IBM Automation Insider

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

July/August 2019

5 articles


17 min

Four data quality steps you must take for AI success

By Melanie Turek

I’m going to go out on a limb and assume your organization is like most others when it comes to AI: you know what it is, you believe in its promise and you’re eager to see what it can do for your business. A recent Frost & Sullivan survey of 1,636 IT decision makers around the world confirms that assumption. We found that 63 percent of companies use AI and machine learning today, and 72 percent plan to up their investment over the next two years.*

There’s just one problem — most companies won’t achieve desired outcomes from AI if their data quality isn’t, well, quality. And, chances are, it isn’t.

Why AI craves good data

At its core, AI uses advanced algorithms and machine learning to better capture, process and act on information. Whether you’re using it in the contact center to improve customer and agent experience, on the production floor to optimize productivity and streamline your supply chain, or in the back office to speed decision making and drive innovation, AI needs good information, like our bodies need good calories, to operate at optimal levels.

Why most data quality is a hot mess

Most companies have had years, even decades, to develop a strong and secure data management system — for structured data. Now there are vast streams of unstructured data and terabytes of data that could be used by structured systems entering the organization from disparate sources like beacons, thermostats, smart cars and wearable tech. Things get messy fast, and the mess grows exponentially nasty — a hoarder’s dream, but a business nightmare. No wonder IT managers say the sheer volume of data is one of the biggest threats to their AI initiatives.

What you can do about it

There are four data quality steps you can take to increase your chances for AI success:

  1. Treat structured and unstructured data equally. AI is most useful when it can analyze a wide range of information — including text, audio and video — from a wide range of sources. But it’s also important to consider all the structured data coming into the enterprise, including inputs from sensors, beacons and the like.
  2. Eliminate the noise. This is one of the hardest, and most critical, elements of a modern data management system: determining what information is valuable, and what is flotsam turned up in the tides. Not all information is worth collecting, analyzing or storing. Make sure the data you’re gathering serves an identified and prioritized business purpose — and put in place metrics for measuring success.
  3. Pay close attention to privacy and compliance. If your AI system is scooping up data from a range of public and proprietary sources, some of that information might be subject to specific rules and regulations that don’t apply to the data you normally collect.
  4. Ensure everyone and every system has access to the data it needs, when it needs it. Information can serve multiple purposes. For instance, knowing that a customer is unhappy with a product serves that specific CX interaction, but it can also feed into product development, channel strategies and more. Take advantage of the advanced analytics AI offers across the organization by surfacing relevant data wherever it makes sense.

How all this data plays out for most companies today

Let’s take one scenario — using AI in the contact center to improve outcomes — and see how it plays out.

One of the biggest frustrations for many customers is explaining the same problem over again, every time they switch channels. They go to your website for help and see nothing in the FAQ. They launch a chat session, then send a follow-up email and finally make a phone call. They expect, or desperately want, your agent and organization to know they took those actions.

Easy, right? Your modern contact center software can offer an omnichannel experience that links all those interactions to the same customer for personalized service. Except, how does that system handle the email and phone call?

That data is unstructured. You need to tag it, classify it and contextualize it — in real time, while the customer waits for an answer. That involves everything from basic translation and voice recognition to advanced analytics that can contextualize key words and phrases for the customer- and agent-side of the conversation. Once that’s done, the system needs to store the data for future use, and then figure out how to surface it for any advanced data mining you might do to continue to improve processes.

And what if the customer reached out on Twitter or LinkedIn? What if the customer finally got exasperated and went to one of your physical stores to speak with a human? Chances are, those interactions are missed. (And, chances are, that customer is already lost to a competitor.)

Now consider the information your organization captures from sensors and beacons — from smart badges to wearable tech, building systems to biofeedback. Even if much of that information is structured, it’s often difficult to know where to put it and how to use it — not to mention where it fits in the overall security posture that requires stringent privacy and control over some (but not all!) information.

Exhausted? Overwhelmed? Absolutely. Fortunately, vendors are developing ways to handle this information overload, from vetting data at its source to applying algorithms that can improve Extract-Transform-Load (ETL) processes. But it’s still incumbent on you, your company, to take the necessary steps — and invest in the necessary technology — to ensure your data quality is ready for AI, or don’t bother.

* Source: Top End User Priorities in Digital Transformation, Global, 2019. To access go to frost.com (links resides outside IBM).

