Loading...

Follow WorkFusion Blog on Feedspot

Continue with Google
Continue with Facebook
or

Valid

WorkFusion recently hosted an informative webinar on the opportunities and benefits of intelligent document processing (IDP) — featuring Anil Vijayan of Everest Group and Arnesh Sahay, a Solutions Engineer for WorkFusion, and moderated by Veronika Andreeva, Product Marketing Manager for WorkFusion.

More than 300 people have attended the free webinar so far, and it’s still available on-demand at: http://learn.workfusion.com/2019-may-intelligent-document-processing-webinar-ft-everest-group

The 30-minute presentation was followed by a Q&A, and many attendees had additional questions that we wanted to answer with some detail here.

Scroll for the answers to these topics:
  • How are ML/NLP algorithms utilized in WorkFusion’s IDP solution?
  • What does the initial training for a WorkFusion ML model look like? After an initial model is trained, how does the software perform quality control against the model’s output?
  • How does WorkFusion’s IDP solution solve for high volumes of data?
  • Can IDP be used to read data from emails? Can we schedule bots to run at varying times for different processes?
  • What does a false positive mean in the context of machine learning processes?

How are ML/NLP algorithms utilized in WorkFusion’s IDP solution?

WorkFusion Process AutoML is a proprietary technology for Cognitive Automation. AutoML automatically finds the best machine learning algorithm and method of automation for myriad use cases. There are several different open-source algorithms leveraged in this process, depending on the type of model being trained.

In machine learning, a hyperparameter can consist of annotators, feature extractors, and various post-processors that are used to control the learning process. In order to produce the highest-performing machine learning models for any given use case, WorkFusion Process AutoML performs a series of experiments to test different “optimal” hyperparameter configurations by systematically varying one or more components and examining its effects on the result. This process is called hyperparameter optimization.

The two most common use cases deployed in WorkFusion software include information extraction: gathering data from unstructured documents to enter it into systems; and classification: classifying documents, transactions and emails into workflows. AutoML enables operations teams to train their own machine learning models without the need for a data scientist.

AutoML SDK goes further and allows Machine Learning Engineers (MLEs) to step in and extend machine learning models for any information extraction or classification use cases. This SDK consists of the following components, each responsible for a specific task in a pipeline:

Parser — remove HTML tags from documents, create an AutoML SDK Document
Annotator — split text into tokens, add boundary elements, add Named-Entity Recognition (NER)
Feature Extractors — analyze tokens in documents and create features
Algorithms — aggregate all field documents and feature extractors, provide a model and its statistics
Post-Processing — apply normalization logic and field grouping

Some examples of machine learning algorithms that are supported in AutoML SDK include Support-Vector Machines (SVM), Deep Neural Networks (DNN), Regression, etc.

What does the initial training for a WorkFusion ML model look like? After an initial model is trained, how does the software perform quality control against the model’s output?

The amount of data needed in the first iteration of training for an ML model depends on the complexity of the use case. An efficient information extraction model for a Prescription Intake use case can be created with as few as 50 documents, whereas a complex multi-class classification model for an Email Intake use case may require more than 100 documents per category. With our enterprise customers, we typically see that ML models are trained on at least 3,000–5,000 documents before they are deployed to Production.

Automatic Quality Control (AutoQC) refers to the use of statistical methods in monitoring and maintaining the quality of products and services. In a WorkFusion automation workflow, the AutoQC sub-process chooses an optimal, cost-effective combination of automated machines and cloud workers that always delivers at or above the acceptable quality level.

The main concept of AutoQC is to take samples from a defined batch of items (records, documents, etc.) and verify the quality of each individual sample. After sample verification, the whole batch is considered as accepted or rejected depending on the Rejection Limit parameter set in the process. Therefore, AutoQC enables customers to maintain a continuous quality check without needing to verify the entirety of the total batch.

Can IDP be used to read data from emails? Can we schedule bots to run at varying times for different processes?

Email Intake is a standard use case that WorkFusion has solved for several different customers. The data within emails (subject, body, attachments) are all parsed via RPA; any attachments are sent to the Optical Character Recognition (OCR) sub-process for conversion to a machine-readable format (XML/HTML); the body of the email is merged with the converted attachments; and a single document is sent to the AutoML process. The case study below outlines how WorkFusion tackled an Email Intake process at a large U.S. bank which was inundated with massive email volumes and riddled with inefficiencies due to human error. (View our Email Intake Case Study to learn more.)

Scheduling bots is a simple and straightforward process in WorkFusion’s Control Tower. The image below shows how process owners can create their own schedules for various processes to kick off at designated times. This is a typical view in Control Tower:

What does a false positive mean in the context of machine learning processes?

When evaluating the performance of machine learning models, we need to be able to differentiate between values that were identified correctly and incorrectly. When looking at a Binary Classification use case, a correctly classified document can be either a True Positive (TP) or a True Negative (TN). Alternatively, an incorrectly classified document can be either a False Positive (FP) or a False Negative (FN).

The definitions below pertain to a simple HotDog/NotHotDog use case, as presented on HBO’s “Silicon Valley” and help illustrate the differences between the four possible outcomes:

TP — Image is a hot dog and the model identified it as a hot dog. (Correct)

FN — Image is a hot dog, but the model identified it as not a hot dog. (Incorrect)

TN — Image is not a hot dog, and the model identified it as not a hot dog. (Correct)

FP — Image was not a hot dog, but the model identified it as a hot dog. (Incorrect)

A binary classification model is relatively simple to calculate for. But when evaluating a more complex multi-class classification model, confusion matrices are often used to map out the four possible outcomes against each possible category, to better understand the performance of a model.

How does WorkFusion’s IDP solution solve for high volumes of data?

