IBM Automation Insider

A bimonthly round-up of useful information to help you automate all types of work at scale

March/April 2019

5 articles


25 min

Why are knowledge workers still doing routine work?

By Brian Safron

Have you ever wondered why skilled workers often spend valuable time doing routine work when there are more automation technologies than ever to automate these repetitive tasks? Is it really the best use of resources for experienced employees to spend time on tasks that could be easily automated? Probably not, if you’re trying to improve productivity or customer service.

So, why are some companies slow to adopt automation software that could help employees be more productive —  that could free them from lower value tasks and assist them in higher value activities? In most situations, in order to go beyond just rote functions and calculations, the software must include some degree of embedded intelligence.  And most automation software lacks that critical capability.

Consider a loan officer named Lisa. As shown in the graphic below, Lisa’s major activities include: 1) inputting data from document and forms; 2) meeting with her current clients; 3) finding and building relationships with new clients; 4) resolving problems, issues and exceptions related to loans; 5) creating compliance and financial reports; and 6) managing the end-to-end loan process and documentation.  

A day in the life of a finance professional

Figure 1. A day in the life of a finance professional

Like most employees, Lisa does whatever it takes to get the job done. Her dilemma is that many of her activities don’t add value, so that much of her time is spent on low value work with little time left for high value work.  

The knowledge worker’s dilemma

Figure 2. The knowledge worker’s dilemma

To make knowledge workers more productive, automation should free them up to focus on inherently human strengths (for example, strategy, judgment, creativity, empathy) while machines should support them by performing repetitive, data-intensive (and often mind-numbing) tasks.  As shown in Figure 3, the goal of intelligent automation is for Lisa to spend more time finding new clients and meeting with current clients (as seen on the right) than she does inputting data and managing loan approvals (on the left).  

How Lisa spends her time before and after intelligent automation — size of circle equates to time spent

Figure 3. How Lisa spends her time before and after intelligent automation size of circle equates to time spent

Using intelligent automation to offload repetitive tasks and assist knowledge workers with their skilled work is what we call “digital labor,” which can be divided into the following three segments:

  • Digital clerk: Intelligent automation can function as a “digital clerk” by performing repetitive, mundane tasks. For example, Lisa shouldn’t spend her time inputting data from documents, but since inputting data requires a certain amount of intelligence, such as knowing which fields on a scanned document belong to which fields in a software application, Lisa and her colleagues often end up inputting the data manually.
  • Digital advisor: Intelligent automation can act as a “digital advisor” by assisting knowledge workers so they can spend more time performing uniquely human activities. For example, when Lisa plans a meeting with clients, intelligent automation should be able to gather the appropriate information and predict which loan products would be most beneficial to the client and therefore most likely to be accepted.
  • Digital self-service: Intelligent automation can enable “digital self-service” by providing end users a nuanced experience equal to, and often faster than, the equivalent nonautomated experience. Customers and other end users often prefer being empowered to serve themselves — on their own schedule — as long as the automation can intelligently navigate the variations and decisions inherent in the interaction. By offloading work directly to end users, employees are freed up to focus on those high-value, mission-critical interactions that require human empathy, creativity and judgment.  
Figure 4. Digital labor — supporting the full spectrum of digital work

Figure 4. Digital labor supporting the full spectrum of digital work

Digital labor mirrors human labor in its ability to perform a variety of skills. The skills are as varied as extracting information, making decisions and anticipating actions. To be effective, the skills must be performed in the specific context of the work being done. Because intelligent automation must mimic human intelligence in order to perform these skills in the context of the work, we call the skills “moments of intelligence.” Six key moments of intelligence are shown below.  

Figure 5. “Moments of intelligence” — skills that can be automated in the specific context of the work being performed

Figure 5. “Moments of intelligence” skills that can be automated in the specific context of the work being performed

In a recent publication, Gartner stated that “by 2022, one in five workers engaged in mostly nonroutine tasks will rely on AI to do their jobs.”[1] Human expertise is always in short supply, and for years we’ve looked for ways to make our experts more productive and take unnecessary work off their plates. Intelligent automation offers a solution for finally getting closer to that goal.

Note: We’re in the process of developing a new automation capability –  IBM Business Automation Intelligence with Watson™ – designed to improve productivity across the full spectrum of digital work. We’re running an early access program to incorporate as much customer feedback as possible into the development of the offering. If you’re interested, click the link above to join the program.

[1] Gartner Magic Quadrant for Intelligent Business Process Management Suites, 30 January 2019 - ID G00345694

Upcoming ROI webcast

Building the business case for an automation platform solution

Join special guest, Sean Owens, Principal Consultant at Forrester, on 1 May 2019 at 1 PM EST. 

