Don’t Create a Monster, Create a Beast

Halloween is right around the corner, and nothing is scarier than transforming your enterprise into a Frankenstein of incompatible tools and techniques. 

You started with the best intentions. You thought you chose the right tools, most attractive processes, and the best minds. 

But you soon found the amalgamation birthed a hideous creature of stitched-together processes, disconnected guidance, and opposing insights. It’s turned your transformation into a nightmare of rotting results, festering budgets, and decaying productivity. Instead of electrifying your enterprise with new life, you’ve created a monster.

Download the full infographic

Consultant’s Brain

Consultants are the experts, and you would expect them to be the brains behind your transformation. But they’re working with limited resources and time, trying to capture process details and then extrapolating anecdotes across your organization. They’re bound to miss things, which leaves your project wandering between options and never really making transformational impacts.

Miner’s Arm

Process mining and discovery techniques are slow, expensive, and highly manual. It’s an outdated approach that only captures a slice of a process based on how users interact with one system. It limits your coverage, misses steps performed in separate applications, and could pull your transformation in the wrong direction.

Software’s Arm

Digging for process details via APIs requires even more connections with IT, developers, and application vendors. It adds time and expense to your transformation project, and even more opportunities for messages to get lost and connections to break. Plus, each application requires a separate API connection, delivers data in different formats, and forces you to manually reconnect and reanimate the insights, adding even more delays.

Data Guts

Most companies truly don’t understand how they operate on a granular user level, and this lack of current state understanding is a major roadblock to transformation success. Process Intelligence helps eliminate nightmares by efficiently creating a dataset of user activity not previously available to kickstart strategic initiatives. Convert your process problems into big data solutions.

Bot’s Leg

RPA is great, but every organization struggles with scaling programs. Process assessment and prioritization delays development. Vendors have promoted screen recorders for task discovery, but they fail to scale and actually create more rework than traditional methods. Scary indeed. The only real answer to scaling your RPA program is adopting real process intelligence. Deploying FortressIQ means rapid assessments, and finding automation opportunities at scale

App’s Leg

Know what’s really scary?Most enterprise processes–especially the valuable ones–require workers to hop between software, move data across screens, or use multiple apps simultaneously. Trying to combine those steps into a coherent process using just log files and APIs is a recipe for disaster. Process Intelligence can give you the insights to start correcting the issue so you can understand today to improve tomorrow.

Stop the Madness! Turn Your Enterprise into a Beast

FortressIQ goes way beyond all of these scary, cobbled-together options to capture every step of every process, without expensive and disconnected tools pulling you in different directions. It brings your transformation to life with real-time, end-to-end process insights captured with DNA-level analytics across all applications, departments, and processes.

Don’t let your transformation turn into a monster. Use FortressIQ to make data-driven decisions that successfully propel your enterprise into tomorrow.

Transforming BSA/AML and KYC with Process Intelligence Technologies

The U.S. Bank Secrecy Act (BSA) of 1970 was one of the first Anti-Money Laundering (AML) and Know Your Customer (KYC) laws. It required companies and financial institutions to establish and report on internal controls and other measures put in place to prevent the facilitation of financial crimes. Other similar laws exist in countries around the world, creating a complex web of potential compliance issues for financial services companies.

Between 2008 and 2018, financial institutions worldwide have paid an estimated US$26 billion in fines and penalties as a result of violations to these regulations. That’s an average of $2.6 billion per year. However, government scrutiny of money laundering is now at an all-time high. Financial Institutions were fined US$5.6 billion in the first half of 2020 alone for non-compliance with AML, KYC, and related regulations. If the trend continues, it would represent a 430% increase over the previous ten-year average.

It is increasingly clear that compliance with these regulations is critical to the sustainability of every financial institution. Unfortunately, the traditional means of transforming your BSA/AML processes are woefully inadequate. But there are new technologies helping accelerate and increase the success of BSA/AML transformation.

Does Your AML/KYC Process Add Risk?

While it is the responsibility of all employees, partners, and suppliers to prevent an organization from facilitating financial crimes, Client Lifecycle Management (CLM) and Compliance are the two departments playing key roles in defining and implementing the required internal controls. CLM is the first line of defense within any organization. Compliance acts as the second line of defense, responsible for policy making, escalation, and resolution, as well as performing independent risk management. Auditors, the third line of defense, ensure any risk governance framework complies with regulatory guidance.

Three Lines of Defense Model

Before taking on a new client, a due diligence process is generally conducted to evaluate the client’s risk rating. It begins with a basic understanding of the client’s identity, the risk involved, and an understanding of their financial habits. Onboarding high-risk customers and politically-exposed persons requires enhanced due diligence with additional assessments of the client’s geographic location, source of funds, and purpose of the transaction, and may require ongoing monitoring.