WorkFusion’s Enterprise Architecture is designed to handle high volumes of data coming in across multiple different business processes. To address the increased load on the systems, single instances can be scaled either horizontally or vertically. Horizontal scaling consists of provisioning more servers for ML, OCR, and RPA as needed to respond to increased volume growth, and vertical scaling refers to dedicating more powerful hardware to each component.

As different business lines within an organization start using WorkFusion, it is important to keep data segregated across different departments. There are two approaches to address this: One is to instantiate new environments that correspond to business units as they are onboarded and have sufficient volume (this is the most common approach); another is to set up a shared multi-tenant instance with scaled ML, OCR and RPA components.

Related: Learn more about a recent Everest Group assessment of leading IDP vendors across the industry and/or download a complimentary licensed copy of the report.

If you have any questions about this webinar content, or if you’re curious how well-suited Intelligent Automation would be for your transformation program, we welcome your inquiries anytime at learn@workfusion.com.

Complimentary research report: Intelligent Automation — Accelerating From Short-term Wins to Long-term Strategic Business Outcomes

IDP Webinar Follow-up: We Answer 6 Bonus Audience Questions was originally published in WorkFusion on Medium, where people are continuing the conversation by highlighting and responding to this story.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 
New Everest Group research explains: For long-term success, companies need Intelligent Automation solutions with integrated AI

Automating purely to reduce cost isn’t transformation. It’s optimization. That’s fine if you’re only interested in short-term wins, but leading enterprises — the ones who beat born-digital competitors, keep customers and delight shareholders — are looking beyond cost reduction and envisioning long-term success.

Led by distinguished analyst Sarah Burnett, the team at Everest Research Group set out to help enterprise executives and automation practitioners understand the landscape and its options. This result is this recent report: “Intelligent Automation: Accelerating from Short-term Wins to Long-term Strategic Business Outcomes: A Guide to Undertaking Automation-led Business Transformation.” A complimentary licensed copy of the report is available for download here, or read on for key points and our responses:

  1. How is Intelligent Automation different from traditional RPA?

Robotic process automation (RPA) follows rules to automate work that has no variation. When you log into your email account, you enter a username and password the same way every time. RPA is great for these types of repetitive tasks, which often string together to form a simple business process, like: log in, click box, move file from Point A to Point B, log out.

Scalability and ROI problems rapidly emerge, however, when variation is introduced. How often does your business process change? How much of the high-volume work that inundates your team involves unstructured data? People adapt, but software bots that only follow rules do not.

This is why AI-driven Intelligent Automation is superior to rules-driven RPA. Intelligent Automation (IA) integrates all the capabilities found in RPA, plus adds capabilities only possible through bots that learn and adapt to data in real time. According to the Everest Group report, these are some critical features of what they call “RPA 4.0” or Intelligent Automation:

  • “Machine learning, computer vision, text analytics and NLP” to process unstructured data, automate tasks that require judgment, and detect and adapt to constant change
  • “Predictive and prescriptive analytics” help leaders plan resources and set achievable KPIs, plus help ensure optimal operational results
  • “Self-managing, self-healing robots” to handle exceptions and reduce bot management

“In order to build a sustainable competitive advantage through digital transformation, it is essential for enterprises to adopt a long-term automation strategy that aims to implement intelligent automation solutions that combine both RPA and AI capabilities.”

2. How many companies are using IA vs. traditional RPA?

Everest Group takes great care to understand technology, and its research team also reports on how the smartest enterprises apply it. “Pinnacle Enterprises” is a term they use to distinguish long-term strategy focused vs. short-term, tactically oriented businesses. (Read more about it here.)

In a comprehensive market survey, Everest Group found that among Pinnacle Enterprises:

  • 44% are “in piloting stages for cognitive / AI” vs. only 9% of other enterprises
  • 45% are “currently in planning stages for cognitive / AI” vs. 49% of others
  • 11% are “currently not considering it” compared to 33% of others

3. How are the benefits different?

Especially in banking, finance and healthcare companies, regulatory compliance has eroded margins, and leaner, fast-moving digital-native competitors have more room for profit variability. Reducing cost is a critical benefit, no matter how strategic an enterprise is.

The difference between IA and RPA in term of cost reduction is simple: AI automates more and automates more efficiently because it processes unstructured data, handles exceptions and continuously learns. In short: IA delivers significantly greater cost reduction.

“Siloed RPA implementations in disparate business processes can provide short-term wins for enterprises around immediate cost reduction (mostly in transactional processes) but may not dramatically improve business outcomes.”

Also: Traditional, rules-driven RPA doesn’t meaningfully reduce human effort. It just changes it. Analysts may eliminate repetitive work, but then have to spend time managing and re-training RPA bots. IA improves employee engagement by reducing both.

Compliance is another concern in these sectors. As Everest explains in the report, “the compliance functions of enterprises from regulation-heavy verticals such as banking, insurance, and healthcare go through multi-step repetitive processes and need to navigate through complex IT architecture.” Navigating this complexity and providing the depth and detail of reporting that regulators mandate is beyond the capability of standalone RPA products. IA helps complex organizations “ensure adherence to quality and compliance.”

And let’s not disregard the customer experience, the factor most responsible for driving revenue. By automating many middle- and back-office tasks connected to customer queries, companies can lower the “average handle time” (AHT), process requests with greater accuracy, and create higher customer satisfaction.

4. Why is an integrated platform superior?

Buyers of automation products who understand the importance of both AI and RPA to delivering superior results have two choices: 1) point solutions that require integration of RPA and AI through APIs and web services, or 2) a single platform that natively integrates all of the capabilities required to automate an end-to-end process or an entire role.

Everest thoughtfully presents the pros and cons of both approaches.

Some of the benefits attributed to the first approach include “reduced single vendor dependency” and “access to a wider partner ecosystem.” In our opinion, these both have their own down sides: more vendor management and more service checks to write.

