Natural language processing (NLP) has been incorporated into everyone’s everyday life. Between Alexa and Siri, NLP is artificial intelligence (AI) technology to understand language as it is spoken and written.
New York-based LivePerson is a company focused on conversational AI, a technology that can help companies communicate better with their customers. LivePerson aims to help companies improve lives while accepting the new digital world.
Datamation interviewed Alan Gilchrest — Chief Technology Officer at LivePerson — who shares his perspective on the development and growth of NLP:
Gilchrest joined LivePerson in 2018 after working at Amazon. At Amazon, he worked on Alexa as the global head of dialog management, AI agent integration, and next action selection. At LivePerson, he is focused on scaling the company’s conversational technology through the approach of fusing conversational building blocks and AI-enabled conversational orchestration to redefine human-to-machine interactions.
See more: Top Natural Language Processing (NLP) Providers
Natural language processing Q&A
The NLP market
Datamation: What is your favorite part of working at LivePerson?
Gilchrest: Every day, we take on a challenge that’s really interesting to me: helping humans and AI work together better and more naturally. We’re always improving our AI to make it feel more human, like talking to someone who knows you and has earned your trust, and we’re working with one of the most impressive datasets in the world that focuses specifically on these consumer-facing challenges.
Datamation: What sets LivePerson’s NLP approach or solutions apart from the competition?
Gilchrest: At LivePerson, we’re focused on Curiously Human AI, which means experiences where AI understands consumers’ intents, connects them to brands, and delivers meaningful outcomes for everyone who is part of the conversation: consumers, agents, and brands.
Of course, building great AI starts with great data, and we have one of the world’s largest datasets for brand-to-consumer interactions, which helps us build outstanding experiences. In fact, our platform is home to nearly one billion conversational interactions per month.
We are also committed to reducing unconscious bias in AI and helped found a nonprofit organization called EqualAI whose mission is to do just that. Together with leaders and experts across business, technology, and government, we’re developing standards and tools as well as identifying regulatory and legislative solutions to increase awareness and reduce bias in the use of AI and are deploying these standards in our own products.
Datamation: What is one key NLP technology that particularly interests you?
Gilchrest: When it comes to AI that actually engages with you as a person, many people think about the Turing Test or the idea that we need to convince people that they are talking to another person rather than an AI. I don’t think we need to trick people or try to fool them, I think we need to make human-machine interactions more intuitive and efficient, so people will want to talk to AI. This aspect of AI is of particular interest to me and helps drive my perspective on building “more human” experiences as we develop our AI.
Datamation: What is one NLP technique that teams should implement?
Gilchrest: Companies can cut costs and drive revenue at the same time if they use conversational AI to scale customer conversations. Let’s face it: it gets very expensive very quickly to scale traditional communication channels with a large enough pool of agents available all day, every day. That means that although contact centers are critical to top-notch customer experiences, they are also cost centers for most brands.
However, leveraging conversational AI to support channels that consumers use every day to communicate with friends and family, like messaging, drives real efficiencies for companies and can be applied across industries. New enhancements, like routing and self-learning technology, integrations with apps that consumers use every day, and better ways to measure conversations means conversational AI is feeling much more like the conversations people have with friends and family on a daily basis. They’re also not only supporting customer care conversations these days, but also increasingly supporting commerce use cases to drive increased conversion rates and sales.
Datamation: What is the biggest NLP mistake you see enterprises making?
Gilchrest: Companies trying to scale their customer experiences sometimes rely on rules-based experiences that can lead to mechanical-feeling conversations, misunderstandings, and ultimately, customer frustration. These manifest as the major “fails” that customers experience with low-quality chatbots today. For example, consumers can get caught in a never-ending loop when a bot doesn’t understand what they are looking for, then asks or instructs the same thing over and over without providing a way to reach a human or try something else. Imagine trying to resolve a problem with your account and you can’t log in, but the bot keeps asking you to log in before trying to solve your problem. Companies should instead keep a human in the loop when needed, implement tools that understand when things aren’t going well, and set up appropriate routing and conversational pathways for complex issues (i.e., seamlessly transferring a customer to an agent when the conversation is out of the bot’s scope of responses).
Datamation: What are some current trends in the NLP market that are promising?
Gilchrest: Through recent surveys of consumers, we’ve found that positive sentiment toward chatbots nearly doubled in 2021 (61%) vs. 2020 (31%). As people are becoming more open to interacting with AI, companies are also evolving conversational AI experiences to feel more like having a conversation with a human.
Datamation: What are the biggest factors that are driving change in NLP/AI?