Note from the editor: If you need help digitizing, classifying and extracting structured and unstructured document content, IBM Content Analyzer may be worth consideration. Designed for business users, it's an intelligent, cloud-based API service you can use to train many document types in minutes with just one sample. Learn how the Bank of Montreal is using it.

Automation 2019, Toronto

On 10 Sep, join us to learn how companies are using automation and AI to satisfy customers, improve compliance, grow revenue and create new business opportunities.

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.

Bank of Montreal: Helping customers offload a tedious life task

By Cheryl Wilson

Despite the benefits of paying our bills on time, such as avoiding late fees or credit score dings, most of us struggle to pay them all on time, every time. The reasons range from forgetfulness to fear of seeing a low bank balance after payment.

Bankers have offered support like online bill pay for a while, but customers admit they still don’t feel in control. Different billers, amounts, accounts and due dates can create a cognitive load that works against successfully managing this life task. Any financial challenges or emotional baggage around money can make it worse.

The Bank of Montreal wanted to help its 8 million+ customers with this bothersome task, which is one of the most frequently used transactions in digital banking. 

In brief, the soon-to-be-launched BMO Quick Pay allows customers to forward their e-statements or photos of paper statements to BMO. All of that information is converted into data that’s automatically fed into their mobile banking engine, and then uploaded as preauthorized or future-dated bill payments. Quick and easy. Fire and forget it for customers.

  • Early results: 9 out of 10 customers said they’ll definitely adopt the service. And, in customer testing, bill payments were completed six times faster.

Given these positive, early performance indicators, we wanted to know more about this automation solution. The BMO Quick Pay team kindly answered the following questions for our newsletter subscribers:

Your early results are promising. Anything you want to share that would help other companies achieve early success?

We transitioned to an agile methodology and co-created this solution with our customers from the ground up. We took into consideration that their bill-paying journey starts with a statement, often a piece of paper, that they need to pay at the bank, online or via a mobile app. We wanted to offload that still-largely manual process of coordinating multiple bills, accounts, amounts and dates by automating it through IBM Business Automation Content Analyzer so customers could focus on other things.

We also needed to ensure a safe process for our customers. When we designed Quick Pay, we also vetted and co-created it with our fraud and risk partners.

The whole thing was a cultural shift for us. Previously, our check-ins and sign-offs with customers, fraud and risk partners happened at the end, after the solution was developed. This ground-up, diverse development process created a better, stronger solution.

Why IBM Content Analyzer?

There are new billers, thousands of them, created every year, so we needed the level of intelligence and machine learning that Content Analyzer provides to keep pace with the number and variety of forms coming through. It also has a highly mature ontology.

Another deciding factor: It was easy to train the predictive model and learn the user interface. No computer science experience required. The entire journey from purchase commitment to production provisioning was done in less than a week.  After provisioning, we were up and running on the same day.

Are you planning to use Content Analyzer for other use cases?

Given the market-leading nature of these solutions, the most we can say right now is yes. Looking ahead, there are more opportunities to come. With all the manual data we encounter every day, Content Analyzer can help parse it into concrete, digestible content that can be rapidly processed to improve any customer experience with automation.

If you want to look even further ahead, imagine waking up and not having to worry about your money or doing banking. Banking will be done for you automatically, alerting you if something needs your attention. That's the kind of banking experience we want to co-create with customers and partners. And automation technology is a key to that evolution.

Watch this 90-second video to hear Peter Poon, head of Digital Innovation and Channel Management at Bank of Montreal, talk about that value of BMO Quick Pay.

To learn more about IBM Content Analyzer, watch the demo video Easily capture and understand your documents using AI. (03:39)

Automation 2019, Toronto

On 10 Sep, join us to learn how companies are using automation and AI to satisfy customers, improve compliance, grow revenue and create new business opportunities.

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 ways to make your chatbots more likeable (and more likely to achieve KPIs)

By Jeff Goodhue

Have you ever met a chatbot you liked but wished it was smarter?

Wherever your favorite chatbot lives – website or mobile app – it's basically a digital program that uses AI to interact with you. Chatbots do their best to provide missing information or make suggestions based on your inputs. Flashback. Anyone remember one of the first chatbots commonly known as Clippy? (10:59)

No surprise, not all chatbots are helpful. They might not have the information you need when you need it or the conversation flow might be awkward and slow. These issues can negatively affect customer perception, Net Promoter Score (NPS) and achievement of key performance indicators (KPIs), such as sales conversion or case completion time.