All the cited challenges of this ecosystem approach are unambiguously negative: “increased complexity in integration,” “potentially higher TCO,” “potentially higher license costs,” and “governance & traceability challenges with multiple vendors,” among others.

On the other hand, Everest Group says the benefits of buying a single integrated platform that provides all critical capabilities are intuitive: “ease of interoperability,” “scalability,” “reduced license cost and TCO,” “native capabilities can be optimized to work well together.” To distill it down, buying a single platform is better, cheaper and faster. It removes from internal IT the archaic burden of integrating disparate applications, which is ironically why RPA was originally conceived. Even more importantly, the data generated from each capability is harmonized into robust analytics that help leaders manage and plan productivity.

Have any questions about the report? Or are you curious how well-suited Intelligent Automation would be for your transformation program? WorkFusion’s Intelligent Automation Cloud offers intelligence, simplicity and scale. We welcome your inquiries at learn@workfusion.com.

If You’re Only Automating with RPA, You’re Playing the Short Game was originally published in WorkFusion on Medium, where people are continuing the conversation by highlighting and responding to this story.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

Robotic process automation (RPA) is one of the fastest-growing sectors in business, but it can deliver even more benefit to users with an injection of breakthrough tech. Enter intelligent document processing (IDP), which uses machine learning (ML) to capture, classify and extract the most difficult to automate data: unstructured. IDP fills the wide capability gap left by rules-based RPA, which pulls and pushes structured data — and this capability is finally getting the attention it deserves as an enterprise technology superpower.

Everest Group is among the first firms to map the market and recognize its leaders, and in their latest assessment of the landscape, they have designated WorkFusion a Leader based on Market Impact and Vision & Capability.

We’ve licensed the report — Everest Group’s Intelligent Document Processing PEAK Matrix™ 2019 — which includes the full vendor evaluation of WorkFusion. Download a complimentary licensed copy here.

Read on for highlights of the report and why WorkFusion was spotlighted.

What is intelligent document processing (IDP)?

According to industry research, about 90% of data within any organization is unstructured, and most of it is locked in documents. Everest defines documents as “email, text, PDF and scanned documents” — which first-generation, rules-based RPA cannot process. IDP has emerged as a critical capability for large organizations because RPA customers have struggled to scale automation programs, most of which require processing a large volume of documents. With IDP, operations are able to automate data capture, classification and extraction.

Why is WorkFusion a Leader out of 16 IDP vendors?

Everest Group defines Leaders as those vendors who “command higher overall satisfaction among their clients than Major Contenders and Aspirants. Leaders’ clients are highly satisfied with their product vision & roadmap, IDP product capabilities, and their ability to help them achieve better business outcomes.” Looking beyond the way our products delight our customers, Everest Group offered these main reasons for assessing WorkFusion at this level:

1. A unified platform combining core automation capabilities

In Everest Group’s words, “WorkFusion has a distinctive vision to enable digital transformation for enterprises by providing innovative AI-driven automation capabilities … within its integrated automation platform that brings together RPA, BPM, OCR, analytics, and AI capabilities.”

Yes, it’s a lot of acronyms, but each one of them is a critical capability in enterprise digitization initiatives, and it’s essential they all work in concert. Put simply:

  • IDP automates extraction and structuring of document-based data
  • RPA automates pulling and pushing of structured data into disparate systems
  • BPM orchestrates the business process that contains many data tasks
  • AI makes cognitive decisions based on the data
  • Analytics says how a process is performing, how it will perform, and what to do to optimize

WorkFusion has created the best iteration of each capability and built them into a unified software platform so that rather than cobble together multiple point solutions, customers only have to buy and learn one great product.

2. Making AI easy and scalable for business people

Not too long ago, AI was complex, expensive and scarce. WorkFusion was founded to make AI easy for business people — and the success that leading banks, insurance companies and healthcare organizations have seen using our product suggest we’re meeting that goal.

How? The key is a proprietary native capability called AutoML, which automates the expensive, complex data science work that used to require a team of Ph.Ds to spend months cleansing data, selecting and testing different machine learning models. AutoML watches people do their jobs, finds and learns patterns, and selects and trains the right algorithm to automate the work… within days.

3. Pre-built use cases and user-friendly process design tools

One example of a pre-built package: Anti–money laundering (AML) efforts are complex and expensive, but necessary to ensure banking customers are not performing illegal financial activity. Along with two major services partners, WorkFusion decided to create an off-the-shelf AI-driven automation solution for AML.

Everest Group noticed and rewarded this innovation, writing that our platform “has the capability to leverage pre-built use cases such as invoice processing and anti–money laundering [AML] to quick-start automation or configure a new use case through its built-in visual BPM/workflow with a simple drag-and-drop interface. This provides enterprises with the flexibility to enable end-to-end process automation with RPA, OCR, IDP, and system tasks with human-in-the-loop to handle exceptions.”

Versatile, complete enterprise automation solution

RPA, OCR, BPM and AI have each been around for a while, but our breakthrough automation capability has combined and delivered them as a unified product for customers. This is truly versatile intelligence: making many different tasks in a process more accurate, more measurable and more predictable. Entire functions are not only automated, but also perform better, faster and more efficiently as volume and variability of work increases. Our recognition by Everest Group as a leader in IDP, in two significant dimensions (Market Impact and Vision & Capability) is proof that one powerful, versatile platform that creatively combines capabilities can outperform specialized, rigid point tools. Our vision since 2012 has been for our customers, and we’re grateful to every company who has bet on us and helped us blaze this trail — and those who will continue to leap forward with us.

Intelligent Document Processing, the New Enterprise Technology Superpower was originally published in WorkFusion on Medium, where people are continuing the conversation by highlighting and responding to this story.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

Robotic process automation (RPA) is one of the fastest-growing sectors in business, but it can deliver even more benefit to users with an injection of breakthrough tech. Enter intelligent document processing (IDP), which uses machine learning (ML) to capture, classify and extract the most difficult to automate data: unstructured. IDP fills the wide capability gap left by rules-based RPA, which pulls and pushes structured data — and this capability is finally getting the attention it deserves as an enterprise technology superpower.

