How was the book?

If you are as a business leader interested about a step-by-step way forward with Robotic Process Automation (RPA) and Cognitive Automation (CA) – then this a great book for you. It also delivers also business cases with a triple win methodology. Bonus part is that you can get easily understanding about the solutions that are available for both domains – RPA and CA.

This book should also be mandatory for all leadership teams, because this way you will get an understanding what other industries are going through.

What are the key learnings of the book? 

Key question to all business leadership teams is that do we have a strategy that links automation with other digital technologies. Do you?

If not – start developing one and simultaneously test your service automation strategy with proof-of-concepts and piloting. There is a need for strategy that ”sees technology augmenting, complementing and amplifying humans skills”. Managers should be able to answer based on the new strategy how they will prepare and train their people to seize the new business opportunities. Can you?

RPA can be a strategic enabler. RPA Strategic usage is built on:

·       Understanding and planning for the mid-term and long-term endpoint

·       Aiming for the ”triple-wins” for shareholders, customers and employees

·       Resourcing RPA as a strategic business project

·       Ensuring the C-suite is completely on board with the strategic vision

·       Identifying and managing change and implementation challenges from the start

·       Centralizing automation capability to accelerate scaling

RPA/CA will be game changers. This phase will take us to year 2027 and the game will be guided by human intelligence. Bigger picture of future of automation is that:

·       RPA & CA are only part of bigger technological developments

·       RPA & CA will have an important role within social, political, economic, legal and regulatory context.

If the RPA/CA is a game changer – how will businesses deploy RPA/CA, Internet of Things and business analytics to gain competitive advantage?

·       Acceleration,

·       Employment,

·       Security,

·       Privacy and

·       Environmental sustainability.

About RPA/CA market

First are foremost everybody should understand that human capabilities are not easy to replicate. Especially if we are talking about capabilities such as leadership, empathy, creativity, sense-making, intuition and so on. Secondly, the vast and popular debate about Artificial Intelligence (AI) is more or less Cognitive Automation (CA). Service automation is divided into:

·       Robotic Process Automation (RPA)

·       Cognitive Automation (CA)

If you want to get readiness to deploy these technologies you should do research around service automation:

·       Why clients are adopting automation?

·       What outcomes are they achieving?

·       What practices distinguish service automation outcomes?

There are 45 companies that provide RPA solutions. Popular RPA companies are Blue Prism, Automation Anywhere and UiPath. Advisors in the RPA space are Symphony Ventures, ISG, KPMG, EY, Accenture. Redwood software provides Accelerated Robotic Automation (ARA) and it is connected with APIs to the ERP. By the way Finnish OpusCapita is mentioned as a decent service provider. Size of RPA market is 250 mUSD (2016) to 2,9 billion USD (2021).

AI or CA solutions provides 120 companies. Popular AI companies IBM Watson, Workfusion, IPsoft, Expert System, Nuance and Digital Reasoning. Size of the CA market is 1 billion USD (2018) to 11 billion (2024). AI is cognitive automation. AI mimics cognitive functions

What happens to workforce? Nearly half (49%) have redeployed the FTEs within their unit. Only third (32%) reduced the number of FTEs. Half (51%) of the users where satisfied with their RPA deployment. And half (52%) where satisfied with their cost savings. RPA might provide 30% improvement and other positive outcomes. Triple-win companies (shareholder, customers and employees) are Associated Press, Telefonica O2 and Xchanging.

The RPA market took off during the 2016-2018 period. 51% of hi-tech and 44% of banking&insurance companies will invest into RPA solutions. RPA PoC are deployed by 51% of companies across industries. Only 32 % rely on their current providers i.e. there is a huge potential for new business. RPA is alive and well plus growing exponentially.

Blockchain is to transaction, what Internet was to information. Blockchain will decentralize, democratize and disintermediate transactions. RPA/CA is that to business operations.

RPA as a business transformation process

Important to incorporate RPA into corporate strategic thinking, service automation, digital developments and business strategic intent. The Royal DSM case shows that RPA is best treated as a strategic long-term investment and not as a one-off tactical initiative. Do not buy technology, buy business solutions that fits into your agenda.

RPA deployments have seven risk areas:

·       Strategy

·       Sourcing selection

·       Tool selection

·       Stakeholder buy-in

·       Automation launch

·       Operations/change management

·       Road to maturity

Action principles to follow in a RPA deployment:

·       Cultural adoption by C-suite for the Strategic RPA

·       Let business operations lead RPA

·       Send the right messages to staff

·       Evolve the composition of RPA teams over time

·       BIG THING: Identify process and sub-process attributes ideally suited for automation

·       Prototype continually as RPA expands to new business contexts

·       Re-use components to scale quickly and to reduce development costs

·       Bring IT on-board early

·       Build robust infrastructure

·       Consider RPA as a complement to enterprise systems (automate systems that ERP does not cover)

RPA is not for quick wins. The evolution goes typically:

1.  Phase is Hype & Fear

2.  Phase is Focus primarily on ROI

3.  Phase is Focus on the Triple-Wins

4.  Phase is Institutionalized

RPA should be seen as a change management program. You should try to combine RPA, BPO and ERP. However, RPA and CA are complementary technologies.

