Jing Huang is currently serving as Momentive's Senior Director of Engineering and Machine Learning. As the leader of the machine learning engineering team, Jing Huang envisions empowering every product and business function with machine learning. Previously she worked at Cisco Systems for six years and was also an entrepreneur devoted to building mobile-first solutions and data products for non-tech industries.
While generative artificial intelligence models like ChatGPT have gained widespread attention, specific and task-oriented AI tools might be significantly more useful for companies and businesses. These tools are trained in specialized tasks, all with the intent of maximizing efficiency and performance.
Momentive, the maker of SurveyMonkey, specializes in this area, offering a range of specialized tools and services that leverage machine learning, a vast database of questions and answers, and experienced employees to help clients make the right decisions.
In this interview with Jing Huang, Momentive’s Senior Director of Engineering and Machine Learning, we discuss the current state of specialized artificial technology, the required balance between human input and AI, data security, and what the immediate future might hold for the advancements of AI.
Spotlight: Can you provide an overview of Momentive's AI technology and how it sets itself apart from other AI platforms in the market?
Jing Huang: At Momentive, we create solutions that help customers understand what matters throughout their decision-making processes. It all starts with asking the right questions, to the right audience, at the right time. Then, collecting responses and extracting insights in a timely manner. Our AI-powered technologies help our users throughout this journey.
Specifically, our SurveyMonkey Genius solution—powered by AI-smart scoring information, completion time and completion rate estimations — guides survey creators to create impactful surveys effortlessly.
Our Audience Genius solution enables customers to collect timely and cost-efficient responses from the right audience through AI-powered targeting recommendations, cost estimation, and completion time estimation.
Furthermore, our AI-powered Automation Insights solutions empower customers to uncover the most relevant insights quickly. It can be the sentiment from customers, employees, friends and family, or the market at large; it can be the themes buried in the long lists of open-ended text responses; it can be the statistical insights hiding in the closed responses; it can be filtering out on responses that are low quality.
Our team is positioned to deliver accurate, widely representative AI-driven insights with millions of new questions answered daily on our platform across more than 100 languages and in 100 countries. This translates to a living, breathing reflection of realities, not biases.
We champion this mindset at Momentive—from our diverse machine learning leadership team to our ML algorithms built on a diverse data set that includes 6 billion global respondents.
2023 is expected to be a significant year for AI. Could you describe the strategies Momentive has in place to leverage the growing popularity of AI and its advancements?
This year, we have seen tremendous excitement around generative AI, especially with applications like ChatGPT. We believe generative AI to be a generic methodology and solution to enterprises' problems that is reaching maturity, but its effectiveness toward specific domains requires significant domain-specific knowledge and data to fine-tune.
At Momentive, with two decades of accumulated data and knowledge in the experience management space, we are best positioned to provide the most effective solutions that leverage generative AI.
How does Momentive maintain a balance between technology and human touch in its AI platform to deliver human-centered solutions?
Momentive weaves together the best of humanity, artificial intelligence, machine learning, and billions of questions and real responses, for a unique approach that quickly drives better, more confident decision-making.
Throughout our 20-plus year history, Momentive has accumulated learnings from billions of questions and continues to collect millions of responses on our platform each day. We’ve already brought AI capabilities to market through product innovations such as our Automated Insights and SurveyMonkey Genius.
Intelligent technologies are only as effective as the data and models they are built on. We introduced human in the loop methodology for our latest open theming text analysis solutions. This innovative approach builds customized models for each customer, leverages human inputs’ domain knowledge, and provides tailored insights to enable our customers to act quickly and effectively with their unique use cases.
By harnessing the power of insights gleaned from millions of respondents to highlight potential biases and errors in survey questions, responses, and logic, we can effectively flag data and/or audience segments that our customers should pay attention to, whilst using machine learning to create industry benchmarks and guide users toward more engaging, less biased question types and answer sets.
Our embedded research frameworks and methodologies leverage the wisdom of human experts. Through a community of diverse minds, we engineer solutions to open the future for all of us, rather than restrict it for some of us.
Can you give an example of a successful implementation of Momentive's AI platform for a client, and what impact it had on their business?
AI is present in every part of our solutions and one strong example would be how the Golden State Warriors used Momentive to advance DEI across their organization.
Prior to the start of the COVID-19 pandemic, the Warriors used annual engagement surveys to connect with their workforce of more than 2,500 employees. This method resulted in feedback for a specific point in time, but the team realized that they needed more employee feedback more frequently.
They began using Momentive to send monthly pulse surveys, which not only fueled their COVID-19 response, but enabled them to parse data by business unit, department, and demographics. Those latter metrics accelerated the Warriors’ DEI efforts by helping them get a deeper and more nuanced understanding of the employee experience within and outside of the pandemic’s unprecedented challenges.
This is an excellent example of how AI is present in every part of our solution. The Warriors were no doubt getting a lot of data from such a regular cadence and analyzing all this data would take away from their day-to-day responsibilities.
AI empowers our survey analysis feature, highlighting the most impactful insights automatically, giving the Warriors the most up-to-date insights, they need to pivot and change, depending on feedback.
These employee check-ins, in turn helped the Warriors HR team keep up with the evolving sentiment and needs of all employees. It is a proofpoint that having up-to-date information on what all employees are experiencing is absolutely critical to properly engage with them and make good decisions.
What measures does Momentive take to protect survey data, and what can companies do to enhance their data confidentiality?
Momentive uses encryption, access control, and need-to-know processes to ensure proper handling of customer data throughout its lifecycle. Our global Trust & Security team works around the clock to monitor and manage our security posture. They are responsible for security compliance, education, operations, and incident response.
Data resides on our infrastructure, which is hosted and managed on public clouds. We select public cloud vendors that demonstrate and adhere to rigorous data protection processes. In addition, we perform rigorous security testing and maintain security incident response policies. These processes help us to adhere to 99.9% uptime of our web services.
One of our most important security strategies is to comply with and expand coverage of the industry regulations that matter to businesses. In addition to our SOC 2 Type 2 certification, we work with multiple third parties to audit and certify our products with ISO 27001, PCI DSS 3.2, and more.
Momentive’s Security Statement shares more details about how we protect data for both our enterprise and individual customers.
How does Momentive plan to continue innovating and developing its AI platform in the future?
Over the past couple of years, we have built a scalable AI platform that is adapted to our technology stack. It is powering more than 900 million predictions each year, supporting 24 unique capabilities.
Because of this platform, we are able to deliver new, innovative AI capabilities into our product suite in just a couple of weeks. In the future, we will continue to improve our platform in terms of scalability, maintainability, and efficiency.
We are also expanding our platform to support online training with human-in-the-loop methodology and expanding our platform to provide generative AI capabilities.
What do you expect will be the biggest AI advances in the couple years?
With the advancements in AI hardware and greater accessibility to large-scale language models, we will likely see more breakthroughs in specialized applications built on top of the generic models to solve problems in a specific domain.
These specialized applications will provide better AI explainability, improved AI security and privacy control, and better focus on fairness. They will be much more effective and powerful in solving domain-specific problems.
I like to think of this as the ‘solving the last mile problem’, similar to power plants and LED lights. When the power plants are built, getting electricity to every house capable of consuming and benefiting from it takes a different level of effort.
Thank you for your time, Jing Huang. Best of luck to you and Momentive!
Keep up with Momentive on LinkedIn and read about other successful businesses here.