Exclusive Talk with Ryan McDonald: Chief Scientist at ASAPP

Ryan McDonald is the Chief Scientist at ASAPP. He is responsible for setting the direction of the research and data science groups in order to achieve ASAPP’s vision to augment human activity positively through the advancement of AI. The group is currently focused on advancing the field of task-oriented dialog in real-world situations like customer care. In such dynamic environments, there are many interacting parts: the conversation between customer and agent; the environment and tools the agent is using; different measures of success; a wide range of customer needs and situations. Optimizing this environment in order to lead to quality outcomes for customers, agents and companies require significant research investment in retrieval, language generation, constrained optimization, learning, and, critically, evaluation.

Ryan has been working on language understanding and machine learning for over 20 years. His Ph.D. work at the University of Pennsylvania focused on novel machine learning methods for structured prediction in NLP, most notably information extraction and syntactic analysis. At Penn, his research was instrumental in growing the fields of dependency parsing and domain adaptation in the NLP community. After his Ph.D., Ryan joined Google’s Research group. There he researched sentiment analysis and summarization models for consumer reviews, which resulted in one of the first large-scale consumer summarization systems consumed by millions of users every day.

Q1. Tell us about your journey in AI.

Ryan: I’ve been working on language understanding and machine learning for over 20 years. My Ph.D. work at the University of Pennsylvania focused on novel machine learning methods for structured prediction in NLP, specifically information extraction and syntactic analysis. After my Ph.D., I joined Google’s Research group focused on sentiment analysis and summarization models for consumer reviews, which resulted in one of the first large-scale consumer summarization systems consumed by millions of users every day. While there, my team was instrumental in the development of Google Assistant as a global technology by building out many multilingual capabilities. After over a decade working on consumer products, I then shifted gears towards enterprise and led numerous NLP and ML projects to improve Google’s Cloud services, including the core NLP API, solutions for Call Center AI, and knowledge discovery from the scientific literature. My research on enterprise NLP and ML now continues at ASAPP.

Q2: Tell us about your role as Chief Scientist at ASAPP and your leadership style and philosophy

Ryan: At ASAPP, I’m tasked with setting the research agenda to realize our vision of augmenting human activity positively through the advancement of AI.

My experience has been that industrial research labs that are successful in the long term are those that have a culture of execution excellence to drive business objectives. As such, I am a strong believer that strong industrial AI research works backward from present or future business outcomes in order to develop a program that can be broken down into a series of short-term objectives, each of these testable. So one of my main jobs is to ensure that we start efforts by thinking about what are the outcomes we care about: are these useful? How large is the impact? Etc. From there we focus on measurement: what data do we have or need to collect? How hard will it be to get that data? What metrics can we measure and optimize that are correlated with the outcomes we care about? Only then should we think about AI models and solutions.

I have a lot of great teammates at ASAPP who are capable of amazing things. This dictates my leadership style, which is more focused on making sure the top-level objectives are aligned with ASAPP’s short/medium and long-term goals. Once that is done, I mainly focus on ensuring we have the resources in place to execute as well as working to remove obstacles.

Q3: In terms of some of the biggest challenges in the customer service and call center space, How can ML, and NLP help to improve the customer support/ agent experience?

Ryan: The customer experience and contact center industry often finds itself in a tricky balance between lowering costs while attempting to improve the quality of customer service. 

Before widestream adoption of AI, companies used a variety of methods to lower costs but suffered a lower quality of customer care in the process. “Containment,” a measure of having customers solve their own issues without human intervention, was seen as a key way to lower costs. This often came in the form of simple rule-based systems such as an interactive voice response (IVR) or chatbot which used FAQs to help customers solve their issues. And, to reduce customer wait times, new agents were given abbreviated training periods. Unfortunately, the confluence of these efforts created scenarios where customers never had their issues resolved through self-service means, and agents encountered high turnover rates from lack of training and support from automation. 