Gilchrest: AI is increasingly powering experiences across channels and uniting them, so consumers can have a seamless experience, no matter when and where they try to connect. A recent LivePerson report found that consumers no longer want to be limited to phone calls; 91% prefer brands giving them the flexibility to toggle between high quality, personalized messaging or voice experiences, depending on where they are and what they are trying to accomplish. AI will help brands deliver the holistic, seamless experiences that customers desire at scale, with the power to help brands manage millions of conversations at once and make them feel just as personalized as a human-to-human interaction.
Datamation: How has NLP changed during your time in the market?
Gilchrest: NLP and AI-led experiences are getting easier to deploy on the brand side and easier to use on the consumer side. New capabilities have been introduced that change the conversation, literally. For example, companies are turning to AI-powered dynamic routing to understand a consumer’s intent and sentiment, then automatically route them to the best qualified bot or agent. Dynamic routing no longer requires a ton of back-end configuration. It can be deployed quickly and easily through no-to-low code interfaces where you drag-and-drop bots and policies directly into conversational flows.
On the consumer side, conversational AI can also now use real-time signals from the consumer, like intents, conversation quality, and sentiment scores, to learn and improve on its own, deploying self-healing strategies to understand users better, reset conversations to a known good state, and delegate to other capable bots and humans. This means smoother and more helpful experiences that get you what you need, when you need it.
Datamation: Where do you predict the NLP market will be in 5 or 10 years from now?
Gilchrest: My hope is that AI-led interactions will become something humans crave and look forward to. We want AI to be a tool that helps us live better and even think better. Through increasingly more sophisticated and helpful AI, we’re going to teach humans how to be better communicators in life and to think and traverse logic more intelligently — if we get it right.
Datamation: What is one NLP technology your team wants customer care professionals to know?
Gilchrest: The ways we measure how conversational AI performs are getting much more sophisticated. Our Meaningful Automated Conversation Score (MACS), for example, is a way to measure how much friction customers experience in automated conversations. It produces a quality score for every single interaction with an automation, combining two critical factors: the customer’s response and the level of effort they have to put in.
Because it pinpoints which specific interactions have failed, MACS makes it clear what actions bot-builders should take next to improve conversations customers have with the automation going forward. Other metrics, like CSAT, only measure conversations in their entirety and don’t distinguish which agent or bot contributed to the customer’s satisfaction.
Long-term, MACS lays the groundwork for automating the entire process of bot building and management by identifying the penalties and rewards needed to train AI. This opens the door to new self-learning loops that eliminate the long hours employees spend on mundane tuning tasks, meaning they can focus on critical moments that need the human touch, plus gain experience with AI tools and systems. It’s a really exciting step forward for measurement and optimization.
Datamation: If you could give one piece of advice to a NLP professional in the beginning of their career, what would it be?
Gilchrest: You want to be where the best dataset is for the challenge you’re trying to solve. After all, you can’t build great AI without having the data to back it up and help it continually improve.
Datamation: With the shortage of tech talent, how is your team finding and retaining professionals to work in NLP?
Gilchrest: With our employee-centric model of working, we’re looking across the globe for the best talent. Being able to work in offices, remotely, or in a hybrid setup means there are no limits to finding the best of the best.
Work and life
Datamation: What is one of your top professional accomplishments?
Gilchrest: I’m proud of helping to establish an AI tech hub in Seattle for LivePerson, where many of our conversational AI and data experts work. I’m also very proud for the whole team for their hard work being recognized by Fast Company, which names us the #1 “Most Innovative AI Company in the World.”
Datamation: What is your favorite part of working in the NLP market?
Gilchrest: I love seeing better outcomes come to life. For consumers, we’ve seen how conversational AI can increase satisfaction and reduce friction. On the brand side, we’re seeing really meaningful outcomes, including 2x increases in employee efficiency, up to 10x conversion rate versus traditional digital experiences, 20% boosts in customer satisfaction, 90% automation containment rates, and 50% decreases in agent attrition rates.
Datamation: What is one of your favorite parts of the work week? How does it encourage or inspire you?
Gilchrest: My favorite part of the work week happens during our staff meetings, when we kick off talking about our good news. Each team member talks about something that is positive in their life (work or personal). It’s a great opportunity to catch up and express gratitude about all the positive things we can share with one another.
Datamation: Do you have a favorite way to recharge during the workday?
Gilchrest: Surfing is a great way to get some exercise in, especially now that we have a “be anyone, be anywhere” mentality as a company that means we respect how every individual works best in their own way, whether that’s in the office, remotely, or both. I moved to Hawaii during the pandemic and have been enjoying surfing more often!
Datamation: What are your favorite hobbies or ways to spend time outside of work?
Gilchrest: I’m an outdoors person and love cycling, swimming, mountaineering, and scuba diving, just to name a few activities.