Making Machines Human: How research-tech startup Quilt.AI is Indexing Humanity at Scale
How can we get people to care about climate change?
What is the next big thing that is going to go viral?
How do we identify influencers more relevant to our brand/work?
Quilt.AI is a fast and scalable technology platform that converts big data signals into human insights. In simpler terms, we exist to help organizations make faster and better decisions using greater amounts of public data available on the internet.
Unlike traditional market research organizations, we are not limited by sample size (like in the case of qualitative research) or surveys (like in the case of quantitative research).
We use artificial intelligence (AI) to analyze big data through an anthropological lens. Our source of data for all insights is human behavior on the Internet. We help organizations engage with these humans (always in an anonymized, aggregated, GDPR safe way), while fulfilling our own desire to index humanity (the way Google has indexed information), and thereby develop greater understanding and empathy for each other- at scale.
Human Insights at Scale
What does this mean for a market research company in the digital age? How can the axioms of offline research be translated to cater to online research? How can we make sense of data when it is available in abundance, and when even ‘making sense’ can be automated?
Before we dive into the above questions, it is important to understand the workings of traditional offline research. We can say that there are three main cornerstones of offline research:
- Authenticity, represented by ethnographic research, is ensured by spending a significant amount of time with respondents — being present and listening deeply before arriving at any conclusions.
- Robustness, in that information is precisely recorded and the hypothesis is backed/affirmed with evidence. Rigour and methodological coherence are important in order to defend and (if needed) repeat experiments.
- Truthfulness, represented by an acute awareness of one’s assumptions and baggage while conducting the research. For example, for a semiotician, everything has layers; for a psychoanalyst everything is symptomatic; for an anthropologist everything is contextual.
Today, businesses need quick insights on impact, accurate feedback, and advice for future programming- and offline research can often be time-consuming as well as expensive to conduct at scale. Further, offline researchers cannot always go back and clarify ideas and details with respondents (if needed)- thus, limiting the ability to iterate. That’s where we come in.
At Quilt.AI, we don’t seek to disrupt traditional anthropology- rather, we continue to study people with the same authenticity, robustness, and truthfulness as traditional anthropology- but using the internet as our ‘field’.
The internet allows for self-storytelling at a speed and scale not possible before.
Multimedia uploads and searches are happening on the internet every second, and the volume of data being managed by businesses such as Google and Facebook is massive. For example, #berlin has 43, 169, 915 posts on Instagram (as of today). Who are the people behind these uploads and searches, and what does this information mean?
Such a large volume of data can be overwhelming, and difficult to understand and compute without context. For this purpose, we ensure the authenticity associated with offline research by layering demographic and psychological information which changes the meaning of data by helping us better understand who is speaking.
Further, knowing that each online platform exhibits a different type of performative element allows us to read/analyze it appropriately, as it provides context to why people speak the way they do.
Listen to our Co-Founder, Dr. Angad Chowdhry, explain this in more detail:
We are also continuously training our machines to detect the cultural context behind the data available online. Our research is designed and conducted by a diverse team of anthropologists, behavior scientists, marketers, semioticians, and data scientists. We use machines to scale our work and build an additional layer of context- very quickly.
For example, in the image below, we show you what a regular AI (on the left) detects in an image: beach, chair, outdoor, sun tanning, etc. Contrast this with what our Culture AI (on the right) detects keeping in mind the ‘context’ of the image: relaxation, summer, community, leisure, etc.
We are doing research with the same rigor and robustness of offline studies, but at a scale, that cumulative ethnographic studies of entire cities cannot match.
This is possible because there are more people online now than ever before (almost 60% of the world), thereby allowing us to study more people than have ever been studied before.
While traditional qualitative research relies heavily on creative storytelling due to limited data at hand, we have access to so much data online that the challenge of interpretation must be addressed by machines at the initial stages.
For example:
Tech-enabled cultural research allows us to use the assumptions and articulations of people’s own definition of the “meaning” of their universe.
We aim to study people through a cultural frame and understand the “meaning of A concept in Y culture”. However, what does “culture” mean? How many cultures and subcultures co-exist? What are the determinants of these cultures? Once we start studying billions of people across multiple cultures, how do we make our findings meaningful?
A limitation of traditional offline research tends to be that the assumptions and articulations that inform the study are that of the researcher, and not of the respondent. This limitation is easily cured in online research, as we build upon people’s own interpretations and articulations of their reality, thus, gaining insights that may not otherwise be as apparent.
Listen to our Co-Founder, Dr. Angad Chowdhry, explain this in more detail:
We are living in a world where machines are slowly augmenting our powers as human insights professionals. What does this mean for the industry when all information and analytical rigor is available at the touch of a button?
We don’t know the answer to this yet, but we are learning something new about humanity every day — and through this learning, moving towards a world that is hopefully empathetic.
Read more stories on Quilt.AI.