Today, almost every stage of your interaction with a call center could be driven by AI or already have AI informing how the issue is addressed. After a customer connects with an agent AI can guide and make suggestions to the agent. What should they say next? What flow should they follow? What knowledge base articles will help solve the problem? While companies should be optimizing AI in these ways, what we are finding is that most still do not. Such models are best trained on historical data and optimized for some key performance indicators, which can handle time (how quickly the problem was solved) or customer satisfaction score (was the customer happy with the experience). 

Once the call or chat is over, AI is still at work. In most call centers the agent will leave structured information and notes about what happened during the call. This is for analytics purposes but also for any subsequent agent who picks up the issue if it has not been resolved. AI helps with all these steps.

Finally, there are supervisors who are there to help assist agents and grow their skills. AI can be critical here. In a call center with hundreds of agents handling thousands of calls a day. How can supervisors identify the issues that need their intervention? How can they understand what happened during the day? How can they find areas of improvement for agents in order to grow their skillset?

At ASAPP, we’ve also found that while real-time dynamic guidance for agents is critical, more structured training, coaching and feedback is also important. Many agents train on new issues or procedures ‘live’. That is, they get a description of the procedure, but then only see it in practice when they take a call with a real customer. Imagine we gave pilots the manual of the plane and then told them to fly 300 passengers to Denver? Because of this, we are focusing on using AI to help build tools for agents to practice procedures and handle difficult situations before they deal with live customers. When this is coupled with targeted feedback (either by a supervisor or automatically) this will allow the agent to grow their skills in a less stressful environment.

Q4: Can you share a little about ASAPP’s AI Services and Platform? 

Ryan: ASAPP’s AI services integrate into existing CX environments and support people in being their best by predicting what an agent can say and do throughout every customer interaction, automate tasks within their workflow and continuously retrain the models to ensure increased accuracy and ultimately impact. 

Our AI platform has a particular emphasis on empowering agents through a host of modular AI services. Ready via API, SDK, or plug-in options, we offer the following services:

• JourneyInsight – Analyzes agent activity in depth, identifies ways to streamline
• AutoCompose – Crafts quality agent responses for digital messaging
• AutoTranscribe – Delivers highly accurate speech-to-text transcription
• AutoSummary – Creates high-quality disposition notes automatically
• CoachingInsight – Provides real-time visibility, tools to guide agent performance
• AutoWorkflow – Automates time consuming tasks for agents during interactions

Each of these is designed and trained to optimize for key business outcomes. Specifically, these services focus on the joint objective of improving the experience of the call center customer and the job satisfaction of the agent. ASAPP customers also derive greater value when multiple services are together. The network effect of using multiple AI services makes every one of them better for you. 

Q5: How is ASAPP bringing its AI technology in a way that is different from its competitors?

Ryan: Our central hypothesis at ASAPP is that AI should augment people in positive and productive ways. This takes shape in our research and product strategy which has a particular focus on the agent and their experience. We combine our domain expertise to create AI models tailored for the contact center and customer experience use-cases. 

The AI-driven results speak for themselves. An airline customer saw agent productivity increase 86% and a rise of organizational throughput (total number of interactions across all customer service channels divided by labor spent to satisfy those needs) by 127%. For a global network operator, their net promoter scores (NPS) (the willingness of customers to recommend a company’s products or services to others) increased 45%. For a telecommunications company, their cost per interaction decreased 52%. These examples show how AI, designed for people, can increase productivity, improve the quality of customer service, and decrease business costs.

Q6: Can you shed some light on the latest employment trends related to ML and NLP? Is your company hiring?

Ryan: Yes, we’re hiring! We can’t comment on behalf of all ML and NLP jobs, but at ASAPP, prioritized areas of ML research and engineering are in speech, ML engineering, and task-oriented dialog. 

Researchers at ASAPP work to fundamentally advance the science of NLP and ML toward our goal of deploying domain-specific real-world AI solutions, and to apply those advances to our products. They leverage the massive amounts of data generated by our products, and our ability to deploy AI features into real-world use to ask and address fundamental research questions in novel ways.

Suggested