This is an important task that typically happens as follows:

  1. Pre-onboarding checks are conducted by working with Sales, Risk Management, Legal, Compliance, and others to collect and review relevant client data, product information, and documents as mandated by the regulatory authorities.
  2. Teams then update multiple systems of record to ensure a client’s readiness to transact.
  3. Post-onboarding processes then include on-going client reviews and continuous monitoring, managing client and counterparty data and records, and potentially, client off-boarding.

This process can quickly become complex, especially at global organizations spanning multiple geographies with various policy interpretations, competing rules and regulations, and related data housed in multiple and disconnected software applications. That last point adds risk, especially when data is not integrated, thereby forcing considerable amounts of manual, repetitive, error-prone work. The result is increased operational, reputational, and financial risk.

Additional risks arise from policy interpretations and potentially incorrect execution of processes, which both depend on the experience of KYC analysts. It is indeed demanding for analysts to make critical decisions that require focused thinking while concurrently performing important yet mundane manual data-entry tasks.

Add it all up and your AML/KYC process is exposing you to more risk, which is exactly the opposite of what it is supposed to do!

Transforming BSA/AML with Success

Transforming any enterprise process can be daunting, for good reason. A study by McKinsey & Company indicates that a staggering 70% of large transformation projects fail to deliver expected results. Reasons may include unclear objectives, lack of leadership, and lack of commitment. But looking deeper, transformation projects are frequently derailed when teams underestimate process complexity. It’s a huge undertaking to identify the appropriate processes, perform detailed current state assessments, develop business requirements, and keep an eye on budgets. Then, for any transformed process, adequate training is required, and even minimal employee turnover can add to the challenges.

When focused on AML/KYC processes, the need for a successful transformation can be critical to your organization’s survival.

But help is available from point solutions such as Microsoft Power Automate, which uses Robotics and artificial intelligence (AI) to help organizations streamline, standardize, and automate routine tasks. Many financial institutions are also leveraging cognitive natural language processing (NLP), with focused solutions such as DDIQ by Exiger, to accelerate adverse media and sanctions screening processes related to clients.

AML/KYC platform providers can help streamline end-to-end processes. But successful implementation of these types of platforms largely depends on the quality of the business requirements and clearly defined compliance policies. It’s also dependent on the prevailing regulatory rules, final user acceptance testing and training. In reality, it takes many months for organizations to fully understand and effectively leverage these platforms, which adds further delays to already complex transformation projects.

FortressIQ is playing a key role in a successful AML/KYC transformation by converting a process problem into a big data problem. FortressIQ performs detailed current state assessments to provide near real-time process intelligence. It then provides the insights to make data-driven decisions.

Using computer vision, NLP, OCR, and deep learning algorithms, FortressIQ will:

  • Capture tasks at the most granular level, with no bias or blind spots;
  • Provide faster time to value by generating detailed, enterprise-wide process insights in just 2-4 weeks and without consuming worker time; and
  • Cost much less than human consultants, including eliminating documentation errors and the related rework.

Insights provided by FortressIQ can be leveraged by functional and transformation teams to collaborate on areas that matter: process enhancement, automation, and training.

Effectively managing your AML/KYC risk is critical to the success and reputation of your organization. Process intelligence and emerging technologies can help mitigate these risks, speed up the transformation journey, and enhance the customer and employee experience. It could also prevent a AML/KYC violation, which is becoming an increasingly expensive prospect.

Top-Down vs. Bottom-Up: Where to Begin Your Process Transformation Journey

Developing a deep and detailed understanding of your business processes lets you root out inefficiencies, double down on operational excellence, and make better, more informed decisions to reach your goals. But simply mining system log files misses the details in every process, while deploying consultants to map processes takes time and disrupts operations. Instead, intelligently decoding how your people and processes really work, across systems and screens, and across your entire business, is a better approach.

Capturing process intelligence to understand business processes has traditionally been tough to obtain because the methods have been manual. The drawback is that it results in static, incomplete process data. Today’s technologies, which evolved from these traditional methods, offer an automated and intelligent approach that’s both faster and captures more detail.

Here’s how process insight capture has evolved:

  • Process Mapping is the traditional, human-based route where business analysts and consultants interview and look over the shoulders of your workers. It’s slow and expensive, the sample size is limited and incomplete, and it can’t realistically cover processes across the entire organization.
  • Process Mining is a back-end, system-centric approach that captures a narrow, step-by-step workflow based on how users interact with specific systems. This method requires access to log files, which limits coverage. It also misses tasks like data collection, calculations, or other steps performed in separate applications.
  • Process Discovery is a modern alternative to mining. It tracks workflow at the UI level, no matter who performs the task or which application is used. It excels at capturing discrete sub-processes, but has trouble scaling because it ultimately requires human evaluation of the results.
  • Process Intelligence advances process discoveryby using computer vision and Artificial Intelligence (AI) to uncover actionable process insights at enterprise scale. It has the speed and coverage to capture, record, and analyze granular steps in complex use cases, plus adds intelligence to quickly identify new opportunities.

