Accessible AI and Value Realization with Communications Mining
Learn how to streamline AI adoption and accelerate automation with existing data across your organization.
By Andrew Pirie, Anna Lampe, and Roger Smedley
Just years ago, the prospect of incorporating artificial intelligence (AI) was reserved for organizations with substantial investments in specialized data science teams. However, the emergence of generative AI and other advanced tools has ushered in a new era, breaking down barriers and enabling a broader range of businesses and functions to tackle their complex challenges.
And yet, many of our clients face a dilemma. With a proliferation of compelling AI offerings and investments — matched with leadership’s responsibility to be stewards of value creation — it’s difficult to understand what an iterative step looks like toward AI adoption and accessing “quick wins.”
At Slalom, we see a heightened demand for what we call “accessible AI” — a solution that quickly turns AI from a nebulous concept into a practical reality. Such an accessible solution would be quick to implement and genuinely effective for businesses or functions of all sizes and technical proficiency.
What is accessible AI?
“Accessible AI” refers to making artificial intelligence technologies more user-friendly and accessible to a wider audience, including those with no specialized knowledge in data science or coding. By breaking down barriers to entry, accessible AI enables a broader AI user base with expedited implementation, contributing to the creation of exponential value across various organizational dimensions. The concept of accessible AI encompasses several aspects:
1. Flexible, turnkey application
Pretrained models that allow businesses to quickly integrate AI into their operations, addressing immediate concerns, while also ensuring the adaptability necessary for sustained and evolving success in a dynamic business environment.
2. Little prerequisite training
Accessible AI also often comes with educational materials, tutorials, and support systems designed to help users understand how to best utilize AI tools, even if they start with little-to-no background knowledge in the field.
3. User-friendly
Accessible AI tools are designed with intuitive interfaces that allow users to interact with AI technologies without needing to understand the underlying complexities. This can include conversational AI, drag-and-drop interfaces, or simple input fields that guide the user through the process.
4. Versatility and integration
Accessible AI tools are developed to be versatile and easily integrated into existing workflows and systems, allowing users to seamlessly incorporate AI into their everyday tasks and operations without the need for extensive customization or technical support.
The goal of accessible AI is to enable rapid value creation by allowing more people to use AI effectively, improving efficiency, creativity, and decision-making processes across various industries and sectors. We’ve seen a convergence of these factors with a new offering at UiPath called Communications Mining and see the implications for scaled adoption across organizations looking for the lowest-hanging fruit in next best AI adoption.
What is communications mining?
Organizations are flooded with communications data, from emails to support tickets to chat and CRM notes. These digital interactions hold a treasure trove of customer insights, market trends, and operational nuances. However, the unstructured nature of these digital communications poses a challenge to organizations looking to derive value from this data. This is precisely where communications mining comes into play.
Communications mining offers tailored AI solutions for not only extracting insights from your communication channels but also automating the tasks surrounding these communications.
By leveraging advanced natural language processing (NLP) techniques to decipher human language, communications mining parses, interprets, and categorizes content from various communication channels, transforming it into machine-readable data that can be used for analytics and automation. Its user-friendly no-code interface, customizable features, rapid training capabilities, and scalability render it seamless to integrate into existing tech stacks. Unlike task and process mining, which concentrate on activity and process step analysis, communications mining specializes in comprehending and uncovering insights from human language.
The insights unearthed by communications mining models drive significant advancements in business automation, particularly in service and dialogue-driven workflows. With its Generative Extractor feature, generative AI is leveraged to produce well-organized information outputs. Each request is an “extraction” that includes not only the primary intention behind the request but also all pertinent associated data. By utilizing Generative Extractor, all different requests in complex messages can be correctly recognized and handled automatically.
What’s the value of communications mining for your organization?
The application of communications mining spans various organizational needs — but at Slalom we see several key direct and indirect benefits:
1. Creating a process taxonomy
Our favorite by-product of communications mining is in the conversations it forces the business to have around process. In order to categorize communication, we need to create a taxonomy around what the business does with those communications. This helps create structure for not only communications mining implementation, but also conversation opportunities regarding other ways to continue optimizing processes and systems.
2. Metric-driven understanding
The face value benefit of communications mining comes through the out-of-the-box analytics providing insights into customer preferences and pain points, facilitating more informed decision-making and strategy development. Moreover, the analytics and reporting capabilities allow organizations to measure and demonstrate the impact of their automation initiatives and support continuous improvement and strategic alignment.