Good news: You can increase chatbot likeability and improve time to customer value with support from business automation. The following three patterns combine chatbots and specific business automation tools to make your chatbots smarter and faster:

1. Combine chatbots with AI-driven document capture so they can answer more questions

To design a personalized chatbot that anticipates your user's needs, it needs data from your user as they type into the chatbot interface and data from unstructured sources.  For example, what if your banking chatbot knew you submitted an application form and could automatically extract the form’s values, looking up the application’s status from another system? That would be useful.

Figure 1. General pattern combining a chatbot with intelligent content analysis

Figure showing general pattern combining a chatbot with intelligent content analysis

If you're a business analyst or developer, check out Content Analyzer’s Get Started page to see how easy it is to configure document classification and extraction of values and paragraphs.

2. Combine chatbots with business rules and Natural Language Understanding (NLU) to improve conversation quality

Building a primary chatbot conversation flow normally goes quickly because many of the branches and leaves can be enumerated, including the rule conditions that control flow. 

But as chatbots take on more responsibilities, onboard new business areas and increase personalization, two functional areas become more important: NLU and business rule management.

  • NLU: If your chatbot needs to understand more complex user intents, such as longer sentences and paragraphs, or unstructured documents, AI-based NLU should be used. A continuous training loop that governs and tracks accuracy of chatbot responses should also be considered.
  • Business rule management: If your chatbot needs to respond more personally with specific recommendations and next-best actions, business rules should be used to set up and control those types of decisions. For example, rules can recommend a new product based on the current conversation and the user’s likelihood to buy, or next-best action rules can determine the best promotion based on the customer’s retention score.

If you’re an architect or developer, check out this reference architecture to see an example of this pattern in action.

Figure 2. General pattern combining a chatbot with declarative and predictive decisioning

Figure showing general pattern combining a chatbot with declarative and predictive decisioning

3. Combine chatbots with RPA bots and processes so they act faster

Once your chatbot can answer more questions, more personally, your users will probably want it to act faster.

Imagine your user wants to change an address, purchase a subscription plan or open a new account, all of which require task and process automation. The chatbot could tell her where to go, leaving her to complete the task.  A smarter chatbot could launch the process for her and provide a status, which requires integration into robotic process automation (RPA) bots and business processes. If your chatbot can only repeat FAQs and provide personalized responses but not take action, the experience is nothing more than interactive help documentation.

Figure 3. General pattern combining a chatbot with task and process automation

Figure showing general pattern combining a chatbot with task and process automation

To wrap up, a chatbot is only as good as its last conversation with you, like a restaurant is only good as its last meal served. To keep customers coming back to your site or app, chatbots can be a key differentiator. The more likeable, the more helpful, the better.

Note: The capabilities highlighted within these patterns are part of the IBM automation software platform that enables you to automate any type of work at scale, but you can find them as single tools, too.

Automation 2019, Toronto

On 10 Sep, join us to learn how companies are using automation and AI to satisfy customers, improve compliance, grow revenue and create new business opportunities.

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.

“Why do I need an all-in-one automation platform?”

By Brian Safron

Imagine your VP’s reaction when you break the good news. “We’ve decided to work with several automation software vendors. Each offers best-of-breed software that doesn’t interoperate with the other vendors’ software. And we’ll need to hire more people to support and manage each of the solutions.”

It’s hard to imagine a VP being thrilled about needing to integrate, deploy and support a random collection of automation software — or having to build and maintain unique skills for each offering.

What would thrill a typical VP in that position? Is it the ability to drive digital transformation by automating as much of the business as possible — as fast as possible — while avoiding a proliferation of automation solutions that demand more people, platforms and skills? Probably. But what’s the best solution for achieving that?

When it comes to automation software, you can choose from four basic options (for more context, see my publication The quick and practical guide to digital business automation, PDF, 294 KB):

1. You can write all code from scratch.

  • Pros: You own and control everything.
  • Cons: The software becomes a black box, where the business side doesn’t have any visibility or understanding of the code. This option requires a lot of IT time and expertise to understand what the code does and to make any changes.