Sign up for the IBM Automation Insider

What automation opportunities are out there? How do you get results?
Every other month our experts share five pieces of strategic content to help you drive growth through automation.

True detective work: A case of unrealized data potential

By Cheryl Wilson

Police work has launched a lot of TV shows and films featuring detectives collecting, cataloguing and analyzing evidence to solve crimes. We watch them catch criminals and turn them — and the evidence — over to attorneys who prosecute the cases.

But TV and film dramas can’t fully capture the scope of work done by agencies like Edmonton Police Services (EPS), in Edmonton, Alberta, Canada, including:

  • The sheer volume of evidence available, in both digital and nondigital forms
  • The skills required of the officers who analyze it
  • The time required to go through all of it to find clues

As Greg Preston, Deputy Chief of the Intelligence and Investigations Bureau at EPS, explains in this case study, “Our policing efforts bring in an incredible amount of data, from almost every source imaginable … but it takes a small army to piece through it all and identify the key pieces of evidence that help police officers solve cases.” 

Like many organizations across industries, EPS needed to find a way to use all of that data and information to make it easier for employees to deliver the best outcomes. Ultimately, EPS wanted to help police officers and other employees improve the safety and quality of life for the citizens of Edmonton.

A case of unrealized data potential

In the Automation Insider interview The race to create intelligence is on, Mike Gilfix, IBM Automation Executive, stressed the importance of digitizing as much data as possible to enable the integration of intelligence – which is built on a foundation of data – into daily operations. As Gilfix explains, “The more digital your operations, the easier it is to automate them.” And the need to digitize applies especially to the extraction of information or how people can get information from unstructured business data.

For EPS, a lot of their evidence already is digital — records, video, forensic evidence and more. But it wasn’t in one place. Like many law enforcement agencies, EPS stored different data types in purpose-built software programs. Each of its teams had distinct ways of managing information, making it difficult for the organization to build a common view of the data.

With millions of pieces of digital evidence, and with data volumes only continuing to grow, EPS needed a solution that would allow it to put its data to good use.

How they planned to solve it

The answer for EPS is building Canada’s first enterprise digital policing platform, offering teams new ways to access, manage and analyze a wealth of data from across its operations. With a digital policing platform, EPS will be able to free teams from searching for documents and digital exhibits, so they can focus more on analysis and investigative work to generate leads and solve cases.

“Our people are very good at what they do, and we wanted to give them the tools to be even more effective,” says Brock Kahanyshyn, Chief Information Officer at EPS. “By bringing all our digital records together in one place and making it easier for people to surface leads hidden in the data, we could point investigators in the right direction faster, saving time and resources, and potentially helping them solve cases sooner.”

At the heart of the new platform will be a centralized data repository that employees can use to search and access digital evidence. EPS will employ additional automation capabilities, such as:

  • Data capture to automatically extract key data from paper records, transform it into digital content and deliver the information to the repository
  • Case management to group related content in a single electronic workspace and streamline case-related activities so teams can access and work with information in a more efficient and controlled way
  • Records management to help ensure that all content is managed consistently from creation to deletion

A foundation for adding intelligence

EPS plans to bring the new platform into operation in a controlled, employee-centric way, aiming to have basic functionality in place this year. The organization plans to evolve the platform over several years (for example, by adding machine learning capabilities to analyze large volumes of structured and unstructured content and reveal patterns).

“Data is one of the most powerful weapons that we have in the fight against crime, and our digital policing platform will help us make the most of this asset,” concludes Preston. “Our police officers put their lives on the line to protect and serve the people of Edmonton, and we owe it to them to deliver the intelligence and tools they need to keep our community, and themselves, safe.”

For more details on EPS’s solution approach, read the full case study.

Upcoming ROI webcast

Building the business case for an automation platform solution

Join special guest, Sean Owens, Principal Consultant at Forrester, on 1 May 2019 at 1 PM EST. 

Sign up for the IBM Automation Insider

What automation opportunities are out there? How do you get results?
Every other month our experts share five pieces of strategic content to help you drive growth through automation.

Three tips to help scale beyond your RPA Proof of Concept (plus two bonus thoughts)

By Gurtej Chawla

Congratulations. Your robotic process automation (RPA) proof of concept (PoC) worked. Now what? Can you take what you saw and apply it to a larger set of tasks within an important back office process to achieve a desired outcome?

Assuming you have RPA strategically positioned with executive support and the right resources in place, there are a few things you can do to ensure a successful post-PoC implementation. The idea is to plan big, start small and grow fast.

1. Maintain a queue of prioritized processes to be automated

To keep moving forward, you’ll want to strike while you have the momentum and attention of your team and stakeholders. Do this by keeping a backlog of future projects prioritized by processes that will yield the highest, and possibly fastest, ROI. That way, you always have the next project ready to go.

