C-3PO and R2-D2 are an odd couple in the Star Wars universe. C-3PO is a cowardly droid who obeys pre-defined protocols and routine tasks, while R2-D2 is a curious and adventurous robot who learns from previous problems, uses logical thinking and larger concepts to solve new problems. But together they do things they could not do alone.
Similarly, RPA (Robotic Process Automation) and Advanced Analytics are an odd but very complementary combination of new business technologies. Like the diligent but unimaginative C-3PO, RPA follows precise rules to execute repetitive business processes; and like the curious and adaptable R2-D2, Advanced Analytics learns to make complex judgments when faced with new situations. Companies can get quick wins with RPA while strategically introducing Advanced Analytics for sustainable repositioning.
Now, almost 40 years since the first film, Star Wars has introduced a newer generation of robot: BB-8. While we could predict how C-3PO and R2-D2 would respond, BB-8 is far more humanlike in how it adapts its behaviour to each situation: sometimes naïve, often talented, at times industrious and occasionally even manipulative. BB-8 is a cognitive automat.
The boundary between RPA and Advanced Analytics is blurring with the evolution of Cognitive Automation. Future cognitive bots will communicate with us in natural language and will watch us go about our daily tasks – they will understand what we need done and offer to take over when they feel confident, while still asking for our input from time to time.
An artificial intelligence transformation can only be driven by a fully empowered C-Level champion.
C-3PO or Robotic Process Automation: Commercially available quick wins
In 2002, during my first summer internship, I met staff whose job was to manually process hundreds of invoices every day on MS Word. By the time I left, we had managed to automate this job out of existence with a simple macro that opened the bills, fetched key information and sent e-mail reports 24/7. RPA is essentially macros taken to the extreme.
Far from losing their jobs, the staff were able to focus on new and existing tasks that were more valuable to the firm. Freed from pre-programmed robotic behaviour, they had more time to function as thinking human beings – much like Anthony Daniels did when climbing out of C-3PO’s shell after a whole day of shooting Star Wars.
Anthony Daniels gets out of C-3PO’s shell after a whole day of shooting
Firms today can easily deploy RPA to help their workforce avoid hours of mundane and repetitive work. For many, RPA may still evoke images of advanced Japanese manufacturing plants where robotic arms assemble cars and flat-screen displays. While manufacturing pioneered RPA with hardware robots, service management industries today – like banking or telecoms – can automate their processes with RPA in commercially available software robots.
Saving time, cutting costs and improving the quality & accuracy of processes
Ideal targets for RPA are processes that are stable, mature, optimized, rules-based, repetitive and high volume. Many back- and front-office processes fit this description.
Think of contact centres, where customer files need constant updating of address changes or service additions. Or banks, where a lost credit card means updating customer records across multiple systems, and keeping customers informed during the process. In addition to relieving service staff of these repetitive processes, RPA can also reconcile failures to charge for services across billing systems, or even automate many of the reporting processes in accounting or finance departments.
As a result, RPA can potentially cut operating costs by 20-80% in certain areas. Beyond these significant savings, companies can make standard processes more predictable and accurate by removing the potential for human error in repetitive tasks, as well as making them easier to scale.
Customer experience can be massively enhanced through reduced cycle times, improved throughput and by redeploying people to activities where they can empathise with customers.
Automation can also improve regulatory compliance – a potentially large drain on resources and productivity in industries like banking.
Finally automation can increase scalability of processes such as data engineering, which a have a growing demand, or improve data security because robots are completely traceable.
Lightweight and cost-effective IT development
RPA involves getting a bot to take over an existing job. This means that the bot will interact with any existing programs required for the job, like ERP, Word, Excel or SAP. Imagine C-3PO taking over invoicing – he sits at the desk of a finance executive, uses all the same software and starts clicking and typing away – except as a virtual robot within the executive’s PC.
Installing RPA is far simpler, quicker and cheaper than traditional forms of IT integration. Instead of relying on machine-to-machine interfacing with APIs (Application Programming Interfaces), RPA is added on top of existing IT systems and processes. Without costly integration, RPA can be implemented in a matter of weeks and pay for themselves within a year, while typical IT implementations take several quarters and often have three-year payback periods.