2. You can buy packaged apps.

  • Pros: This option is ready made, and some apps will fill specific needs.
  • Cons: If your business doesn’t fit the mold of the packaged app, it’s not going to do all the things you want it to do. Packaged apps don’t have a lot of flexibility, so you have to work within their limits.

3. You can buy a collection of point solutions.

  • Pros: You can select from a wide variety of vendors.
  • Cons: You act as the integrator as you buy different automation applications from different companies. The products don’t always work well with each other, and there’s no underlying foundation for things like analytics, machine learning and AI.

4. You can adopt an automation platform.

  • Pros: This option is an integrated set of foundational applications and cross-platform capabilities with which you can customize any automation solution.
  • Cons: You don’t get to choose a different vendor for every application. With this option, you’re dependent on a single vendor for support across the platform.

If you write all your code from scratch, you’ve accepted extreme complexity because you have extreme needs. In terms of packaged apps, you may be using them where appropriate, but there are many things packaged apps don’t do well, especially when you’re trying to differentiate from your competition.

Right now, we’ll focus on the fourth option – an all-in-one automation software platform – and explain why you might need one of these.

The value of an automation platform

The ideal automation software would provide a repeatable way to build custom solutions while reducing the time and resources needed to build those solutions. An integrated platform can help drive revenue and new business opportunities while avoiding the hidden costs of point solutions.

Driving revenue and new business opportunities

In a recent analysis commissioned by IBM, The Total Economic Impact (TEI) of the IBM Automation Platform for Digital Business, Forrester highlights how standardizing on an automation platform has increased revenue and business flexibility at a large US-based bank.

Revenue growth. Forrester analyzed the financial benefits of using an automation platform; in this case, the IBM automation software platform.

According to Forrester’s evaluation, one percent more loans were closed due to faster loan reviews and approvals enabled by the platform. While one percent may not sound like a huge number, the incremental revenue driven by that increase – USD 16 million in a three-year period – is significant.

Even after Forrester adjusted for risk, the additional revenue just from faster loan processing was more than USD 13 million in three years. For the four automation projects included in the evaluation, Forrester calculated a return on investment (ROI) of 675 percent.

Flexibility. In addition to quantifiable financial results, Forrester highlighted increased business flexibility driven by the automation platform. They noted: “The bank’s standardization on the automation platform will help enable new automation initiatives to be added without the significant expense of deploying and customizing a new automation solution, thereby helping the bank unlock many more future opportunities than it expected.”

Avoiding the hidden costs of point solutions

In the same TEI study, Forrester shared the following 10 factors that may come into play when choosing and deploying a new automation solution:

1.   Searching for vendors that offer the right solution

2.   Gathering requirements to create a request for information (RFI) to send out to vendors

3.   Forming a committee to narrow down RFIs to a list of finalists

4.   Creating criteria and bringing in each vendor to perform a proof of concept (POC)

5.   Issuing request for proposals (RFPs) and selecting a vendor

6.   Negotiating pricing, finalizing an agreement and starting the procurement process

7.   Planning for integration with existing environments

8.   Identifying issues, risks or software vulnerabilities and documenting anything that needs remediation

9.   Standing up environments, such as development, test and production

10. Hiring consultants to help with the first project and learn best practices

Standardizing on an automation platform can help you avoid spending time and money on each of these steps for each new business solution.

According to Forrester’s TEI study, “Setting and communicating automation platform standardization across the organization has allowed the bank to avoid spending time and money on research and implementation for one department or another — as well as the time and cost of integrating as many as 10 disparate automation systems together.” As the bank explained, “We have one standard platform we can keep building on, versus having to figure out every project with analysis of every vendor.”

The best of both worlds: The evolution of the IBM automation software platform

There was a time when organizations had to choose between full-featured enterprise software and narrowly focused point solutions. A platform can solve that problem by providing integrated, best-of-breed capabilities across the full spectrum of automation.

The scenario at the beginning of this article — with the less-than-thrilled VP — dramatizes the client problem IBM Automation addressed when we began architecting our automation platform in 2016. We decided the platform needed five core automation capabilities to enable digital transformation – tasks, workflow, decisions, content and capture (see Figure 1). And they needed to be integrated.

