The global health crisis has resulted in massive disruption to businesses across industries. Shelter-in-place guidelines have altered customer behaviors, resulting in dramatic impacts to the top and bottom lines for companies across a variety of sectors.
As a result, many IT organizations have faced budget cutbacks, with CIOs being asked to identify opportunities to streamline operations. These trends are prompting heightened interest in applying artificial intelligence (AI) and machine learning (ML) to help companies run faster and smarter. One third (33 percent) of CIOs and technology executives polled by HMG Strategy say that identifying practical use cases for AI and ML is their top priority.
For a growing number of companies, including Slack, LinkedIn, Broadcom, and DocuSign, AI bots already handle around 40 percent of employees’ IT issues automatically — without any intervention from the service desk. HMG Strategy recently caught up with Bhavin Shah, Founder & CEO of Moveworks, a pioneer in this space, to learn more about the pain points that Moveworks addresses for its customers.
HMG Strategy: What was the white space that you saw in the market that Moveworks addresses?
Bhavin Shah: We saw that AI was rapidly becoming a prerequisite for modern business — it unlocks a level of speed and scale that’s simply not possible for people alone. But AI had yet to reach the domain of IT support when we founded Moveworks, so we saw an opportunity to fundamentally transform a universal enterprise process.
The previous status quo was less than ideal. So far, we’ve analyzed over 75 million IT issues from across many organizations, and we found that the average one takes three days to resolve with a manual approach. Given that we’re now completely reliant on tech to stay productive, successful companies know they need to address tech issues in seconds — rather than days.
That’s where Moveworks fits into the picture. We use machine learning to automatically fix about 40 percent of IT support issues, which allows our customers’ employees and service desks to stop worrying about routine problems and focus on what matters.
How is Moveworks differentiated in the market?
Shah: IT support can be frustrating. And tons of effort have been put into fixing what is essentially a broken system. But many of these new solutions involve manually pre-scripting dialog flows, which, to us, are equally problematic. A solution should help the IT team — not encumber them with yet another time-consuming, effort-intensive process.
What we realized early on at Moveworks is that the key to fixing support was actually understanding the language employees use when talking about their IT issues. That’s why our engineering team has spent the last four years building hundreds of advanced machine learning models that all work in tandem to first understand — and then resolve — support issues. It’s this interaction layer that sets our approach apart.
A central component of our ability to understand language is recognizing that different employees have different IT needs. This approach was the impetus for creating our identity system which has up-to-the-minute information on a company’s IT environment. This system provides the extensive context needed to customize answers to each employee — putting an end to many of the problems associated with IT support. Our identity system provides the same contextual awareness that service desk agents use to naturally solve IT issues, but our chatbot can mirror that intuition instantly.
The last thing I’ll touch on is time to value. We know that service desks don’t have years to spend building dialogs and training machine learning models. So, we’ve taken the approach of pre-training our natural language understanding (NLU) to correctly identify 99% of entities from day 1. At the same time, our models continuously improve as they acclimate to our customers’ environments, meaning that the Moveworks chatbot quickly makes itself at home — increasing the percentage of autonomous resolutions with each employee interaction.
What are the top challenges that Moveworks solves for its clients?
Shah: Today’s employees are expected to be increasingly self-sufficient when finding the IT support they need — from troubleshooting VPN issues to updating their OS. The problem is, in many organizations, resources are scattered across dozens of back-end systems, while answers to IT questions are buried in multiple knowledge bases. Our role at Moveworks is to surface the right information to employees directly on their enterprise messaging tool. Rather than digging for what they need, employees just ask the Moveworks chatbot right on MS Teams or Slack to get help the moment they need it.
Beyond solving issues for employees, Moveworks also addresses a core challenge for service desks and companies as a whole: visibility over the IT environment. Visibility is a challenge for large organizations with many disconnected, legacy back-end systems. But Moveworks shines a light on what’s working and what’s not within this IT environment. With AI, the service desk has a clear picture of their IT ecosystem, possibly for the first time, helping them be strategic and find new opportunities for automation.
What are the primary benefits derived through the use of deep learning in the Moveworks model?
Shah: As I mentioned, our chatbot’s effectiveness is predicated on understanding the language of IT. And at its heart, NLU is a product of deep learning, without which our conversational approach would be impossible.
Most chatbots use “if-this, then-that” logic to determine next steps. The problem is that language is inherently ambiguous, and without context, no one knows what’s going on. Standard chatbots that take a deterministic approach, relying on manually pre-scripted conversations, are completely unable to absorb and respond with context in mind. This is why it’s incredibly hard for machines to understand natural language.
In contrast, Moveworks’ probabilistic approach opens the door for our chatbot to have naturally high-level and complex conversations with real users. Using deep machine learning models, our chatbot can adapt to ambiguous, contextual, and dynamic language. This probability-based decision making ensures that Moveworks can determine the right response to a user’s request on the fly, rather than in advance. We’ve spent years creating and operationalizing hundreds of machine learning models — to eliminate the language barrier between us and our machines. Because beyond just our IT support chatbot, we’re building toward a truly conversational future.
Why is Moveworks so well positioned in the work-from-anywhere environment?
Shah: During the initial shift to work-from-home, Moveworks allowed customers like Autodesk and Western Digital to weather increased support demand without increasing headcount. And moving towards 2021 and beyond — it’s become clear that we’ve shifted to an enduring new normal where location is no longer a factor when it comes to getting things done.
We’ve built our tech to complement this new normal. For employees scattered across time zones, rapid support appears unfeasible. But our bot works 24/7, picking up employee issues regardless of how or where they are submitted.
Crucially for our WFA future, a Moveworks bot answers employees’ questions — from IT to company policy — while accounting for location and other relevant data. Because our AI integrates deeply with a company’s IT tech stack and internal knowledge bases, we can surface snippet-sized answers that directly and quickly answer employee questions.
During our own company-wide shift to remote work, our internal Moveworks bot has been crucial, and we know that it will continue to learn, becoming increasingly indispensable far into the future. For our part, we see AI as an essential tool to empower employees, no matter where they work.
In this interactive digital roundtable where participants can ask questions and share insights, Bhavin will share examples of leading enterprise companies that are using artificial intelligence to provide real-time tech support to remote employees, autonomously resolve IT tickets via deep integrations and dramatically reduce the mean time to resolution of IT issues.
To learn more about this digital roundtable and to register for the event, clickhere.
