digital workers can be shared across the business service. If you look at where RPA is really being invested in now, it's really... Except for that 15% line of business, these are all shared services. They're departments that every company has. Mostly. Some have really big contact centers, but other don't. But everyone's got HR. Everyone's got finance and accounting. There are a lot of rule of five tasks to be automated in these areas. This is where the industry really got its start.
Now, when you have this sort of movement from remote work, office space work to remote work, back and forth, trying to figure this out, it's nice to have bots that can augment and support and be a source of resilience and a source of sharing work across this sort of dynamic environment that we're all trying to figure out. So that's a good aspect of how you can use the RPA intelligence to really augment work. This is a finance and accounting area. I think any real CFO will cringe at this summary, but kind of gets to the major spots.
The dark areas are where RPA has shown the most value in terms of FTE offset. We'll talk about that offset at the end of this talk, but really reconciliations, consolidations, monthly and quarterly close. A bank I have spoken with several times had 1,500 people that were involved in quarterly and monthly closes, and they felt that probably 30%, 30, 40%, a bot could augment that staffing and provide resources to do other things, right? So you can imagine that as we move into this recession, which is going to be, by most...
I'm not an economist. I was by training, but I'm not now an economist. But most 80% or so, are projecting a level of recession at least starting off to be 1930-ish like. And you can imagine in that environment that any of these types of tasks that can drive costs out of an organization are going to be looked at very, very different, very, very hard. And just doing the remote business, we'll need many technologies. Here you have intelligent capture. We've talked about that.
It's a combination of the OCR text analytics we talked about that is that phase two RPA. But it's also e-signature, e-forms, e-notary. This is where it's really showing value. I've had more calls about e-signature and e-notarization in the... Particularly e-notaries. I get e-signature calls all the time. But e-notary, it's been four or five years since I've talked about that. Then now all of a sudden now, real estate entities, brokerage companies, advisor companies that need face-to-face notarization now need a solution. It's just all of a sudden.
So these are, again, ability to do remote business, new forms of collaborations that are domain-specific. So I want a mortgage and I need to talk with a mortgage specialist. Well, it's not the one trying to sell you the mortgage, it's somebody else that really understands the process. Let's have a work domain or a collaboration session, a domain workspace that enables that. We're going to need to build these things. And we're going to need the process automation to do it. We're going to need digital process automation, DPA wide and DPA deep.
They're on the chart, but DPA wide is kind of low code process capability for citizens to develop process. So it's very wide in the organization. Whereas CPA deep is more what you'd build if you were redoing the claims management process or your customer service environment. It's deep expertise needed in the platform that is actually building that process. So basically remote business is going to get accelerated and here are the kind of components of it. Now this is an example of that phase two, and to the extraction classification and routing.
Look at this insurer here being able to extract information, but not just in the classic OCR way, maybe with some validation for fields on what are extracted, but using machine learning and these evolving technologies to computer vision to be able to really get a much more reliable and richer set of data. And with that, you can classify and do analytics on, as I've talked about. This is a real-life example of that, occurring with routing of emails. It's a lot of talk about human-in-the-loop. What we've tried to do with this framework, this is a variation on what our defense department came up with as human-in-the-loop framework.
What this means is human-in-the-loop with machines. So as we get more intelligent machines, we go to this non-deterministic, probabilistic world of machine learning, it becomes more important to understand exactly how the humans interacting with technology that's becoming more autonomous, becoming smarter, it's becoming smarter independently based on data that it's absorbing, right? So how humans interact with this is, it's been around for a long time as an area of study, but it's really been raised in terms of its importance.
Now, this framework, which was, again, derived from some defense frameworks which were designed for managing weapons. So this is, how does the human interact with an AI-based weapons system? And then, so the level, they actually had like eight levels. This is being tailored by us to focus on a regulated financial services environment, where thankfully the consequences of a bad process aren't as devastating as any defense context. But still, I have quick examples of each of these, but a level zero is basically, it's all human.
As you move up these... Oops, sorry. As you move up these levels, you start to have more and more interaction. The bot takes on more and more of a stronger role. Basically, the human gets pushed to the side more. And so, I'll show you examples of one, and two, and four, just to give you an example of what that means. So it's very important to understand responsibility of humans as they work more and more with intelligent machines. And hopefully, this framework will help.
This is a level one, which I talked about as attended RPA. So here, again, it's that simple. If you look at ROI land grab, this is address and customer account updates. Very common automation in RPA automation in contact centers. I have an update I just got from the customer. It needs to go into six different systems. So I'm going to pull down a menu and say, "Update address," and a bot is going to take the address I've entered and load it into four different old versions of Siebel, and three different new versions of Salesforce, and so forth. That's good.
The bonus territory is the one I like because it has a customer experience angle to it. In other words, you're improving customer experience and you're also gaining productivity. And this is very common in contact centers to use this type of level one, human-in-the-loop. Basically, the human is still making all the decisions. The human is still controlling the bot activity.
Now, when you move to this where you actually have business process management or what we call DPA wide or DPA deep, you're really interacting with bots in more of an end-to-end process. So here are the humans more in the loop in that the human's going to be a workflow queue that's going to get work after the bot's done. Or the bot's working side by side as in step one, with the bot, or the BPM solutions routing around failed bots to a human queue. So this is very much a traditional workflow task management view of human-in-the-loop. And that's very much what...
