Before founding TestGorilla I ran a social enterprise called Sukhi. We were an online store for handmade carpets and as we grew, we found ourselves needing to bring on more team members. Even though we were small we were inundated with applications, and we quickly discovered we had no reliable way of weeding through them to discover the best talent. The second issue was that as a social enterprise, we felt very strongly about making our hiring process fair. We didn’t want to give into biases and gravitate toward shiny brand names many of these applicants had on their resumes. Our initial solution was very simple. We built our own little handmade tests for job candidates in Excel — these included questions about rugs, creativity, cognitive ability — just to have something more objective to dig into at the top of the hiring funnel. It worked surprisingly well for something we threw together ourselves. And then I switched hats and went looking for a proper software product that did this process, but even better and at scale. Much to my surprise, I could not find one. That was around 2017 and that gap is what became TestGorilla. What makes TestGorilla different is the scientific foundation behind our solutions. Every assessment on the platform is developed by psychometricians, IO psychologists, and data scientists, and complies with the Uniform Guidelines on Employee Selection Procedures. Many of tools in this space are built around convenience, or simply going through the basic motions of vetting talent through the paths of least resistance. Ours are built around validity, hyper-focused on developing assessments that functionally predict who will actually bring the skills to perform in a specific role. TestGorilla was launched around the time that skills-based hiring was becoming a recognized term, and has scaled as that concept has seen more adoption. The fact that everyone from the BBC to Forbes is now talking about skills-based hiring is something we take some pride in having contributed to.

TestGorilla has 3 business priorities, and don’t ask me to pick which matters the most – they’re all my babies which means I love them, and we pursue them equally. Our core growth engine has been PLG: people sign up for the free plan and upgrade from there based on the value they see and their needs. Because everyone can start for free, the free plan motion is extremely important for us as well. Smaller companies could continue using the product and find value in their hiring with it, without needing to buy a paid plan. We’ll wait until they get to a certain size. And we’re one of the few B2B companies to offer a free plan which is adequate by itself. We also have a sales motion but it’s not a traditional enterprise motion. We don’t cater to companies with tens of thousands of employees. Of course, we have made the odd exception in the past but our sales motion is also in the 750-2500 FTE range. We initially didn’t know our limits – we now do, and operate successfully within them.

TestGorilla has fully handed over screening rounds to our conversational AI avatar. These interviews screen every applicant fairly – structured, science-backed, consistent questions for everyone. Our studies show AI delivers closer alignment to expert scorers than the average recruiter, providing fairer and more consistent evaluations. Before you pushback on candidates losing the opportunity to speak to humans, we didn’t build AI interviews to remove humans from hiring. We built them to remove inconsistency and give teams their time back – every candidate gets a fair shot at being considered and hiring teams get to spend more time with the strongest candidates When it comes to human way, the human in the loop is a compliance floor. We put humans at the bookends: setting the bar, and holding the verdict. Our AI interviews don’t select anyone and don’t reject anyone. They produce a score, a record, and a video, and then a person decides. AI’s job is to produce evidence. A person’s job is to decide. Internally as well, nothing we publish ships without a person putting their name to it. We hold ourselves to the standard we sell.

We’re noticing two major trends: – The buying journey is becoming AI driven, through a combination of citations and direct clicks – With AI providing thorough information about a product or service, comparison, risks and benefits, the multi-touch buying journey is slowly disappearing. Specifically, we’re tracking prospects coming to us via AI at two key stages: the research stage, where they’re using AI assistants to scope out options, and the final decision-making stage, where AI is helping validate or narrow down a choice. The volume and quality of inbound has shifted as a result of this.

The one thing I would tell people to actually do: define what AI fluency means for each specific role before you begin hiring for it. It’s very important to clarify that this is not the same as establishing a company-wide definition. Instead, AI fluency needs to be defined on a role-to-role basis. Think about what AI fluency means for a data scientist compared to what it may mean for a customer success manager or finance analyst. Defining fluency requires a tremendous understanding of role-specific nuance. Once you have a real definition, skip having candidates describe their level of fluency and instead have them demonstrate it. Name-dropping AI tools doesn’t work anymore. Instead, ask them about a specific moment when AI failed them. What they caught, what they changed. That one adjustment to the conversation surfaces more signal than most interview processes produce in an entire cycle. The one thing I would tell people to ignore: the idea that posting "AI fluency required" on a job description constitutes a hiring strategy for AI talent. The demand for AI-capable workers is real and it is not going away. The organizations winning that competition are the ones that have gotten specific about what AI fluency actually means in a given role and have embraced a process that can reliably identify it. The supply of genuinely AI-capable candidates is smaller than the demand — according to a recent Indeed and YouGov survey, only 19 percent of workers describe themselves as fluent in AI while 59 percent of employers say finding that talent is essential over the next year. In a market that tight, the organizations with the most precise evaluation process will hold a massive advantage over the ones still screening on the basis of keywords and basic credentials.