Robin Ducot, chief technology officer:
Reliability and accuracy are huge for us—because otherwise, what are we doing?

With our AI features and machine learning solutions, we continuously monitor output and follow best practices to maintain validity. We periodically retrain and compare prediction accuracy against the ground truth, otherwise known as known data. That allows us to assess quality and make adjustments when necessary to ensure our customers can rely on their data. 

Zoe Padgett, senior research scientist:
When we started developing Build with AI, I collaborated directly with our engineering team. I have a master's degree in survey methodology, and I used everything I learned in my program to inform our product. 

I created a checklist of well-established survey methodology best practices—things like making sure the response scales match the question, avoiding double-barreled questions, and optimizing question order. That checklist was vital as the AI team iterated on our feature, and anytime we make a big change to the models or prompts, I make sure nothing deviates from those best practices.

We also constantly evaluate our expertise. One of my colleagues on our product research team recently interviewed internal and external survey experts to make sure that we're following the most up-to-date practices in the field of survey research. Those insights guide our processes and help us identify gaps and areas for improvement.

Meera Vaidyanathan, chief product officer:
Trust is really the backbone of everything we do, and it’s why people continue to come to us to get the best insights out of their surveys. Transparency is key. Our customers need to know when we’re using AI, how our models are trained, and what data is used to train them. We want them to understand not only how their data may or may not be used, but also how we arrive at the insights that we give them.

And as always in software, it’s important to give customers the controls, should they want to turn AI support on or off. 

Zoe Padgett, senior research scientist:
I think that our response quality feature is so cool, and really easy to use. You can just toggle it on, and it filters out poor-quality responses automatically. The SurveyMonkey research team uses it a lot as we conduct our original research and reports, and it improves our data quality and saves us a lot of time. 

Eric Johnson, chief executive officer: 
Analyze with AI allows me to take hundreds or thousands of survey responses and turn them into something I can take action on. I can understand overall sentiments easier, and see what themes are arising over time. 

Meera Vaidyanathan, chief product officer:
The biggest hurdle for a lot of people is just getting started. If you aren't an expert in survey research or survey design, that is where our AI capabilities can really shine. Crafting the perfect survey question is an art and a science, and our AI makes it easy to get it right on the first try. That way, you know you’re set up to get the best responses. 

Robin Ducot, chief technology officer:
Our AI gives you a head start by automatically applying the lessons we’ve learned from billions of real survey responses. What sets SurveyMonkey apart is that we sit on a goldmine of response data across industries. Our algorithms have the benefit of 26 years’ worth of real data and experience with surveys. We're able to use that proprietary data set to train models to understand survey design best practices, predict response quality, and generate industry-specific benchmarks. Essentially, we’re using our history to make sure you get high-quality, professional results right out of the gate.