BLOG

AI and ML in UC: What You Need to Know

Written by: Ellie Loehrer

Updated: 09/14/2021

Introduction

Wherever you look, artificial intelligence (AI) and machine learning (ML) is looking back at you. If it weren’t for artificially intelligent machines, machine learning in UC could not exist. In fact, a direct correlation between those who adopt AI into their UC systems and those who achieve UC success has been found in Nemertes’ 2018 Unified Communications and Collaboration report. To go even further, artificial intelligence and machine learning are present in many aspects of our lives: opening your phone with face recognition, your banking system detecting fraudulent activity, or even using a GPS to find your way to the nearest coffee shop.

When discussing AI, machine learning must also be brought into the conversation. One common misconception is that AI is the technology running the show. Take a closer look and you will see that it’s the other way around – ML is one way of achieving AI. “Artificial Intelligence” is an umbrella term for a broader concept where machines can execute tasks in a strategic way. Machine learning relies on the assumption that machines will be able to learn and adapt through experience. If you were to ask a human and a machine a question and you can’t tell which is a machine, then AI has been achieved. In this blog, we will discuss the power that ML and AI hold in the UC world.

AI and ML for Unified Communications

ML and AI play a significant role in UC. Data can be analyzed from systems and users with ML algorithms – signaling events and errors, endpoint characteristics, device characteristics, network characteristics and server inputs – pinpointing and addressing issues before they become incidents. ML can also be used to enable targeted user engagement. In creating a persona profile, data such as typical communications usage and quality, endpoints and devices used, frequent contact and session profiles, and common communication times and locations are all useful input considerations. As a result, more automated user engagement is possible, which can in turn enhance UC quality. By following user activity closely, the ML algorithm can deduce the root cause of an issue and propose a solution. AI then steps in to take this information and incorporates decision-making to provide monitoring and management software with recommendations for how to resolve issues. An example of this is Unify Square’s  patented technology, which enhances visibility and provides actionable insights into UC environments.

Without artificially intelligent machines, machine learning in UC would not be possible. For example, a simple day-to-day task such as a conference call is artificially intelligent. The active speaker quality and probability of the call are calculated using ML to optimize the experience. With this calculation, ML can proactively identify and resolve call issues by incorporating data collected from the users at the end of the call. This in turn, helps prevent similar problems from recurring in the future.

Artificial Intelligence Improving the Meeting Room Experience

While AI meeting analytics are still growing more robust, their impact and upside today is already enormous. There are several ways a company can take advantage of AI and ML for UC meetings. Along with several updates, Microsoft revealed a new feature that proves how AI can be put to significant use; a top-of-the-line whiteboard feature uses computer vision to display the part of the board behind the writer. Studies have proven that organizations who adopt AI into their UC systems see success in communication, collaboration, information access and UC endpoint control. This successful outcome is a result of the increased efficiency and effectiveness that is brought about by improved interaction with data.

One example of how companies can augment the meeting room experience is Natural Language Processing (NLP). This technology translates raw human speech into meaning, powering voice assistants and translators. NLP-powered virtual assistants are omnipresent in the consumer world, and they have also rapidly expanded to business. AI-powered chatbots rely on core ML technologies like NLP and Information Retrieval (IR) techniques. By applying such methods, tech giants like Facebook and Google have released open-domain multi-turn chatbots- Meena and Blender, which have advanced to be able to provide human-like conversations.

In recent news, Microsoft announced a new Teams feature that combines recording and transcription in tandem. This live transcription with speaker attribution feature can identify each speaker, capture voice inputs automatically as the meeting is happening, and is available during and after the meeting. Teams improve accuracy and recognize meeting-specific jargon for each transcript automatically by utilizing the meeting’s invitation, participant names, attachments, and other components, without any human involvement.

With employees working across the country or even the world, we are bound to have different terminology. Localization allows participants to hear phrases with which they are familiar. For example, while participants in a meeting room in San Francisco talk about “highways,” their counterparts in Philadelphia hear “turnpikes.” Furthermore, it provides international teams with the flexibility to use idioms that they are comfortable with and confidence that the meaning is transferred to the other team.

The Future of work eBook 2021

How is hybrid & remote work progressing? Uncovered key findings about workplace communication, collaboration, and productivity.

Sentiment Analysis and Digital Management Solutions

An additional subfield of NLP is sentiment analysis (SA), a method for detecting emotion based on a person’s choice of words. Building on sentiment analysis, emotion recognition also incorporates tone and computer vision-based facial analytics to better provide an objective assessment of emotions. These automated systems can look for several specific properties in facial images, including the six basic or universal expressions: happy, sad, angry, surprised, scared, and disgusted. Systems such as FaceReader can also recognize a ‘neutral’ state and analyze ‘contempt.’

SA is also a core building block for Digital Experience Management (DEM) solutions; DEM can provide a holistic view of a user’s entire digital experience, including remote workers. It can also help IT align technology KPIs to business metrics, such as revenue, churn and Net Promoter Score (NPS). As customers demand a smooth, effortless, and high-speed omnichannel experience, a digital experience management platform helps companies align their channels, manage their content, and keep their digital ducks in a row. According to a recent study by Gartner, by 2025, 70% of digital business initiatives will require infrastructure & operation leaders to report on business metrics for digital experience.

Practical AI Applications in PowerSuite

Going even further with NLP, Unify Square’s data science team introduced department mapping for PowersSuite functionality. NLP powers this feature, which can standardize departments into 16 functional groupings. You can find department standardization in PowerSuite, within the Policy Management dashboard.

UC systems are continuously generating a significant amount of data. As a result, it is important that any collaboration monitoring software being used by IT can pinpoint security risks. Rather than dig through each workspace to manually alter the settings, PowerSuite’s AI algorithm determines if there is a substantial risk associated with a particular team or channel based on a variety of factors. By doing so, IT can focus on areas that are most vital, instead of wasting time on low-risk areas.

Security issues can also lead to high-cost toll fraud situations for UC. Approximately 50% of IT professionals surveyed stated that they spent more money on investigations and toll interventions than they did on the fraud itself. To solve this issue, Unify Square’s data science team developed a one-of-a-kind process to establish patterns of fraud across multiple tenants. To date, PowerSuite has already detected over half a billion dollars’ worth of toll fraud instances for various PowerSuite customers.

PowerSuite Demo Request

Industry Leading UC & Collaboration Management. Receive Actionable Insights & Analytics for Zoom and Microsoft Teams. 

The Future of UC with AI and ML

From what we have discussed above, it is important to point out that machine learning is not necessary for AI in UC. AI and ML have an interdependent relationship that shouldn’t be mistaken for interchangeability. Both have their own functions, but they are also better together. For example, if an employee encounters problems in a conference call, IT can use ML to identify the problem and propose a solution. This is all made possible with the help of an artificially intelligent software system such as our PowerSuite™ software.

While the current set of advantages of AI and ML in meetings offer unique opportunities for innovation within UC and collaboration, the possibilities for AI in UC are endless. AI and ML in Unified Communications is more powerful and advanced than ever before, thanks to increased data and computing power.

To learn more about artificial intelligence and machine learning in UC, check out our On Demand webinar covering the growing role of AI in UC from Alan Shen, VP of Consulting, as well as this article in Computer Weekly, coauthored by Shen and two of our closest partners (Arkadin and Jabra).

Share on facebook
Share on twitter
Share on linkedin
Share on email

Read Our Other Latest Blogs

Shopping Basket