Everest Group is among the first firms to map the market and recognize its leaders, and in their latest assessment of the landscape, they have designated WorkFusion a Leader based on Market Impact and Vision & Capability.

We’ve licensed the report — Everest Group’s Intelligent Document Processing PEAK Matrix™ 2019 — which includes the full vendor evaluation of WorkFusion. Download a complimentary licensed copy here.

https://www.workfusion.com/everest-idp-peak-matrix-2019/

Read on for highlights of the report and why WorkFusion was spotlighted.

What is intelligent document processing (IDP)?

According to industry research, about 90% of data within any organization is unstructured, and most of it is locked in documents. Everest defines documents as “email, text, PDF and scanned documents” — which first-generation, rules-based RPA cannot process. IDP has emerged as a critical capability for large organizations because RPA customers have struggled to scale automation programs, most of which require processing a large volume of documents. With IDP, operations are able to automate data capture, classification and extraction.

Why is WorkFusion a Leader out of 16 IDP vendors?

Everest Group defines Leaders as those vendors who “command higher overall satisfaction among their clients than Major Contenders and Aspirants. Leaders’ clients are highly satisfied with their product vision & roadmap, IDP product capabilities, and their ability to help them achieve better business outcomes.” Looking beyond the way our products delight our customers, Everest Group offered these main reasons for assessing WorkFusion at this level:

1. A unified platform combining core automation capabilities

In Everest Group’s words, “WorkFusion has a distinctive vision to enable digital transformation for enterprises by providing innovative AI-driven automation capabilities … within its integrated automation platform that brings together RPA, BPM, OCR, analytics, and AI capabilities.”

Yes, it’s a lot of acronyms, but each one of them is a critical capability in enterprise digitization initiatives, and it’s essential they all work in concert. Put simply:

  • IDP automates extraction and structuring of document-based data
  • RPA automates pulling and pushing of structured data into disparate systems
  • BPM orchestrates the business process that contains many data tasks
  • AI makes cognitive decisions based on the data
  • Analytics says how a process is performing, how it will perform, and what to do to optimize

WorkFusion has created the best iteration of each capability and built them into a unified software platform so that rather than cobble together multiple point solutions, customers only have to buy and learn one great product.

2. Making AI easy and scalable for business people

Not too long ago, AI was complex, expensive and scarce. WorkFusion was founded to make AI easy for business people — and the success that leading banks, insurance companies and healthcare organizations have seen using our product suggest we’re meeting that goal.

How? The key is a proprietary native capability called AutoML, which automates the expensive, complex data science work that used to require a team of Ph.Ds to spend months cleansing data, selecting and testing different machine learning models. AutoML watches people do their jobs, finds and learns patterns, and selects and trains the right algorithm to automate the work… within days.

3. Pre-built use cases and user-friendly process design tools

One example of a pre-built package: Anti–money laundering (AML) efforts are complex and expensive, but necessary to ensure banking customers are not performing illegal financial activity. Along with two major services partners, WorkFusion decided to create an off-the-shelf AI-driven automation solution for AML.

Everest Group noticed and rewarded this innovation, writing that our platform “has the capability to leverage pre-built use cases such as invoice processing and anti–money laundering [AML] to quick-start automation or configure a new use case through its built-in visual BPM/workflow with a simple drag-and-drop interface. This provides enterprises with the flexibility to enable end-to-end process automation with RPA, OCR, IDP, and system tasks with human-in-the-loop to handle exceptions.”

Versatile, complete enterprise automation solution

RPA, OCR, BPM and AI have each been around for a while, but our breakthrough automation capability has combined and delivered them as a unified product for customers. This is truly versatile intelligence: making many different tasks in a process more accurate, more measurable and more predictable. Entire functions are not only automated, but also perform better, faster and more efficiently as volume and variability of work increases. Our recognition by Everest Group as a leader in IDP, in two significant dimensions (Market Impact and Vision & Capability) is proof that one powerful, versatile platform that creatively combines capabilities can outperform specialized, rigid point tools. Our vision since 2012 has been to for our customers, and we’re grateful to every company who has bet on us and helped us blaze this trail — and those who will continue to leap forward with us.

Intelligent Document Processing, the New Enterprise Technology Superpower was originally published in WorkFusion on Medium, where people are continuing the conversation by highlighting and responding to this story.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

To help financial institutions understand why Intelligent Automation solves problems that earlier RPA solutions couldn’t handle, IDC’s Sneha Kapoor has published an insightful report that defines cognitive automation and makes a compelling case for why it’s a game-changing leap forward for enterprises with aggressive transformation and cost-reduction initiatives.

Notably, this report — “Robotic Process Automation Game Changers Advance Financial Services Institutions Toward Intelligent Digital Workforce” — takes a firm stand on how to define “intelligence” in automation software. It’s a strong validation for operations who choose AI-driven automation products like WorkFusion, and it’s a great guide for those who are transitioning from first-generation RPA to Intelligent Automation.

Let’s consider IDC’s six defining characteristics of the intelligent digital workforce.

1. Simple, usable and reusable by business users

Before RPA, integrating applications required IT expertise, and lowering the cost of operational work required business process outsourcing (BPO). An intelligent digital workforce gives business people— the people in charge of functions like customer onboarding, trade settlements, and anti–money laundering efforts (AML) — the power to automate, on their own.

Intelligent Automation software with enhanced ML and QC capabilities automates cleansing data and training machine learning models. The IDC report mentions “improving data quality, redesigning processes and automation workflows, interoperability and integration.” A unified platform approach simplifies integration of critical tools such as OCR, BPM and rules engines.

