Improving Traditional Market Research

The #1 question customers ask us is “How do I know my market research is right?” You know it’s right the same way that we know any scientific research is right: It is repeatable.

When scientists do drug trials, they test the same drug in multiple markets and they know they have a solid treatment when each population responds well to the drug. If it succeeds in Boston but fails in New York, you don’t have a marketable drug.

With the advent of social media, researchers are now able to do repetitive market research trials on different platforms and when each platform gives the same result, they can trust that research.

The Hitch?

You need the ability to create representative samples in online platforms that match the offline population.

That is what our patented AI, “Polly” is able to do.

Polly’s Advantages

Since Polly uses much larger sample sizes than traditional market research, she has a number of advantages:


Polly can see into the future – much more accurate trending and forecasting.

Near real-time measurement

Polly is always available and always working. Your market research is up to date every day, not once a year.

Eliminate Bias

Direct questions can contain bias, and people often avoid answering controversial questions, especially when they are asked by a complete stranger. Polly doesn’t ask direct questions, eliminating bias and delivering accurate results and insights.

Scenario Testing

Because all studies are longitudinal, Polly can accurately determine how people will react to different situations and how their behavior will change.

How We Do It

We strive to be open and transparent about AI and how we use it to improve upon traditional market research. For details about our methodology and patents, check out these technical documents.

Who We Are

Erin Kelly

Erin Kelly

Erin Kelly is President and CEO of Advanced Symbolics Inc. (ASI), a company that uses Artificial Intelligence for human behaviour research.

ASI is known for its work in HR, consumer research, health care, tourism and government. The company is active in over 15 countries as their AI, “Polly”, speaks every language. Polly is particularly well known for her pioneering work identifying suicide risk in populations.

ASI gained acclaim after Polly correctly predicted the BREXIT vote, the 2016 American Election and the 2019 Federal Election in Canada. In 2018, Polly was chosen as the official pollster for Canada’s major network covering the election of Doug Ford. Polly was able to correctly predict outcomes for each of 124 electoral districts.

Erin Kelly is featured regularly in national and international media for the work her company is doing to advance research into human behaviour.

Erin in a Chartered Professional Accountant (CPA) with over 20 years experience in market research and advertising.

Advanced Symbolics holds two U.S. patents on population sampling for online media and works with major companies including GM, Mastercard, Fidelity, Greenpeace, The World Wildlife Fund and Health Canada, among many other companies and organisations around the world.

Kenton White

Kenton White

Kenton White has been working in the AI space since 2003, when he was co-founder and Chief Technology Officer for Distil Interactive. Distil Interactive used artificial intelligence to evaluate worker performance in training simulations. In 2009 Canadian Standards Association, North America’s largest corporate training provider, acquired Distil Interactive. Following the acquisition, Dr. White was a Professor of Computer Science, first with Carleton University and then with University of Ottawa. Today, Dr. White is co-founder and Chief Scientist of Advanced Symbolics, an artificial intelligence company, and Adjunct Professor of Computer Science with University of Ottawa.

Dr. White studied Physics at University of California, Berkeley (go Bears!) and holds a Ph.D. in Physics / Applied Mathematics from The University of Arizona (go Wildcats!). He is a past chair of the Canada’s National Science and Engineering Research Council’s Electrical and Computer Engineering committee and a Special Advisor to the Defence Science Advisory Board. Dr. White has published over 80 peer reviewed journal and conference articles, 5 patents, and has been a featured and keynote speaker at several international AI conferences.

His list of international awards includes:

  • US Department of Education Fellow
  • US Air Force Office of Scientific Research Fellow
  • SPIE International Young Investigator Award
  • Most promising new company of the year, Canadian New Media Awards
  • Best Serious Game, Arcademy Games Award
  • I/ITSEC Serious Games Showcase & Challenge Finalist
  • Best Product, DevLearn
  • IDC Canada’s Top 10 to Watch

Frequently Asked Questions

Are you complying with privacy regulations?

Yes. We access publicly available information shared on social platforms by consenting individuals. We do not access private communications.

Isn't online research biased?
We measure natural engagement on different subjects shared by online users. We do not ask questions, we measure authentic engagement in the subjects at study.
The samples of our study are representative of the entire population, allowing us to learn about how the general society feels about the same subject.

Is this the same as putting out a web survey?
No. We measure natural engagement with subjects. Unlike web surveys, we do not ask questions. Instead, we study how people feel about subjects without having to ask them directly. This measures real engagement and eliminates opt-in bias.

But not everyone is online.
True, but there are enough people online to generate representative samples. As technology evolves, fewer people are using landlines or are willing to respond to pollsters on their mobile phones. Online platforms attract different age groups and interests, offering a wide range of users who are actively engaging with a large variety of topics.

Is this the same as Social Media Listening?
Social media listening is a customer intelligence tool that informs on subjects being discussed. The results don’t allow us to draw conclusions on a larger population because the gathered information does not come from a representative sample, and therefore cannot lead to scientifically accepted generalization.
We study topics based on information gathered from large samples from targeted demographics. From this we can draw scientifically approved conclusions on the larger population.

135 Laurier Ave. West, Suite 100

Ottawa, ON


K1P 5J2

(613) 518-1644

Business Hours


9am-5pm EST