RPA also does not require complex and time consuming process-reengineering. However customer facing processes can be re-engineered before being automated.
However, companies still need to be strategic about how and where they deploy RPA. An unplanned proliferation of bots can threaten the integrity of the overall IT architecture. Additionally, although bots are much faster than humans, they are still orders of magnitude slower than fully automated IT systems.
Powerful and intuitive drag-and-drop automation tools
Solutions range from off-the-shelf tools customized for specific processes like finance or web scraping, to scalable and reusable enterprise software suites.
The market also offers powerful and intuitive drag-and-drop tools for developers to customize bots for specific tasks involving extremely detailed and company-specific business rules and logic.
A workflow controller is also required to assign jobs to robots, store credentials, and most importantly, to manage handoffs between robots and humans.
While solutions are usually generic, every one of them has spikes in specific in processes or industries.
Managing a workplace of humans and robots
The great concern about RPA among employers and employees alike is that it will push many people out of the workforce. This is a very unlikely scenario.
Simple bots can take care of repetitive business processes more rapidly and accurately than humans. Today, bots can already execute most of these processes: McKinsey estimates that about 64-78% of all time spent on data collection, processing and predictable physical work can be automated.
Instead of making workers redundant, RPA can drive business value by supplementing human activity within a much broader job, or by freeing workers to focus on new and more rewarding tasks altogether.
A transition from automation to intelligence occurs when humans intervene in a rules-based processes; these transitions require deliberate and systematic governance.
For example, a simple rule-based customer service bot can independently handle about 10% of queries, allowing customer service representatives to focus on the other 90% that require empathy or better contextual understanding. Customer service representatives can be refocused on higher-value activities that are often more rewarding, which can reduce outsourcing in many cases.
Paradoxically, automation and Artificial Intelligence actually make us more human because they allow us to focus on human relations and abilities.
How RPA complements Advanced Analytics
RPA blindly follows rules, and it is limited by an inability to learn. When rules conflict with reality, or when faced with unexpected events, RPA bots will continue to execute a flawed process until a human realizes and steps in. This is where Advanced Analytics comes into the picture for most companies.
R2-D2 or Advanced Analytics: An ongoing transformation in today’s enterprises
R2-D2 is a very smart robot. Some of his most impressive achievements include repairing the shield generator of Queen Amidala’s Royal Starship, fixing the Millennium Falcon’s Hyperdrive, and shutting down the garbage mashers of the Death Star. How did he learn to solve such complex tasks? The answer: through machine learning.
R2-D2 actor Kevin Baker resting during shooting
Machine learning is what differentiates Advanced Analytics from RPA. This technology enables a machine to learn how to solve a complex problem without needing the extremely precise kind of step-by-step instructions RPA requires.
Most Advanced Analytics algorithms used today in business identify and predict patterns in well-structured data, through three machine learning techniques called supervised, unsupervised and reinforcement learning. The article “What a CEO needs to know about machine learning algorithms” covers the state of the art in detail.
Advanced Analytics can be applied to every function of the enterprise:
- Within marketing, companies can target digital advertising to each individual customer, or predict the products each customer is most likely to buy.
- Within operations, companies can identify safety and quality issues by automatically processing images from a production line, or detect service anomalies in telecom networks.
- Within human resources, key attributes of leaders and managers can be assessed to better understand behaviours, develop career paths, and plan successions.
- Within back-office functions, machine learning can help detect fraud in real-time in credit card transactions or in insurance claims.
BB-8 or Cognitive Automation: A family of emerging technologies
When Rey freed BB-8 from the scavenger’s net, she did not for a single moment intend to keep it: “Don’t follow me!”. As BB-8 persisted, Rey roared angrily: “No!”. We do not know what BB-8 said then because its beeps were not subtitled – but with empathy and charm, BB-8 convinced Rey to reluctantly keep it until the next morning. They never separated from each other afterwards.
Its ability to empathise with and influence humans by adapt its behaviour in response to the human’s own emotions is what makes BB-8 a cognitive automat.
