Sophia Xing: The Product Manager Crafting AI-Driven Products That Balance Efficiency and Sensibility

As we see apps and programs like Lensa and ChatGPT making big waves in the online world, it is clear that AI is here to stay and that it brings incredible innovation with it.

Of course, these tools are not without controversy. Many people see AI technology as a threat that will eliminate jobs and destroy the uniquely human creative and emotional perspective.

However, for expert product managers like Sophia Xing, the key is in how this technology is implemented to work with humans.

When brought to appropriate industries, AI holds immense potential for improving efficiency and maximizing profit margins.

Furthermore, when humans and AI can actually work together, the possibilities are endless.

A Background

Originally from Australia, Xing first encountered the powers of artificial intelligence and machine learning while conducting academic research during her time at Princeton University.

She ran topic modeling and sentiment analysis on newspaper articles that covered public equities, and then used the results as a foundation to construct a dynamic and risk-managed investment portfolio.

This experience gave her an understanding of the scalability of insights that can be garnered from using machine learning techniques on unstructured data, which is so prevalent nowadays.

In 2015, Sophia completed her BA in economics, statistics and machine learning, and applications of computing.

She then took a position with Goldman Sachs, as a portfolio manager in quantitative investment strategies, which was a very informative period of her career.

In this position, she consistently witnessed individual data vendors supplying sources of verticalized data to quantitative analysts, who would then comb through to determine the robustness of the data source to deduce useful insights.

Although there was great potential in the unstructured data, she was also blown away by the sheer quantity of noise that most industry practitioners had to work with before arriving at even the possibility of true insights.

Finding Direction

After five years with Goldman Sachs, Xing decided it was time to take the next step.

She was admitted into the highly-reputable MIT Sloan School of Management, where she excelled and graduated amongst the top 5% of her class.

The program provided some incredible opportunities, and Xing’s quantitative background enabled her to leave a lasting impact on many projects.

In 2020, she was brought on to MIT’s Covid Analytics Task Force, where she worked to construct a risk calculator in order to optimize hospital ventilator allocation in Massachusetts and neighboring states.

Xing also developed TalkAI, a unique product that provided an AI-powered audience in the AR space that reacts to the speaker’s tone and content, in order to help users overcome their fear of public speaking.

This project won the Best Tool category of the MIT Reality Hack in 2020 – one of the world’s largest and most reputable mixed reality hackathons.

Throughout her coursework at MIT, she mastered advanced natural language processing (NLP) and computer vision, laying the foundation for future aspirations. 

Outside of the AI space, Sophia’s financial and business background also enabled her to contribute immensely.

She served as the Partnership Director of MIT’s Clean Energy Prize of 2020, as well as the Sponsorship Director of MIT’s FinTech Conference of 2020. 

AI Meets Customer Support With Great Success

Sophia had the opportunity to work with some of the biggest and most innovative tech companies in the industry: Google, Bodeswell, and Curated.

Her passion for conversational AI helped her to land the role of Lead Product Manager at Forethought, an AI SaaS platform for customer service teams.

She has led the product management, from ideation to launch, of the company’s software that allows customer service managers, who may not have technical knowledge about machine learning, to be able to easily construct an AI-powered customer support agent for their company.

Xing’s expertise in NLP and software product development have been key to the platform’s success.

These AI agents are different from standard chatbots because traditional chatbots are decision tree-based and require extensive behavioral programming, whereas the AI agents that Forethought has developed can learn from a handful of input phrases and can then dynamically learn to improve their capabilities.

These AI-powered customer support agents are able to resolve customer needs with lower cost, faster turnaround, and they lead to overall higher customer satisfaction.

For Xing, the idea is not to eliminate human customer support agents.

Instead, by working hand in hand, AI technology can help to reduce the noise for human agents by resolving the tasks that can be automated and, in turn, allow them to give their time to the cases that require human sensibility.

She explains: “In a customer support context, AI is usually good at tasks such as finding the most relevant and recent information, processing tasks are routine and require layers of validation, conducting automation with integrated systems, and gathering feedback at scale to optimize business decisions. On the flipside, humans are usually better at exercising discretion when offering similar solutions, serving as a sounding board to make sure customers feel that their feedback is heard by a representative from the company, as well as processing one-off cases that may not be detailed in company customer policies.”

From her experience, there are certain steps that a company can take to provide its unique AI-powered customer support agent with the greatest potential for success.

For example, investing time and resources to create strong supporting knowledge articles helps to provide a solid foundation of knowledge for the technology, and can be easily linked without the need for human engagement.

Narrowing down inquiries based on the highest ROI can help to focus resources on engagements that will pay off the most.

Evaluating key terms, like “password reset” or “address change”, can help to take the easiest tasks off the plates of human agents as quickly as possible.

Lastly, by considering the customer persona, companies can help to direct their AI towards the lower-stakes customers first and leave human agents to engage with VIP or repeat customers that would require more personalized attention.

At a time when there are such strong emotions on both extremes about technology, it is refreshing to hear from someone finding the middle ground.

Sophia Xing is perfectly positioned to utilize the benefits of AI technology to improve the human experience in the customer service industry.

In reflecting on her work, she shares that “Customer satisfaction improvements at scale require thoughtful automation.”

Of course, the truth is that AI software can only be as thoughtful as the people who design them.

Thankfully, with Sophia Xing leading the way, we can look forward to a future with tech tools that make our lives a little bit easier.