2. Create reusable low-code assets

In order to scale quickly and effectively, it’s important for business users to have the ability to automate their repetitive work without a heavy reliance on IT. You can do this by empowering them to develop bots with reusable low-code assets. This makes new solution builds efficient, reduces risk and increases compliance by using previously approved designs.

3. Combine RPA with other automation capabilities

To get even more out of RPA, you can begin to extend its value by augmenting task automation with additional automation capabilities. For example, you can make your bots smarter with decision management, giving you the agility to change rules on the fly with no downtime.

Watch this two-minute video to see how additional automation capabilities can help your bots orchestrate workflows, integrate with business rules and decisions, manage content and capture data.

Bonus thought #1: Include IT early and often

Like most tech-related endeavors, a successful RPA project or initiative is anchored in strong relationships and coordination between the business and IT. Partnering early provides the following benefits:

  • A clear understanding of who does what
  • An appropriate amount of time allotted to IT to build a solid foundation
  • An agreed upon RPA solution that works for both the business and IT

Bonus thought #2: Establish an automation center of excellence (CoE)

In the last issue of this newsletter, my colleague Benjamin Chance covered the seven features of a successful automation center of excellence saying, “It’s the job of the automation CoE to ensure you have the muscle to start to work in new and disruptive areas — to embrace the resultant human change and ensure that automated processes run smoothly and are strategically improved with new technology.”

The benefit of an automation CoE, whether you call it that or not, is having a group in place to champion your RPA efforts, to act as a singular force in charge of your governance framework, including project evaluation and program analysis.

To sum it all up, there's no one right way to succeed with RPA, but to help ensure success:

  • Plan big by ensuring all stakeholders (business and IT) are on the same page and champion resources are in place.
  • Start small by moving beyond your PoC with a “can’t-miss” project from your prioritized list.
  • Grow fast by empowering users with reusable low-code assets.
  • Extend the benefits of RPA by combining it with other automation capabilities.

To learn more about extending RPA’s value, watch these demo videos to see IBM RPA with Automation Anywhere at work with:

Upcoming ROI webcast

Building the business case for an automation platform solution

Join special guest, Sean Owens, Principal Consultant at Forrester, on 1 May 2019 at 1 PM EST. 

Sign up for the IBM Automation Insider

What automation opportunities are out there? How do you get results?
Every other month our experts share five pieces of strategic content to help you drive growth through automation.

“Should I redesign my process before I automate it?”

By Katie Sotheran

As automation programs gather steam, the question of how automation fits alongside existing improvement initiatives often comes up. Should you attempt to automate the existing process – mess and all – or improve and redesign it first?

Unfortunately, it’s not a simple yes or no answer. Depending on your objectives and process complexity, you could take either approach. For example, you could:

  • Automate a couple of bottleneck or error-prone tasks, making the process more customer-pleasing. By digitizing those parts of the process, you’ll generate meaningful data that can support more in-depth analysis and understanding of where and how the process can be improved overall.
  • Entirely reengineer a process to create a new, automated, intelligent workflow. In this case, improvements likely won’t be incremental but exponential, with a new distribution of labor.

Either way, determining the right approach comes down to answering the following two questions:

1. How broken is the process?

A standardized and relatively efficient process may not need redesigning and could be automated entirely with technology standing in for human activity at each step – from data entry or document ingestion to decision management, task execution and communication activities. The result will be a change in the way work is distributed between people and technology (for example, bots, data capture tools, rules engines and so on).

At the other end of the “not broken to broken” spectrum are workflows where the only way to deliver value is to completely reimagine it for automation. This is often the case when the process is designed around a set of human or system constraints that are no longer needed, such as a process that was designed with legacy systems in mind, where new microservices can remove those constraints. Tinkering with task-level automation in these cases will deliver only a small portion of the potential benefits – and, as Mike Hobday, VP of IBM Automation in Europe puts it, you run the risk of embalming a dated operating model by removing the need for more radical changes.

2. What’s the business case for improvement?

Whether you automate or redesign your process first also depends on understanding the strategic purpose of the change. Will improvements reduce operating costs or affect customers and drive growth? Is it part of a department-sponsored improvement initiative or a broader enterprise-wide digital transformation?

Here are some quick, general guidelines:

  • If the process is part of an enterprise-wide digital transformation with potential for high financial impact, the business case for entirely redesigning for automation can be made more readily.
  • If the process is part of a department-sponsored initiative to reduce costs, you might consider making some small-scale process improvements first, including task-level automation. This will enable you to standardize, capture some improved performance data, and pave the way for more extensive reengineering using automation as a next step.

When weighing the merits of incremental changes versus fully fledged redesign, bear in mind that an automated workflow – done right, with appropriate orchestration, flexible hosting and scalable tools – is an agile one, ready to change as business conditions evolve. The question of whether to improve or automate won’t need to be answered again as the automation will drive the improvement.