2. Ability to deliver enterprise-wide scale

IDC acknowledges the “rapidly growing demand for an automation solution that can offer a single unified platform with centralized view and management of enterprise-level automation across various IT systems and technologies.” In other words, neither end users nor executive sponsors want rogue projects, cobbled together with disparate tools. To ensure automation success at a business-wide level, the smartest businesses select one primary platform to automate cross-operational functions and generate robust predictive and prescriptive analytics.

3. Security and governance as foundational tenets

IDC highlights the importance of “institutions’ ability to address the issues around data quality, data usability and data governance.” The inability of first-generation RPA products to monitor data quality is one of the biggest reasons complex businesses have struggled to scale their automation programs. Also, no one should overlook the growing challenge of regulatory compliance — essential for banks, insurers and healthcare providers. Intelligent automation products with native OCR, AI and RPA can process documents at massive scale. They also provide audit trails and explainable algorithmic decisions, which let compliance teams give regulators proof of success and comprehensive documentation of process.

4. Availability of real-time operational analytics

To measure ROI and achieve strategic resource planning, it’s critical to know both the performance of individual bots and of the functions they automate. Enterprise customers who choose products with unified platform models — i.e., “all-in-one” products — can see real-time performance and predicted performance of bots, people and functions. The latest generation of intelligent automation platforms also generate prescriptive insights: recommendations toward optimal operational performance.

5. Intelligence powered by cognitive/AI technologies and innovative tools

IDC points out that “progressing to more advanced automation technologies will stand a higher chance of delivering better business value and success.” Though there are many success stories, there are also numerous automation programs that have stalled — and those have first-generation RPA products to blame. Products with native AI make it possible to automate complex, non-standardized, and less repetitive tasks. IDC says: “Most vendors offer these capabilities through third-party partnerships; only a few have proprietary solutions” — and we are proud to say that WorkFusion pioneered native cognitive/AI automation, plus it’s the only product that delivers end-to-end process automation without third-party vendor integrations.

6. Strong support extended by the ecosystem

Despite the increased simplicity of modern products, most large enterprises work with partners to deploy and scale intelligent automation. Products with a unified platform model and native AI eliminate the brute force integration work that partners are often paid to do and allow expert resources to focus on the most valuable work: identifying processes to automate and deploying the product across the business.

In every growing technology category, buyers have many choices. Enterprise leaders with a long-term, strategic view of their business want to reduce cost, drive revenue, and improve the experience of both their customers and employees. Intelligent automation is key — and knowing what to look for when it comes to taking their automation efforts to the next level helps leaders make the best decision for their organizations.

Note: This article responds to only part of IDC’s “Robotic Process Automation Game Changers Advance Financial Services Institutions Toward Intelligent Digital Workforce.” In another article, we discuss the report’s definitions for an “intelligent digital workforce” and “cognitive/AI” solutions. WorkFusion has licensed the report, so if you’re interested in reading more, you can freely download a copy here.

IDC Research: 6 Defining Characteristics of an Intelligent Digital Workforce was originally published in WorkFusion on Medium, where people are continuing the conversation by highlighting and responding to this story.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

Recently, WorkFusion and our partner PwC hosted a webinar on automating AML featuring Michael Lammie, a leader in PwC’s Financial Crimes Unit. During the webinar, available on-demand, we received many smart questions and we are happy to answer some of the most interesting ones here.

1. How easy is it to implement a use case like automation of Negative News Adjudication?

An AML process like Negative News Adjudication can be implemented into production within 4 weeks.

WorkFusion’s AML Expert Bots help automate complex processes quickly. This approach includes prebuilt workflows, quick connections to common systems and pre-trained but customizable bots. Implementation can be compared to hiring a colleague from another bank, one who is already trained on AML and comes with useful experience that speeds up how quickly they are productive.

A more bespoke AML process will typically take 4 weeks to pilot and 4–8 more to put in production (8–12 weeks total).

2. How are differences from bank to bank addressed in AML processes?

WorkFusion’s SPA is a flexible platform, so automations can be tailored to the requirements of a specific bank or financial institution.

Learning bots are trained by your staff, learning from real-world examples. They can be trained in real time and on premise (or to the cloud), which creates your own intellectual property. This contrasts with many AI tools, which are trained once, remotely and only in the cloud.

Customers can also easily add their own connections. Using rules (APIs or RPA), one can quickly integrate with proprietary systems and apps.

This customization can be easily implemented by the customer, with WorkFusion training available through our Automation Academy. Trusted partners including PwC also have a ready-trained bench of WorkFusion experts to help guide and accelerate if needed, and WorkFusion can also provide support directly.

3. What other news sources can be used? What can we change?

WorkFusion can use any news source. Typical sources are pre-built, like Google News and Factiva. A customer can also add niche providers or their own internal data sources.

Moreover, WorkFusion works in 135 languages, and with non-Latin characters and alphabets, so global sources are encouraged.

Anything about the Negative News Search process can be changed. While the pre-built components help accelerate deployment and are based on best practices at top 10 global banks, every step is open to customization to meet bespoke requirements.

4. How do I address all other AML use cases?

AML is a core focus area of WorkFusion, with a dedicated product team, developing specific functionality.

The core WorkFusion product, Smart Process Automation, is a flexible and process-agnostic platform. It’s used across the world and in many industries to automate manual work (with customers ranging from Spotify to Canadian Tire!).

This means that it isn’t limited to just those AML processes with Expert Bots, but can be applied across AML, KYC, Financial Crimes, Compliance and Risk processes.

We have customers who automated over 100 processes in AML and who continue to scale their AML transformation programs at pace.

5. How do Learning Bots work? How do they analyze news?

Think of an employee you hire to perform investigations. There are skills you expect them to already possess on Day One, and there are skills you expect them to learn. For example, you expect them to be able to read the news and check for key words. They may need additional training on how to judge different risk factors and summarize findings. Over time, as they work on more cases and receive guidance from their manager, they will become better at the job.