This spectrum of process insight techniques is referred to as a top-down manual approach, versus a bottom-up intelligence-driven approach. They all help you gain a better understanding of processes, but the bottom-up approach offers more granular insights, faster and across a wider range of processes. The result is more business impact in less time.

Choosing an Approach to Gathering Process Insights

If you have process insight experience, you may lean towards combining multiple approaches to address specific requirements. But if you’re tackling a project for the first time, it can be difficult to determine the best approach.

A top-down manual approach adds the perceived expertise and guidance of a team of consultants. That can be helpful but adds more cost and time by a few orders of magnitude. On the other hand, a modern bottom-up approach offers deeper insights and faster results but puts the decisions in your hands.

So, the question is, do you take a bottom-up or top-down approach?

Top-Down Misses the Detail

Top-down process mining technologies piece together a process within a single system, but since they do not capture granular user level activity, they don’t capture all the key steps of users or systems. For example, process exceptions and variations are not reliably identified because they may involve activities outside the analyzed system.

Additionally, enterprises often run hundreds of applications. Many of those—including Excel and common email clients—don’t generate usable log files. So, any use of those tools will not be captured, and the resulting process maps won’t represent the complete business process. The missed steps then aren’t included in automation efforts, resulting in costly rework once deployed.

Bottom-Up Provides More Insights and Speed

In contrast, process intelligence and process discovery technologies capture detailed user activity across all systems and tools, covering every granular step, including task interdependencies and connections. There is no need to access APIs or log files to create the process maps, which speeds the entire project, and the more complete insights keep you from automating broken or inefficient processes. Analytics can also quickly compare processes and system usage across teams and tasks.

Speed is what really separates the top-down and bottom-up approaches. You can expect to receive business value in weeks with Process Intelligence instead of months with Process Mining.

Bottom-Up
Process Intelligence
Top-Down
Process Mining
ACCURACY
Completeness
Full capture of sub-process activity and variations Limited view of end-to-end functional workflows
SCALE
People & Process Coverage
Full coverage across all systems and teams Limited coverage to systems with log files
DETAIL
Degree of Specificity
Level 5 step-by-step process and sub-process details Level 3 general workflows
SPEED
Time to Value
Weeks to deploy to desktop sensors and collect data Months to integrate back-end systems and map data
EFFICIENCY
(reduced rework)
Granular activity data is comprehensive and actionable System-only data leaves process gaps

 

Start at the Process Itself

If you are exploring process insights for the first time, Process Intelligence is the logical starting point given its more comprehensive view of operations. It offers a faster time to value, takes less IT resources since you are not integrating with APIs or capturing log files, and is substantially less expensive and disruptive than unleashing a team of consultants into your operations.

As you eventually dig deeper for process insights, Process Mining could complement Process Intelligence, however. Combining both approaches provides the most comprehensive insight on the implications of a process. Interested in how FortressIQ provides process intelligence quickly and efficiently? Read our solution brief to see how you can get started.

Large-Scale Business Process Transformation Starts from the Top-Down

“If You Can’t Measure It, You Can’t Improve It.”

– Peter Drucker

Measurement and improvement – easier said than done, especially in the enterprise. When it comes to large-scale, strategic transformation, it’s also a continuous journey. Every enterprise company is in some stage of digital transformation, and there are challenges at every stage.

For the past decade, companies have been trying to make decisions with data collected from all lines of business in an effort to move the needle on a successful digital transformation initiative. Efforts to integrate departments such as finance, HR, supply chain, procurement, and marketing with various technology solutions have seen varied levels of success with ample opportunity for improvement. Individually it may be possible to collect data from different areas of the business, merge into a single place, and see what insights can be extracted, but without the right solution(s) in place, near impossible.

Several companies are in the process of standing up teams to tie data together from all lines of business into one single repository to be used for analysis and decision making. When you consider that an undertaking of this magnitude requires data from several different systems to be filtered into and stored in one place, and the IT systems and processes are constantly changing in parallel, it may be time and cost-prohibitive. In addition, the surge in popularity of automation and RPA have companies implementing bots without truly understanding the overall business impact.

Instead of shuffling data from one system to another, and trying to compile insights from disparate systems and applications, you can capture the work and tasks being executed across all applications and systems – for multiple users and with zero business interruption – using a cognitive process discovery solution. AI can be used to study, map, and deliver process information to multiple department heads so they can make data-driven decisions that improve efficiency and productivity. Additionally, automation and process improvement initiatives can move from changes to pockets of the business and expand to full departments, shared services, and centers of excellence to have the greatest effect on the company’s largest, strategic transformation initiatives. Common examples of process optimization in the enterprise often start by using AI to identify improvements to back-office systems and business processes such as finance, HR, and customer support – small changes in these areas can see huge increases in ROI and deliver quantitative results to the enterprise.