RPA can deliver:

·       Processes automated

·       FTE savings

·       Process time savings

·       Increase productivity

RPA could enhance:

·       Quality issues

·       Regulatory compliance difficulties

·       Customer dissatisfaction

·       Labor shortages

·       Widespread use of temporary workforce

·       Cost & process inefficiencies

The RPA value proposition is real but different in a complex, low volume service environment:

·       Changing the content of labor

·       Upskilling

·       Driving focus from back office data processing into front office

Cognitive automation

CA systems do not think or understand – they manipulate symbols based on algorithms. Nowadays we are talking about how BPM (business process management) and RPA joins forces, but CA is still for early adopters. Period 2018-2020 will be a major period for automation strategy and cognitive automation. Cognitive automation is about machine learning, image processing, natural language processing and fast computes.

Three specific CA tools:

·       IBM Watson

·       IPSoft Amelia

·       Expert System’s Cogito

Machine Learning:

·       When supervised Machine Learning gets new data the algorithm instructs the computer to match the new input to the closest pattern.

·       In unsupervised Machine Learning the algorithm extracts patterns based on the data, but it needs vast amount of data to perform competently.

·       In deep learning the algorithms build a hierarchy of equitation’s for processing the data inputs and the data can be labeled or not.

Image processing and machine learning:

·       For example, alphabetic characters and image processing / ML is a good case example. Too many characters and handwriting styles.

·       In supervised training human teaches that which character is which letter. In unsupervised the algorithm commands the computer to categorize the data by finding patterns among the elements. This requires massive amount of data – upwards of tens of thousands of examples.

Natural language processing

·       This more complex than image processing, because the algorithm must be extracting the semantic intent of text or speech from the relationship among words and sentences.

·       This is relatively old technology, because the NLP algorithms have been around for 50 years. NLP machine learning has been available since the 1980s

·       Data extraction is used for example in Zurich Insurance to extract data from handwritten medical report with the help of OCR and CA

·       Classification is used in Virgin Train to filter and categorizes 470 types of email correspondence with a CA tool by Celaton called inSTREAM. Daily processing time dropped from 32 man-hours to four. It means six (6) FTEs.

·       Language translations is used in SEB and they used few weeks to train IPsoft

·       Amelia to execute Swedish in banking processes.

The wrenches hinder the usage of RPA or CA technology.

Dara wrenches are:

·       Data wrenches are difficult data (fuzzy images) – OCR can recognize 98% of images,

·       Dark data is 80% of organizations data (emails, text messages, videos etc.)

·       Dirty data is missing, incomplete, incorrect etc.

The Algorithm Wrench

Limitations of today’s Machine Learning Algorithms

·       We don’t know how to build algorithms for ”one-shot” learning

·       The algorithm cannot explain its actions and there is no audit-trail for humans to supervise the actions

IBM Watson and Blue Prism

Year 2011 Watson won from Jeopardy one million dollars. It was triumph of Machine Learning when IBM Watson won the Jeopardy using supervised machine learning (SML). Before SML the hit rate for IBM Watson was 10%, but after SML and upload of 10.000 Jeopardy questions the hit rate was close to 90% and similar to humans.

Triple WIN Business case – Zurich Insurance and Blue Prism:

·       5 mUSD yearly savings

·       39 000 hours saved per annum or freed up to other activities

·       Time to analyze a medical report was taken from 58 min to five seconds!!!!!!

Action principles from Zurich Insurance:

·       Strategy drives CA investment

·       Use RPA as forward reconnaissance

·       Manage as an innovation process

·       Put in place a strong in-house team

·       Don’t look for a Swiss army knife

·       Test the providers tool with a controlled experiment

·       Manage expectations

·       Create new process flow

Automation and the Future of Work Revisited

Are going live in an Automotopia, a well-run technologized world? Alternatively, are we going to live in an Automageddon where humans are replaced with robots?

Problem with predictions about future of work by Frey & Osborne, 2013:

·       The study do not take into consideration the jobs that are created by the change.

·       The study focus on job and occupation, but not with the activities nor the amount of work that needs to be done.

·       The long road to diffusion of dilemma is ignored.

Three engineering barriers to computerization and these qualities are needed at work:

·       Complex perception and manipulation tasks

·       Creative intelligence

·       Social intelligence

In work we need 18 human capabilities and with the help of automation we can perform 11 capabilities at human level within 15 to 50 years. New technologies will enable new products and services alas new work. For every 20 jobs that are lost another 13 are created. An overall job loss and creation estimate is that 12-14 percent across the top 20 economies of the world within the next 15 years.

When looking at RPA/CA the key questions are:

·       Will the technology itself provide advantage? (Key success factors etc.)

·       What will be the innovation implementation process? (silos, legacy systems, organizational politics)

USA is short of workforce and Japan’s workforce is already shrinking. To achieve required aggregate GDP per capita growth of 2,9% we need to grow the size of our workforce. The automation could increase economic growth by 0,8 – 1,4 % annually.

We are seeing sources of more work comes from:

·       Exponential data explosion leads to massive amount of work to be done with collection and analysis of the work

·       The cross sectional explosion of audit, regulation and bureaucracy combined with need for transparency

·       Technology’s double-edged capacity to provide solutions that also create additional problems such as cybersecurity

The diffusion of new technologies typically take 4-5 years in major organizations. This is what we are looking for in RPA and CA. It will take eight to 28 years to reach 90% adaptation level.

How should we change according to the book?

If you believe in game changers, now it is time to act.

What should I personally do? 

Draw few pics based on the book.

Summary

The book in six words – ”RPA is a business transformation process”.