Scaling the Intelligence in Automation Deployments to Survive Digital Darwinism
Having started a new gig at arago, a German pioneer of Intelligent Automation and AI, it feels a little bit like gamekeeper turned poacher. For the last 5 years I had the immense pleasure of helping to shape many discussions and frameworks on the notion of Intelligent Automation and RPA as an industry analyst. As an analyst, you have access to a broad set of stakeholders as you don’t compete, thus you get a holistic view of the market. While moving to the “dark side” is meant to change your perspective to promoting the technology and approach, the access to stakeholders might be limited. In my new role as Head of Strategy at Arago, one fundamental aspect hasn’t changed from my analyst days and that is the common goal of helping to educate the market on the direction of automation and AI and conveying the lessons learned from the early deployments. Truth be told, it is that goal which attracted me to work for Arago. Chris Boos, our charismatic CEO, always will be a tireless educator on the implications of AI, and I am excited to be joining him. So much so, that I often had to remind him over the years to talk also a wee bit about Arago itself. Suffice it to say, over the last 5 years many things have changed, not least the maturation of the RPA market. But many issues remain the same. Here are some of the main challenges that continue to be highlighted in client discussions:
Lack of definitions: Neither Intelligent Automation nor AI have commonly accepted definitions or a clear understanding of the main approaches. Despite a marked maturation of the buy-side, a blurriness around the terminology continues to linger.
Lip service on automation: Despite all the marketing noise around automation and AI, clients tell us about their frustration with their (outsourcing) partners as many of these seek to keep the resulting cost savings for themselves rather than helping their customers build out their automation capabilities.
AI is not a single market: Just like Intelligent Automation, AI should be seen as a continuum spanning disparate technologies and approaches that deliver different benefits. The AI technologies leveraged by RPA vendors and others focus on static and rule-based processes matched to the problems they are solving but that do not match the market hype of AI. Buyers remain confused as to which AI is for them and seek advice as to where and how to start leveraging AI technologies.
Scaling automation in business processes:While we see a maturation in the way RPA is being deployed, scaling projects remains a huge challenge given the need to manage both exceptions (changes to the process) and maintaining the rule-based logic as to which bot to apply to complete a given task. This creates challenges in progressing further on the digital journey beyond the core RPA benefit of automating data transfer.
Breaking down organizational barriers to progress on the digital journey: “Digital” should be about routing a customer request all the way from the interaction to the execution at the backend. Yet, progress in managing this transformation given the disparate number of systems involved and changing business environment has been painfully slow.
Against this background, how does Arago fit into those discussions? First and foremost, having highly differentiated AI technology capabilities focused on Intelligent Automation is a double-edged sword. While it is undoubtedly great to visibly stand out in the crowd, the other side of it is that we run the risk that in a noisy market place we don’t get heard as folks gravitate towards monikers and concepts that they feel they can relate to. In this climate, it makes sense to educate the market less about how we do things using AI technology and more on the automation benefits and ROI we deliver, thus bridging the gap between differentiated AI technologies and Intelligent Automation. Being part of the discussions on Intelligent Automation should get us access to the places where sourcing decisions on automation are being taken, because that is what we are doing.We are offering clients the ability to orchestrate and automate operational processes dynamically at scale. This might sound as convoluted as it might sound unspectacular. Yet, these capabilities are at the heart of operating a digital organization, and more importantly, point to the limitations of traditional automation approaches such as runbooks or RPA in terms of scaling and adapting to changing environments.
HIRO™ AI platform automates by capturing and codifying the knowledge of specialists that run operational processes.The codification is done such that it can be reused and recombined to automate similar but different processes without requiring a new script, a new bot to be recorded to be built. By focusing on the capture and reuse of knowledge, it is designed to be a universal automation engine applicable across industries and processes. We aim for our AI technology to be able to automate any process in any industry – moving from automation toward autonomous processes. The focus on knowledge and experience is crucial for two reasons:
- Automate end-to-end processes to enable a true digital experience spanning the customer interaction at the frontend to the execution of their requests at the backend.
- Allowing incumbents to leverage the power of AI for automation without having to make significant investments in developing their own AI technology or hire expensive data scientists.
