Gathering quality, insightful, unbiased market research is always a core challenge for new product and service designers.
Existing market research methods used to gather user sentiment are mostly based on qualitative, subjective assessments such as interviews, surveys, and focus groups. These collection methods have their benefits, but they are inherently limited in scope and rely on human experience. Therefore, they are susceptible to bias and a lack of diversity and inclusion, which is a growing problem in today’s globalized market.
Existing online-based sentiment analysis methods are also limited, relying on things like general favorability metrics from users, such as like/dislike ratings left in reviews. These ratings don’t give a complete picture and leave a lot of unknowns about the user’s needs. Furthermore, there is a lack of computational methods available to translate insights such as ratings into more comprehensive new design knowledge, creativity, and success.
To help solve this problem, Assistant Professor Mohsen Moghaddam and his team, including Associate Professor Tucker Marion and Professor Paolo Ciuccarelli, have founded Advanced Design Augmentation (ADA) Technologies, LLC—a start-up founded from research performed at Northeastern University. The goal of ADA is to use AI algorithms and machine learning to better inform new product and service development by more deeply understanding customer needs and enhancing human innovation.
This research has earned them selection as one of the Spring 2022 Spark Fund awardees.
Performing Market Research with an AI-powered SaaS Platform
The ADA team is working to create an end-to-end AI-powered SaaS platform. Their platform builds on recent advances in machine learning and AI to draw new insights about latent user needs from the internet and then translate those needs into new, high-quality, desirable, and feasible products or services—creating more innovative and socially-aware products and services.
Their algorithms and software link user data with new and novel digital product designs through new generative design and concept evaluation methods. This link allows designers to include a much more inclusive and diverse set of user data, fine-tune specific attributes from that data, and automatically generate design options without active designer input.
“Our technology will allow a much richer and more diverse evaluation of data and resulting concepts, says Marion. “That may help designers create new products and services that more closely match the needs and desires of users.”
The team is creating scalable and computationally efficient Natural Language Processing (NLP) algorithms that will enable large-scale, user-generated data processing from online reviews. Their recently patented technology, Deep Multimodal Design Evaluation (DMDE) model, does this by leveraging large-scale user and product data to accurately predict how users will receive and rate a new design concept.
The team is also working on advanced Generative Adversarial Network (GAN) algorithms that will automatically process latent user needs to help improve the quality and diversity of the design concepts generated by human designers.
These AI algorithms will help designers draw insights from hundreds of thousands of user reviews, identify patterns between various attributes of a product and their contribution to success or failure, predict how people will react to a new design, and generate hundreds of unique user-centered design concepts.
“Ultimately, however, human designers are still essential in making the fine-grained design decisions,” says Moghaddam. Therefore, this AI technology will never replace designers—it will only enhance their work and break any traditional barriers such as a lack of knowledge about user needs at scale, bias, and creativity blocks.
The system is designed to increase the quality, desirability, and feasibility of design options, thereby reshaping how designers and engineers approach the front-end of the new product development process.
Commercialization with the Spark Fund
Through commercialization, the ADA team strives to foster designer-AI co-creation and innovation centered on empathy with users and bias mitigation to help bridge the gap between user need discovery, social impact, and design.
With the help of the Spark Fund, the team is actively developing a first-generation platform to be tested with initial industry partners in Fall 2022.
“The level of support and mentorship Northeastern provides is invaluable as technology like this move from the lab to venture,” says Marion.
In addition to the Spark Fund, the team has also filed two patent disclosures with the CRI and executed a licensing option agreement.
Learn more about Assistant Professor Moghaddam’s research and the other Spark Fund Award grantees here.