A quick client example

We worked with a global automotive manufacturer to assess the feasibility of, and optimal approach for, introducing robotic process automation (RPA) into a back-office finance process. Process discovery highlighted some issues around the use of scanned images within the process that suggested optical character recognition (OCR) might deliver further improvements. Additionally, unstructured data was found to be inhibiting quality and flow, pointing to the need for a cognitive rules engine.

By identifying these two additional areas for improvement, we changed the business case for automating the process first. By adding two additional automation capabilities to RPA – OCR and a cognitive rules engine – the client could potentially quadruple the business benefit, necessitating a more significant process redesign than had previously been considered.

Note: The process discovery wasn’t a series of lengthy mapping workshops. It was a rapid, automated approach that helped uncover the complexity of the process, range of improvement opportunities and the potential impact of recommended technology solutions.

Three final thoughts

  1. Technology runs more efficiently on an optimized process.
  2. A low-tech improvement to rebalance workload or reduce process exceptions might be a good way to deliver an initial business benefit while grander plans are being made.
  3. Before you determine whether automation should be introduced with incremental improvements or as part of a radical redesign, get to know the process first.

To learn more about how other companies are approaching automation, take a look at our case studies at www.ibm.com/automation.

Upcoming ROI webcast

Building the business case for an automation platform solution

Join special guest, Sean Owens, Principal Consultant at Forrester, on 1 May 2019 at 1 PM EST. 

Sign up for the IBM Automation Insider

What automation opportunities are out there? How do you get results?
Every other month our experts share five pieces of strategic content to help you drive growth through automation.

One of the best places to start with AI

By John Davies

In the Forbes Insight article Intelligent Automation: How AI and Automation are Changing the Way Work Gets Done, KPMG predicts that overall investment in intelligent automation is expected to reach USD 232 billion by 2025, comparing that to an estimated USD 12.4 billion today.

I'm not surprised by that prediction as automation that doesn’t embed intelligence will be missing the automation opportunity. And the opportunity is three-fold:

  • Free up employees from low value tasks
  • Assist employees with high value work
  • Drive growth and new business models

But how do you actually do that – embed intelligence in automation? In February at Think 2019, we announced the in-development of IBM Business Automation Intelligence with Watson™ to help bridge the gap between business automation and AI and enable a new class of intelligent automation. At a high level, this solution is an automation capability for creating, managing and governing AI across the enterprise and applying it to business operations using IBM Watson™. It’s currently being built with the help of sponsored clients and will be available later this year.

Why do we think Business Automation Intelligence with Watson will be one of the best places to start with AI? It nets down to the four following reasons: 

1. It’ll address the top barriers to AI adoption:

  • Business people don’t always know how and where AI can be best applied to their problems.
  • AI algorithms are often disconnected from daily business operations.
  • AI is difficult for business people to trust, control and monitor.

Read the blog: How to remove the top 3 barriers to AI adoption in business automation

2. It’s being built around six foundational “moments of intelligence” that occur across all types of work – from clerical to knowledge work. By automating these, you can have a significant impact on customer experience and employee productivity:

  • Extracting information
  • Entering data
  • Gathering critical information
  • Making decisions
  • Prioritizing work
  • Anticipating action

You can think of these “moments of intelligence” as those basic, readily identifiable elements of work, related to tasks or skills, that can benefit the most from intelligent automation. Automating these moments of intelligence can free employees from low value work by offloading repetitive tasks (for example, extracting information and entering data) or assisting employees with high value work by augmenting expert work (like making decisions and anticipating action).

Automation agent

 

With IBM Business Automation Intelligence with Watson, you’ll be able to build automation agents to intelligently automate these fundamental tasks and skills with built-in business controls so you can trust the agents with this critical work.

3. It’s being designed to help you identify tasks where people spend most of their time. Because successful automation requires prioritizing what needs to be automated, this solution will help you do this by visualizing “hotspots” where automation can be applied.

4. Finally, it’ll access and act on the operational data generated by the IBM Automation Platform for Digital Business. If you’re a platform user, you’ll be able to apply AI across that operational data using this solution and potentially automate more than you could before.

When can you start using Business Automation Intelligence with Watson?  As mentioned earlier, this solution will be available later this year, but you can sign up for the early access program now. Your feedback will help make it the best it can be out of the gate.

Upcoming ROI webcast

Building the business case for an automation platform solution

Join special guest, Sean Owens, Principal Consultant at Forrester, on 1 May 2019 at 1 PM EST. 

Sign up for the IBM Automation Insider

What automation opportunities are out there? How do you get results?
Every other month our experts share five pieces of strategic content to help you drive growth through automation.