To learn more about our automated process discovery and documentation solution, you can request a demo here.

AI in the Enterprise: A Brief Intro on the Technology Driving Digital Transformation

Whether you’re thumbs up or down on artificial intelligence, it’s here to stay, and it’s here to change how we do business. At FortressIQ we are big advocates of using AI for what it’s good at, and alternatively, having humans focus on what they’re good at. Implementing AI effectively gives workers the time to spend on those job functions where AI cannot add value and can increase employee productivity and satisfaction.

Using AI technology to enable better, more effective business outcomes all sounds great but where do you start? This AI mega trend means that business executives (and other non-technical roles) are expected to evaluate and make decisions on where to implement AI in the workplace. For many employees it’s a task just to decipher the jargon, what it all means, and how it can be used to address digital transformation initiatives.

AI technology addresses 2 key areas in the enterprise:

  • How to make sense of the mountains of data collected
  • How to make better decisions based on that data collection

Your current systems – as well as your people – have a lot of knowledge on current processes, customers, suppliers, etc. As businesses expand, the data explosion continues. To enable better decision making through data-driven insights, a few different AI technologies can be deployed, each with a different attribute to address these challenges.

  • Computer Vision
  • Machine Learning
  • Deep Learning
  • Natural Language Processing

Computer Vision

Computer vision provides the ability for a computer to gain a high-level understanding of digital images and videos so that machine can then recognize and make decisions based on the set of images produced. The technology has grown to include facial recognition and the identification of objects such as traffic signals, stop signs, and pedestrians.

Computer vision is used in the automotive industry to create anti-collision detection technology for better vehicle safety. It’s also very popular in healthcare to improve patient diagnoses through enhanced detection on MRI, X-ray and other scanned images. In finance departments, it can quickly identify and process invoices, improve cash flow, and build better relationships with vendors and suppliers.

While machine learning focuses more on making sense of a large amount of data, computer vision and deep learning technologies are focused on training a computer to be able to understand its environment and make decisions similar to a human brain.

Machine Learning

Machine learning is the ability to create meaning from mountains of data. In business, this is often referred to as data mining. Machine learning technology can rapidly make inferences from a large amount of data, whereas if a human performed the same task it could take them thousands of hours. This field of computer science gives the computer the ability to learn without being explicitly programmed.

Companies can use machine learning to accomplish anything from targeted marketing to revenue forecasting. For example, online advertising companies use aggregate user data collected from companies like Google, Twitter, and Facebook to serve up targeted ads to people identified as more likely to purchase. Credit card companies can use machine learning to quickly process thousands of applications and monitor user purchase and payment history to serve up offers such as a credit limit increase.

Deep Learning

Deep learning technology, a subset of machine learning, uses algorithms to learn in a supervised or unsupervised manner; the algorithm does not need to be task-specific. For example, it can be used to classify a large data set or identify and analyze patterns within that data. It can then use those patterns to predict possible outcomes. In business this is often referred to as predictive analytics. In short, deep learning replaces the traditional intuitive aspect of decision making with more data-driven decision making.

In a supply chain scenario, deep learning can be used to reduce the number of product modeling scenarios, and laser-focus on those models that will drive the most revenue. In finance departments a scanned invoice with an abnormally high dollar amount listed will be flagged as an error and automatically sent for review.

A system that can process data faster than a human, while simultaneously learning and applying that knowledge, can increase the overall productivity of an organization and reduce risk. And in the example above, when the task is finance related it could result in quicker revenue recognition.

Natural Language Processing

Natural language processing is the ability of a computer program to understand language as it is spoken. Natural language processing can be used when the text is provided. When text is produced, the computer will use algorithms designed to extract meaning associated with phrases and sentences and then collect essential data from them.

Although very intuitive to humans, aspects of natural language processing can be difficult to implement properly and haven’t been fully resolved. Sarcasm is a good example here – most humans can identify sarcasm immediately, but a computer or chat bot has a difficult time.

When big social media campaigns are launched, natural language processing can be used to track trends and customers’ pulse in real time, and campaign interactions can be addressed directly and be personalized, a critical element to successful brand marketing.

Until very recently, these more sophisticated embodiments of artificial intelligence have mainly been used for academic and scholarly research. Organizational efforts to stay competitive and remain a market leader (such as the race to build the best self-driving car) has forged a quantum leap in AI technologies, making them tangible and cost effective for the enterprise.

Even knowing the basic differences is a good starting point for researching where these technologies might be applicable for your organization. For additional information, check out our on-demand webinar “AI in Business: When and Where to use Artificial Intelligence in Your Organization.”