It’s a phenomenon that those in the startup world are well aware of. A founder has a great idea and a cool website, but they end up shutting down soon after launch because they couldn’t find paying customers.
Perhaps the idea was too early or hit the market during an economic downturn. Maybe the business was doomed because of co-founder infighting or lack of funding.
But more likely, the founder was “not asking the right questions or not hearing what customers were actually saying,” says Atlanta entrepreneur Matthew Burris.
Most startups fail because founders fall victim to confirmation bias when talking to customers, adds entrepreneur Toly Shilman. And simply put, bad questions lead to bad answers lead to bad decisions.
Eliminating bad questions is the focus of CogBias AI, a new Atlanta-based startup co-founded by Shilman and Burris.
CogBias AI’s first tool QuDisc (Question Discovery) is helping users detect hundreds of different kinds of biases that can creep into a question – from confirmation bias to desirability bias to anchoring.
It is best to think of the tool as a “copilot” for teams. After users drop their work into QuDisc, the AI breaks down each sentence and pinpoints where different forms of bias are found. Just as Grammarly shows how a user might restructure a sentence with better syntax, QuDisc gives suggestions on how users might rephrase a question to reduce different kinds of bias.
The tool can also be impactful once users get back responses, since the tool can help discern how bias crept into different answers.
Of course, the AI platform cannot completely get rid of bias. But it is designed to aim human bias in the “right direction,” said Shilman. That’s crucial for questions around developing products, hiring people, or conducting product research.
The ultimate goal is to help founders and companies get to their “product truth” faster by giving them actionable data, Shilman added.
There’s a reason it is hard to answer a question like: ‘Do you think my startup idea is great?’
The reality is that a lot of questions are posed to reinforce the asker’s beliefs. And in business, that can cause teams to build the wrong products without understanding customer wants and needs.
While still in its beta phase, the tool has found some interesting use cases already. The team said VCs have run their investment thesis through the platform to learn what phrases could be biasing the types of founders that reach out. A venture studio used it to understand if an investment memo had any flawed arguments.
Hypepotamus even took the QuDisc beta for a spin and asked the platform to analyze how common questions we ask founders during interviews could be skewing responses.
Get To Know The Team
Shilman moved to the US from Ukraine in the early 1990s, and built up his career in corporate healthcare before turning towards startups. He was an entrepreneur-in-residence for the Flashpoint accelerator program at Georgia Tech when he first met Burris, an electrical engineer by training who worked across industrial design before moving to Atlanta to join payments giant NCR in 2017.
The two entrepreneurs ultimately started working together as partners at Savvy Tinkers Studio, a local venture studio focused on growing B2B startups. Over the years, the two saw firsthand that cognitive bias can undermine team cohesion, cause a company to build the wrong products, and ultimately hurt a company’s bottom line.
The team has bootstrapped the product to date.
Now with QuDisc’s beta officially launched, the team is ‘practicing what they preach’ by listening to early customer feedback to evolve the platform.