Similarly, WorkFusion comes with NLP capabilities — the bots can read the news and have a keyword dictionary. AI is then used to help the bot interpret the news, judge the different risk factors, make better decisions over time, and produce ever-better summaries of its findings.

The specific AI technology is supervised machine learning. The bot learns from people. Its effectiveness and speed are further supported by a cutting-edge AI technique called “Weak Supervision” which reduces how much training is required, so the bot can learn faster.

6. How do we justify use of AI to regulators?

Fortunately, these days, regulators are encouraging the use of “innovative approaches” in AML. U.S. federal banking agencies and the Financial Crimes Enforcement Network (FinCEN) issued a joint statement in December 2018, specifically welcoming Artificial Intelligence to “strengthen compliance approaches… enhance transaction monitoring system… [and] … maximize utilization of banks’ BSA/AML compliance resources.”

WorkFusion offers a range of capabilities to explain what AI is doing and how it is performing:

  • Learning bots are supported with AutoQC, which allows people to dynamically review bots’ work, or, where appropriate, make the process fully attended or unattended
  • For sensitive processes, this human-in-the-loop allows for a “maker checker” model, where the final decision is always made by a person
  • Learning bots are trained based on business logic, and are easy to configure to use “white box” models
  • Preventive controls, audit trails and AI-era security help keep the process compliant
  • Analytics through the end-to-end process make it easy to monitor the performance of both bots and staff, down to a decision/field level
Watch our Anti-Money Laundering webinar here. We’ve also summarized 8 of the best points in a recent blog post.

To learn more about WorkFusion and our AML solutions, email us at learn@workfusion.com.

You Asked, We Answered: 6 Top Questions From Our AML Webinar with PwC was originally published in WorkFusion on Medium, where people are continuing the conversation by highlighting and responding to this story.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

Three years ago, the prospect of business process automation was often met with the question, “Will it work?” Now that nearly every enterprise has deployed some form of robotic process automation (RPA), its feasibility and value have been proven. The critical question has become, “Will it scale?”

According to IDC analyst Sneha Kapoor’s latest research report, Robotic Process Automation Game Changers Advance Financial Services Institutions Toward Intelligent Digital Workforce, many financial institutions that embraced RPA have not been able to scale their automation deployments. Along with highlighting the problem of scalability, the report clearly differentiates a “digital workforce” from an “intelligent digital workforce.”

Building on research of financial institutions across APAC, this excellent report will help any enterprise across the globe better understand the intelligent automation market and learn what to look for in a technology partner to help drive scalable operations transformation. We’re also proud WorkFusion was highlighted as a strong example. Download a complimentary licensed copy of the full report, or read on for some highlights.

What makes automation intelligent

The term “intelligent automation” is starting to pop up more frequently, even from software companies that have historically billed themselves as just RPA providers. But is their automation actually intelligent? IDC offers this definition: “software robots that can perform both deterministic and non-deterministic tasks by continuously understanding and analyzing structured and unstructured data.” Note these three critical terms:

  • Non-deterministic — The task isn’t precisely repetitive and requires judgment, and a software robot must be able to vary outputs according to variable inputs.
  • Continuously understanding — The software robot must ingest new, variable data and continuously adapt while maintaining a consistently accurate output.
  • Unstructured data — Refers to 90% of the data a typical business handles, and most is from documents (particularly within banks).

IDC says these intelligent software robots, “like their human counterparts, are both self-learning and self-healing workers that can discover patterns to predict decisions and even offer recommendations to improve them.” Whereas first-generation RPA products used rules to follow repetitive patterns, such as opening an application and logging in, intelligent automation has progressed to “augmenting human intelligence, as well as evolving quickly to achieve the potential of autonomously emulating this intelligence.”

And the key capability that makes automation intelligent? Artificial intelligence.

Defining “cognitive/artificial intelligence”

“Cognitive” and “AI” are favorite buzzwords for enterprise software marketing departments. The two terms are used so often, by such a wide array of products, that it’s difficult to know what each vendor means and whether or not the products are actually built with intelligence. IDC helps end the ambiguity with a short, clear definition: “IDC defines cognitive/artificial intelligence as systems that learn, reason, and self-correct.” If an RPA bot does not have this capability natively, it isn’t intelligent. Within native AI, rules-based or deterministic bots require frequent retraining — which precludes users of first-generation RPA products from scaling their automation programs to perform complex functions autonomously, enterprise-wide.

Spotlighting WorkFusion platform’s capability, simplicity, scalability and analytics

After spending time with WorkFusion’s intelligent automation platform, the report authors remark:

“Even its AI model can scale up with no need for additional programming and can run on standard hardware to make automation portable across functions, business units and geographies. Much of this simplicity is due to WorkFusion’sAutoML, a capability that analyzes process data using input and output data from completed tasks to pick algorithms, train models, and insert models into the automation in a seamless, nondisruptive fashion.”

Recognizing the importance of measurement and reporting to all large enterprises, IDC analysts noted that WorkFusion’s “predictive analytics provide full visibility and centralized view of automation across bots and processes so that users can better plan cost and capacity.”

We appreciate how this research by IDC supports the work we’re doing to lead our industry, and offers a strong testament to how our Intelligent Automation platform helps make our customers game-changers in their industries.

Note: This article responds to only part of IDC’s “Robotic Process Automation Game Changers Advance Financial Services Institutions Toward Intelligent Digital Workforce.” In another article, we’ll explore the report’s 6 characteristics of an intelligent digital workforce. WorkFusion has licensed the report, so if you’re interested in reading more, you can freely download a copy here.