3. Data environment viability
Unlike other discovery technologies related to organizational system data, like process mining, data requirements for communications mining are based on data in the form of emails and chat that all companies have readily available to them. And more importantly the data formats are highly standardized and easily ingestible with little to no transformation, which ensures rapid identification of viable use cases and reduces engineering overhead.
4. Automation trigger point
The downstream potential value from the technology comes through automation of processes originating or using these communications midstream. By converting unstructured communications into structured data, robotic process automation (RPA) developers can now automate complex communication-based processes, thereby unlocking new automation possibilities and extending the scope of RPA projects.
Communications Mining through UiPath
With the 2022 acquisition of Re:infer, UiPath dove headfirst into the AI-assisted analysis and automation of communications data. Integrating the tool into their automation suite enables a coupled functionality of communication tagging and information extraction to trigger and enhance new automations. This integration enables businesses to automate end-to-end transactional requests and workflows, such as email triaging, customer information updates, and case creation, all in real time and across major business communication channels. Built on their SaaS platform, the tool is able to be rapidly deployed and tested across different business process use cases. It has three components:
Discovery
UiPath wants to focus on turning business conversations into actionable insights. By extracting data from any message, discovery gives businesses insight into previous processes and unstructured data channels. It helps them better understand things like reasons for customer contact, customer sentiment, and where data fields exist within unstructured text. This enables them to identify problems, inefficiencies, and new opportunities for elimination and automation.
Monitoring
Gain a custom-filtered, real-time view of all service channels — extracting the information you care about most. Built-in analytics allow you to track trends, as well as measure service quality and performance at all levels. Custom, real-time alerts enable teams to respond to breaking issues in an agile way. Driving continuous improvement, being predictive instead of reactive.
Automation
UiPath robots use the data created by communications mining to extend automation into service and conversation-based processes. This enables you to automate transactional requests and workflows. Tasks like triaging emails, updating customer information, and case creation are now automated from end to end by UiPath.
So what does this look like in reality? In the section below we illustrate an implementation with one of our clients to help understand how to streamline triage and accelerate automations in the service desk.
A real-world example
In a recent collaboration, Slalom worked with the City and County of Denver (CCD) on a project piloting communications mining. CCD is a dynamic municipal entity that in recent years has leaned in to develop a highly mature automation center of excellence (CoE) within technology services that drives process optimization across their business functions. Their industry leadership has led to recent recognition from the Colorado Technology Association APEX Award as well as being listed in the most recent CIO 100 Organizations. This mature automation program also ensured that sustainment of the tool and potential automations stemming from discovery could be easily supported internally by the organization.
Like most organizations, their accounts payable (AP) team faced a significant workload managing inquiries on invoices and purchase orders, resulting in overwhelming demands on their resources. With the substantial influx of daily emails received by the AP help-desk mailbox, our client was looking to accelerate grouping and interpretation of communications and identify areas of automation opportunity. This use case also opened up the opportunity to rinse and repeat activities across multiple other help-desk teams with similar functions.
In a span of four weeks, together we embarked on a communications mining implementation to comprehensively understand the communications being received, identify pain points, and facilitate automation of classifications, routing, and downstream processes.
Outcomes
The AP insights derived from the communications mining implementation marked a significant achievement, setting the stage for enhanced operational efficiency, data-driven decision-making, and quicker customer responses — all accomplished in just four weeks.
Using the automation opportunity list we identified from the help-desk processes, estimated time savings, and extrapolating across help-desk functions, we were able to identify a sizable opportunity for efficiency and time savings initiated by communications mining. We identified estimated opportunities totaling in 41,018 hours of automation with a 604% five-year ROI and $1.4MM five-year net benefits across service desk functions, starting with accounts payable.
The communications mining implementation delivered crucial insights into the complex processes surrounding invoice and purchase order inquiries, and our findings also served as a roadmap for implementing targeted improvements, fostering a more agile and responsive government for the benefit of the entire community.
What comes next?
Accessibility to implement AI is a real thing. Through the marriage of simple UI and powerful value creation in insights and automation, communications mining presents one of the most accessible opportunities we’ve seen to apply AI using data and processes that look the same in just about every modern company. Utilization of this type of technology will be ubiquitous in the not-distant future as embedded “copilots” and AI engines are attached to just about every structured data source in the organization.
Organizations that jump first, with a strategic tie-in for the technology to business objectives and automation goals, will be able to clearly demonstrate the ROI — not just anecdotal benefits — for their organization.
Slalom is a next-generation professional services company creating value at the intersection of business, technology, and humanity. Learn more about our approach to AI or read our thought guide.