The reason why we presented this framework is because if you're talking to an RPA provider, they tend to use human-in-the-loop to refer to the level one that I talked about in contact centers, attended mode. You talk to a BPM provider, they tend to talk about this, where the BPM's in charge and it's using bots really just to augment the human labor. If you're a more advanced analytics environment, you might talk about employee robots, digital assistants. And so, this is where the digital assistant is helping knowledge workers.
You might have an RPA task master type environment, but here the robot is doing a lot more of the decision that... They're selecting the context. They're maybe giving you recommendations. But once you pick one, then the bot's doing all of the work and may inform you later that they've done something. That's just another way to look at it. And then, finally, the bright spots. With all the talk of the pandemic, I want to end on something that's very positive.
This is very much a part of the framework in the invisible robots, was basically I took these... I developed a 12-persona future of work model that had... And I'm highlighting five of the personas here. But basically, three levels of knowledge workers across domain, single domain and function specific. And the top two are really, you call them expert work. The function-specific knowledge worker. It'd be like an underwriter in insurance. It's a knowledge-oriented job, but still it has a large segment of administrative work.
Then you have coordinators which are generally running around with a clipboard or managing Excel to coordinate schedules and so forth, maybe a fleet manager. And then you have cubicle workers. Now, in each of these categories in invisible robots, I have numbers for all of these and their rate of progression and interaction with technology. But basically, if we apply these intelligent automation technology, they have specific roles that they can perform and help these knowledge-based personas on the left. And this is a great opportunity.
Cognitive search, digital assistance, we've talked about. Text analytics and of course process automation, that's more of the BPM, DPA world, and task automation. And of course, conversational intelligence has really, really helped cubicle workers. If you do that, what you end up with is a pretty significant extraction of economic labor value. And that's what we're calling it. $130 billion, this is just if you apply those IA technologies, and pick the year of 2022, the number of hours that you extracted multiplied by the value of those hours, it's massive. $130 billion of economic value, which we all will need in terms of where we're going.
What a company does with that value is dependent. It was different a month and a half ago. Companies were saying, "Oh, I'm going to shift FTEs from the back office to the front office. I want stronger customer touchpoint. I'm going to take a pile of those hours and I'm going to put them into transformative AI projects. I'm going to create innovation labs, and then maybe I'll bank 10% of them. In other words, I will reduce costs by 10%. What those dynamics look like, how that 130 billion of extracted value is going to be managed as we go into this period, it's anyone's guess. But our guess is that there's going to be a lot more of the banking of the value as companies try to fight off a lower revenue, particularly in some industries that may just not be the same for years and years and years.
Positive note, there's a lot of value that you can provide to your companies. And one of the best ways to do that, in my last slide, and I appreciate you staying on as long as you have, is to provide a process internally that allows ideas for this type of intelligent automation to surface from the business. Getting scale for this type of automation really requires you develop a process that allows you to develop them in parallel, that you have ideas going through different chevrons or gates to get down to the ones that really make a lot of sense, because many of them will not.
You have an ideation phase, an initial assessment, and a process for that. And an automation strike team is our word for a center of excellence, which is a bit of a tired term mainly because it's been associated with larger initiatives and companies, and maybe some bureaucracy with that. And a lot of the automation is being federated. It's being developed and spurred by the business. So what you want is this notion of more agility, and more rapid, and more federated in its environment. So strike teams is always a more appropriate term for that.
The strike team would have templates for a business case for maybe using some analytics, digital worker analytics, which is a very hot area right now where you record desktop human input and output and you move it into an environment and analyze it, and it pops out a heat map of different automation opportunities. Very, very powerful future technology, and basically, get bot development going. So if you're not focused on strike teams, on organizing centrally, you're focused on automation, I think this pandemics going to push you in that direction. Companies that had already built these types of structures are in a much better position to accelerate automation without a lot of discretionary spending, which I think is going to be really critical. Our last polling question.
Pat Cameron: All right. Just a little more feedback from you. Which processes would your organization like to automate or need to automate the most? Data entry, excuse me, data extraction, which is always a big one, report generation, web browser, and GUI automation, automating legacy systems and processes, or other? I worked with a customer a couple of weeks ago that is a manufacturing company, and they make hand sanitizer. And what happened to them was that their orders went from a customer that used to order 10 cases is now ordering 150 cases.
And so, they had to automate some of their data entry to change the quantities on what people were ordering so they could keep up. We helped them out. They were able to keep up with the orders that they were getting. All right, give you just another second to answer. Thank you for your feedback. Thank you for joining us. Craig, thank you. I think I'm going to steal your automation strike team terminology.
Craig Le Clair: Okay.
Pat Cameron: All right. Let's take a look at what people are going to automate. Data entry, data extraction, number one, always. Report generation, number two. Always. And some web browser and GUI automation, and then legacy systems and processes. And like Craig said, the beauty is you don't need to change those legacy systems. Automation can work with them just as they are. So thank you very much for joining us today. If you have any questions, you can send them to the question box on your screen. We'll hang out here for a couple of minutes to see if any questions come in. Otherwise, we'll give you back a few minutes of your day. We appreciate you joining us, Craig. Thank you. Very good insights into what's going on in the world out there. Lots of unknowns.
Craig Le Clair: Well, thank you for the opportunity.
Pat Cameron: All right. All right. Well, thank you, everyone, for joining us today. Have a good rest of your day, and hopefully we'll see you next time.