The technologies represented by these capabilities include:

  • Enterprise content management
  • Business process management
  • Case management
  • Business rules
  • Digital decisioning
  • Data capture
  • Robotic process automation

In addition to these core areas, IBM Automation focused on several cross-platform capabilities that are also critical for digital transformation:

  • Analytics
  • Machine learning
  • AI
  • Business modeling
  • Low-code/no-code authoring
  • Governance
  • Unified user experience
  • Hybrid, flexible deployment options

As highlighted in the last bullet, this platform would run everywhere. We would host it on IBM Cloud™, or it could be hosted on any cloud or in any data center.

Illustrating how IBM Automation Platform for Digital Business can provide integrated, best-of-breed capabilities across the full spectrum of automation

Figure 1. IBM Automation Platform for Digital Business

We recently received positive feedback in The Forrester Wave: Software For Digital Process Automation For Deep Deployments, Q2 2019 — achieving the highest score for strategy among evaluated vendors.

Forrester noted: “IBM has consolidated its content management, decision management, and process automation offerings under a single executive and engineering team with a unified go-to-market execution. At the same time, it has done some of the most pragmatic integration of IBM’s Watson AI capabilities to drive very process-specific business value.”

The last word

The future of work will be based on a spectrum of integrated automation capabilities, supported by embedded machine learning and AI, that enable people, robotics and computer systems to “divvy up” work in the smartest and most optimized way. As a result, knowledge workers can focus on higher value work and businesses can serve their customers better and more efficiently.

An automation platform is one of the best ways to provide the full spectrum of capabilities required to industrialize the digital transformation needed to stay competitive in a hypercompetitive world.

Automation 2019, Toronto

On 10 Sep, join us to learn how companies are using automation and AI to satisfy customers, improve compliance, grow revenue and create new business opportunities.

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.

Automate any type of work on any cloud platform

By John Davies

You’re probably using three or more clouds to run your businesses like a lot of companies these days. And the reasons for doing so vary — reducing operating costs, avoiding dependency on one provider, improving customer experience.

When you’re trying to quickly build and run automation applications and services, deployment flexibility is a competitive advantage. Beyond having your choice of essential automation capabilities, you want the freedom to deploy and run them in your choice of environment: your cloud, managed cloud, on premises or some combination of the three.

Now, with the introduction of IBM Cloud Pak™ for Automation, you can build and deploy automation applications on any cloud using an integrated and managed collection of containerized business automation services.

IBM Cloud Pak for Automation provides the following four advantages:

1. You can deploy and run on any cloud platform. IBM Cloud Pak is certified for Red Hat® OpenShift® on IBM Cloud™ – and runs on any Kubernetes.

  • Avoid vendor lock-in.
  • Run on one or more private or public clouds.
  • Migrate existing automation runtimes without application changes or data migration.

2. You can see and measure the performance of your business operations.

Wherever IBM Cloud Pak automates work, it also collects and consolidates large amounts of data for real-time operational performance visibility.

  • Business managers can monitor performance across the entire platform – not just at an individual product or process level – using predefined or user-configured dashboards.
  • Machine learning and AI can then be applied to your pre-curated operational data to derive recommendations for process and decision optimization.

3. You can deploy faster.

With IBM Cloud Pak for Automation, you get essential, pre-integrated automation software and low-code tools to help you quickly develop modern applications – at scale. Deployment is simplified by having:

  • Common governance and infrastructure management
  • Topology-aware containers that simplify the operation and administration of automation capabilities and common configuration
  • Multicloud support for scaling up or down quickly and moving between clouds for cost savings

4. You can buy it easily.

IBM Cloud Pak for Automation is one flexible package with simple, consistent licensing. You only pay for what you need at the time. You have the flexibility to reallocate licenses to other platform capabilities in the future.

  • Choose perpetual or monthly pricing.
  • Buy new, exchange, or trade-up from eligible stand-alone entitlements.

In sum, IBM Cloud Pak for Automation is our new containerized automation software platform with pre-integrated capabilities such as workflow and decision automation, content management, document processing and operational intelligence. It empowers business users to rapidly deliver applications at enterprise scale for greater cost savings and operational efficiencies. It also integrates with IBM Robotic Process Automation with Automation Anywhere to automate highly repetitive tasks.

To see how to easily deploy IBM Cloud Pak for Automation, watch this demo video. (04:31)

Automation 2019, Toronto

On 10 Sep, join us to learn how companies are using automation and AI to satisfy customers, improve compliance, grow revenue and create new business opportunities.

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.