BB-8 convincing Rey to keep it, with puppeteer in green in the background
Cognitive Automation combines a number of disciplines such as machine learning, natural-language processing and generation, unstructured information analytics such as text and images, data mining and semantic technologies. The integration of all these technologies allow automation in processes where RPA alone could not play.
Whereas RPA and Advanced Analytics are well established commercially, cognitive automation is less mature. Having said that, conversational bots are improving fast, thanks to free tools such as Google’s Dialogflow which enables human–computer interactions based on natural language conversations.
Future generations of cognitive agents will watch humans in action and communicate with them. They will learn humans do tasks and will manage to perform them even better by learning from data. When they reach confidence, the automats will take over these tasks, making decisions based on judgment and sometimes intuition, but still requiring human input from time to time.
For example, a workforce of cognitive automats could offer round-the-clock customer service in perfect English or even a mixture of Sinhalese and English as spoken on the streets of Sri Lanka. Cognitive automats would be able to address complex queries in a broad and growing array of issues from technical support to product and service recommendation, which would increase personalization, engagement, and sales. Unstructured information such as customer interactions would be easily analysed, processed and structured into data which can be used for predictive analytics. Moreover, teams of cognitive automats could train new employees on how to perform face-to-face with customers.
Conclusion: driving an A.I. transformation from the C-Suite
Luke Skywalker (Mark Hamill) with Harrison Ford and Carrie Fisher
In Star Wars, Luke Skywalker is fated to lead the fight against the Empire and later the First Order. Nobody else could have done it but him. But he was not alone. He had a team.
Likewise, in a corporate environment, only a fully empowered C-Level champion can drive an Artificial Intelligence transformation. The CEO, unless in a smaller organization, has too many things on his plate; VPs are not influential enough; the board is non-executive; and the incentives of management consultants and vendors are misaligned. Nobody else can do it but a C-Level champion, with a dedicated team collaborating tightly with other C-Levels.
1. This team articulates a strategy which creates a competitive advantage beyond pure cost savings. This strategy is clear about the target state, the journeys to focus on, the roadmap and the exact heat map of opportunities plotting value against feasibility.
2. This team builds up momentum and captures value by running a successful pilot after another successful pilot, following a flexible roadmap, avoiding a process-oriented big-bang approach to the extent it is possible.
3. This team designs solutions for each pilot along the full portfolio of technologies, delivering quick wins through RPA, strategically introducing Advanced Analytics for sustainable optimization, and watching out Cognitive automation as an emerging area.
4. This team goes beyond the traditional internal data and integrates external sources.
5. This team progressively evolves into a centre of excellence, ideally in marketing or operations. The team includes data scientists, data engineers, business analysts, developers, designers; with skills to assess business units’ needs and proposals, to assemble data from different sources, to create advanced analytical models, to configure robots and, to deploy robots and models in the business.
6. This team scales up the pilots progressively delivering sustainable and scalable solutions, balancing between nimble and legacy systems.
7. This team persuades the organization to adopt artificial intelligence, managing resistances and fears from employees, helping them to acquire new skills, communicating with them, setting values and purpose.
8. This team democratizes data in the organization and redesigns business processes to incorporate data insights into actual workflows, delivering insights to mid-level decision makers across the organization.
9. This team defines new operating and governance models where machines and humans interact in an agile environment of constant learning with numerous transitions between each other.
10. And most importantly, this team sets up an automation and intelligence first imperative within the organization.
Are you the Luke Skywalker of your organization? May the Force be with you!
Disclaimer: Opinions in the article do not represent the ones endorsed by the author’s employer.
“What CEOs need to know about Machine Learning algorithms”, February 2018, Pedro Uria-Recio
“Intelligent process automation: The engine at the core of the next-generation operating model”, March 2017, Federico Berruti, Graeme Nixon, Giambatista Taglioni, Rob Whiteman, Mckinsey.
“What’s now and next in analytics, A.I. and automation“, May 2017, James Manyika, Michael Chui, Susan Lund, Sree Ramaswamy, McKinsey
“Where machines could replace humans and where they can’t yet”, July, 2016 Michael Chui, James Manyika, and Mehdi Miremadi, McKinsey
All images from Star Wars