IDC Research: Toward an Understanding of “Intelligence” in Process Automation Software was originally published in WorkFusion on Medium, where people are continuing the conversation by highlighting and responding to this story.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

In a recent joint webinar, “AML: Improving Compliance and Reducing Cost with AI,” experts in the field discussed newest innovations in the fight against financial crime. WorkFusion’s pre-built AML Expert Bots and Smart Process Automation (SPA) platform is used by leading banks and financial institutions to automate AML processes.

Speakers were Michael Lammie, a director from PwC’s Financial Crimes Unit; James Lawson, Director, Strategic Markets and a leader in WorkFusion’s AML efforts; and Mikhail Abramchik, WorkFusion SVP for Product, who leads the team who built the SPA AML package.

During the discussion, available on-demand here, several key points were made, which we are happy to sum up for you here. James also offered a detailed demonstration of one automated AML process, Negative News Screening. Additional questions asked during and after the webinar will be addressed in a follow-up blog, so check back soon!

8 important takeaways:

1) Crime uses cutting-edge tech, so shouldn’t you?

A root cause of AML compliance violations is that bad actors are overwhelming inadequate systems and technology. Traditional AML processes are manual, labor-intensive and hard-coded, which makes them rigid and difficult to adjust in response to a dynamic problem. By comparison, WorkFusion learning bots allow you to automate this work and continuously improve.

2) Manual work is a root cause for slowdowns, gaps and high costs

The tasks required for effective manual AML are time-consuming, repetitive and unfulfilling. Your employees can only do so well, so quickly, yet customers have high expectations for speed and teams are under pressure to conform to SLAs. Even at the highest-performing institutions, prevention is deficient, resulting in periodic breaches and thus incurring fines and other punishments.

3) The answer cannot be to hire more people

Due to the manual nature of previous solutions, a natural response to increased complexity was to hire additional staff, often offshore personnel. But this is outdated. Like PwC says: “As technology continues to rapidly evolve in the space, decades of proven labor arbitrage and shared services are no longer the most constructive approach.”

4) WorkFusion platform is agnostic & non-intrusive

Financial institutions and banks already have AML operations, delivered by various tools and processes. As these solutions develop over time, the IT and connections linking them are often rather complex. WorkFusion works with any existing processes, tools and people already in place — but with a focus on making them all more efficient, reducing manual work, and improving compliance.

5) Expert Bots accelerate automation, with flexibility

WorkFusion’s pre-built AML Expert Bots are set up to quickly start automating complex processes — yet each is still customizable. It’s like hiring a team member who’s already experienced with AML, but just needs a bit of extra training toward your specific needs to be fully productive.

6) AML includes many opportunities for automation

In addition to Negative News Screening (part of institutions’ KYC obligations, which was discussed and demonstrated in the webinar), PwC recommended many fertile opportunities to automate processes, including: Customer ID & V (identify and verify), Customer Risk Ranking (factor analysis), Beneficial Owner Analysis (identifying hierarchies and business relationships), and Customer Documentation (document collection and data extraction). Payment Sanction Screening and Source of Wealth analysis are also popular automation targets.

7) Explainable AI is more regulator-friendly

Financial crime units within institutions are risk-averse, in part because their decisions and actions must be auditable and transparent. Fortunately, WorkFusion offers various capabilities like Analytics, AutoQC and preventative controls to aid with process management and explainability. Regulators are now more welcoming to such innovations too. But don’t just take our word for it.

8) Benefits speak for themselves — instantly

Automation reaps short- and long-term benefits. Obviously, there are immediate hard results, such as 70% less manual effort… yet 3x coverage. For one global commercial bank using WorkFusion for Negative News analysis, the savings are multiple millions of dollars each year. In other banks that are smaller, more fragmented or less mature, with only a few FTEs focused on Negative News searches, they will still see ROI within a year and can expect steady scaling up in line with their business.

In addition, customers are experiencing softer benefits, such as better decision-making, higher employee satisfaction, flexibility to scale up and increased speed — all of which leads to greater compliance.

To learn more about WorkFusion and our AML solutions, email us at learn@workfusion.com.

Top 8 Lessons from PwC and WorkFusion’s Anti–Money Laundering Webinar was originally published in WorkFusion on Medium, where people are continuing the conversation by highlighting and responding to this story.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 
Now that 2019 is upon us, we thought it’d be fun to briefly look back on an amazing 2018 for our company, our team and our customers. It’s been a year of breakthrough innovation, incredible growth, and community-building worldwide. Here are the highlights:Breakthrough innovation

In 2018, we released LUMEN, a big upgrade to our industry-leading Intelligent Automation platform, bringing critical enhancements such as built-in predictive analytics to automation, as well as major leaps in performance, security and user experience. We also launched RPA Express Pro, a subscription upgrade option to our free RPA Express product.

Incredible growth

Our team doubled in size (and plenty more in smarts and strength), with operations now spanning the globe — from New York, London and Paris to Munich, Minsk, Singapore, Tokyo and Bangalore. Our customers now span six continents (Antarctica, here we come!) and most industries. Our largest customer has over 200,000 employees and the smallest under 50. We also crossed the 40,000 mark for RPA Express users, and 20,000 students in our online Automation Academy. To support this amazing growth, we raised $50 million in funding.

WorkFusioners celebrate summer and great teamwork during BreakAway, our annual all-company retreat in Minsk, Belarus. Join our team!Awards, events and recognition

In 2018, we continued to be recognized by the industry, including the prestigious “The Forrester Wave™ and Capgemini’s “World Fintech Report 2018”. The “Artificial Intelligence for Business” special report was published in The Times in the UK. WorkFusion ranked as the №1 AI-driven software vendor by Oliver Wyman’s Celent Research arm in its report, “The Cognitive Corporate Bank.” Plus, our COO Monica Jonas was honored in The SaaS Report’s 2018 Top Women Leaders in SaaS.