Drilling down to the specifics, to solve operational tasks knowledge is transferred to HIRO™ by our customer’s employees through an intuitive chatbot. However, as opposed to recording these as a sequence of steps (like a script), HIRO™ records the knowledge as discrete steps, allowing its AI engine to recombine these steps to automate similar but different tasks without the recording of additional knowledge, scripts or bots. In this way, incremental knowledge added to HIRO™ is leveraged to provide a much higher return in automation rates on time invested than in other options. Figure 1 summarizes the three basic principles powering HIRO’s AI automation:
Figure 1: Basic principles of Arago’s HIRO™ AI platform
Source: Arago 2018
The three core principles of HIRO’s AI:
- Semantic map: Mapping the customer’s processes to HIRO’s semantic data graph so that HIRO™ has a contextual representation of the real world that it can understand.
- Knowledge:Transferring knowledge to HIRO™ via a chatbot in discrete steps.
- Decision-making engine: Combining its understanding of the customer’s context from the semantic map, available knowledge AI (a combination of machine reasoning, machine learning, rule- based approaches, NLP and other algorithms), the engine determines a which knowledge to apply and in which order to resolve and automate a given task. Crucially, the engine allows HIRO™ to deal with incomplete, ambiguous or even contradictory information dynamically. Thus, it is a far cry from the rule-based, static (if not clunky) world of RPA.
To summarize what I have learnt so far from talking to clients, partners as well as to internal stakeholders, four clusters of differentiation stand out for me:
General applicability:HIRO™ is process and industry agnostic and can therefore be scaled throughout an organization across multiple verticals and horizontal layers within a company. This is realized by the AI engine being generalized to solve a task by processing customer specific knowledge related to a customer context as opposed to an AI that is a model or set of rules generated to automate a specific process in a specific industry that is then applied to a customer’s process.
Dynamism:At the heart of Arago’s thinking is the ability to capture and recombine knowledge and experience dynamically to accelerate the digital journey. Whereas static rule-based processes create an exception report with each variation in context requiring a new script, HIRO™ automatically generates the solution minimizing exceptions and overhead.
Scalability:Once the system has learned an initial pool of process steps, with its dynamic task resolution capability, HIRO™ scales significantly faster than other automation tools. Typically, it achieves an 80% automation rate over time without customers having to invest in the significant resources and operating expenses that typically limit the scaling potential of RPA and script-based automation solutions.
Ecosystem mindset:By reinforcing an ecosystem mindset, Arago safeguards our client’s investments in disparate sets of technology by providing connectors to the leading applications and systems. We are also investing in growing our user community and support a marketplace where partners can publish their connectors and applications accelerate the available knowledge and ROI of HIRO™ clients.
Figure 2 summarizes the main differentiators to other automation approaches:
Figure 2: Arago’s differentiators
Source: Arago 2018
What this all comes down to is that we must stop trying to retrofit all these wonderful innovations such as Intelligent Automation and AI into the old models. Rather we need to start reimagining the potential of AI beyond scripts and standardized processes and vision an operating environment where AI is leveraged to automate processes in today’s business world where processes and business are constantly evolving and changing. In that respect, working with our HIRO™ platform is a leap into the unknown, or more precisely, about applying a change in mindset.
There are no longer “standard processes” andHIRO™ delivers automation of things that cannot be standardized. Organizations need to start working with HIRO™ which, unlike most other solutions,allows organizations to adapt quickly to change. For many organizations this is the missing link to accelerate their digital journey. Fundamentally, for companies of the old economy to survive the Digital Darwinism and outlast the disruptors, they must adopt new (and bold) approaches encapsulated in HIRO’s AI engine.
If organizations want to invest in yet another point solution in order to address yet another automation task and stay with the tried and frustrated solution, we are as blunt as we are transparent in that they are better off looking for an optimized runbook rather than HIRO. However, if organizations want to overcome some of the challenges they face scaling their static systems and seek a solution that can both automate across the business and orchestrate existing automation investments, then HIRO™ offers a unique and consequently vastly different approach. With our upcoming HIRO™ 6.0 release, clients can build on the lessons we learned in our heritage in IT use cases and drive the platform into business process use cases. Thus, we aim at the heart of digital operations: managing and automating processes across organizational boundaries. Critically, this allows organizations to route customer requests all the way from the interaction to the execution at the back-end.
With that the loop gets closed to my earlier point that we need to learn from the early automation deployments in order to progress on the digital journey. We at Arago are committed to engage with all stakeholders to accelerate the understanding of the impact of technology innovation and in particular AI on operational processes. On a personal level, I am thrilled that my experiences as an analyst will actually come in quite handy in that regard, even though I have moved to the dark side.