We also reached new heights with Ascend, our flagship annual conference, welcoming over 1,200 attendees in New York, plus hundreds more at Ascend events in London and Bangalore.

Community building

One of our passions is to support tech education and community. WorkFusioners host meetups for the tech community, with topics including Agile processes, product user discussions, RPA innovations and more. Students are welcome to visit our offices, and we are proud to collaborate with higher education programs.

Students are welcome to visit our offices, and we are proud to collaborate with higher education programs.

THANK YOU!

After a celebratory holiday season to close out the year, we are refreshed and ready to reach even higher in 2019 — and more than anything, look forward to working with you.

Excelsior! Onwards and upwards!

Snapshots from WorkFusion holiday parties around the world: New York, Minsk and London.

WorkFusion 2018 Year in Review was originally published in WorkFusion on Medium, where people are continuing the conversation by highlighting and responding to this story.

  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 
Kyle Hoback, Director, Market Enablement, WorkFusion

The countdown is on for Santa Claus’s annual sleigh ride, when he brings Christmas cheer to (nice) girls and boys. What you may not know is there’s a new factor shaping the North Pole’s 2018 toy delivery: Artificial Intelligence (AI).

Like all modern executives running data-heavy enterprises, Santa has been intrigued by AI and has been considering how it will transform his business (making and delivering toys). But unlike 95% of his peers, the Big Guy in Red has extensively incorporated AI into his day-to-day operation.

How has Santa been able to not only develop, but also act on an AI strategy? It starts with a business-oriented approach, an understanding of data, and a “get to production” methodology.

Maintaining Focus on Business Objectives and Challenges

With centuries of experience as an early adopter of new technology, Jolly Old St. Nick knows to maintain focus on his business objectives. He knows he cannot treat technology as just a novelty, but as a tool to improve the key mission: maximizing the happiness of nice boys and girls around the globe by delivering toys on Christmas.

Santa’s main objective is child happiness, but operational challenges make achieving it tricky.

While he may have some magic on his side, when Kris Kringle considers his operation at the North Pole, he notes key challenges:

  • Increasing kid-to-elf and toys-to-elf ratios, plus elf work hours: The global population of kids is nearing 2 billion, increasing faster than his elves can keep up with, leading to longer work hours and a disgruntled production team.
  • Increased variety of toys: The growing variety of toys increases the difficulty of mapping options correctly to each child.
  • Increased range of materials, and assembly time per toy: Each toy requires different materials, and construction times can range dramatically. Lincoln Logs (simple wooden tools on the market for 100+ years) are much easier to produce than the Anki Cozmo robot or the (ever-dreamier) Barbie DreamHouse.

Santa is using AI to address these challenges, ensuring a cheerful Christmas for the kiddos — and happier elves in the North Pole.

Understanding Available Data: Format, Location, Processes

Father Christmas has done his research and understands that AI solves defined tasks by learning from data. Which means he needs to understand his data and the tasks performed with it.

At a high level, North Pole data sounds simple enough: a list of each child in the world, detailing where they live, whether they were naughty or nice, and which toys they should receive. Santa works with his elves to make that list and check it twice — but the devil is in the details, as they say.

One of Santa’s key data sets comes from in-person encounters with children.

To better understand his data, Santa commissioned an analysis, summarized below:

Based on the study, it’s very clear that Santa has data, lots of data. But it’s also clear the data currently requires work and governance to make it useful for each current year and referenceable in years to come.

Executive Learnings While the Machines Are Learning

Santa Claus has been reading up on AI case studies, and a common theme is that productivity gains and true ROI come from getting AI into production. His goal, then, has been to find practical applications that drive impact he can improve upon, rather than just expecting the machine to learn everything. In essence, he strives to avoid long-term moonshots in favor of continually improving useful AI applications.

Father Christmas does have eyes on his moonshot: An attempt to throw all the Children, Toy, Letter, and Notes data into a master model that could spit out an annual delivery list with no elf effort (and thereby be able to find different work for the elves). But, thus far, it hasn’t proven practical to do so, especially because the best data doesn’t arrive until just before the big night and the processing power required might actually melt the North Pole.

His primary focus, then, has been on Intelligent Automation use cases (as we initially discussed last year) — which put AI to work on the manual aspects of his operation, such as the data collection and processing work of the elves, including:

  • Letter Processing: Identifying children and toys from scanned paper mail, emails and social media posts (fax is no longer used by the children).
  • Shelf Notes Processing: Classifying the nightly reports of those elves that sit on shelves into naughty/nice (with severity levels), along with the occasional identification of a toy request.
  • Toy Popularity Planning: Proactively analyzing hot toys of the season, specific toy demand, and assumed childhood excitement levels from open source news and social media.

Value has been achieved in each of these three examples, all retaining some level of elf involvement to assist the machine with validation and updates to the predicted data points to complete the work and re-train the machines. This has decreased elf work hours while improving the toys-to-elf ratio.

Elves can take breaks from all the North Pole’s data work.

AI now underpins all the data work in the North Pole, a continual interaction between elves and machines. But it’s not just the machines that are learning: every new model that is introduced or enhanced is also a teaching moment for Santa to improve his understanding on how AI can transform his operation.

AI at the North Pole

Santa Claus is well along his journey of implementing AI within the North Pole. Keeping his business objectives in mind, St. Nick has a strong grasp of his data — what it is, where it is, and how it’s used — and is on a continuous learning path so that he and his elves can make the most of AI.

(Photos captured from Unsplash.com)

How Santa Claus Is Using AI Before He Comes to Town was originally published in WorkFusion on Medium, where people are continuing the conversation by highlighting and responding to this story.

Read for later

Articles marked as Favorite are saved for later viewing.
close
  • Show original
  • .
  • Share
  • .
  • Favorite
  • .
  • Email
  • .
  • Add Tags 

Separate tags by commas
To access this feature, please upgrade your account.
Start your free month
Free Preview