Sprockets' AJ Richichi on Using AI to Make Intelligent Hiring Decisions

An interview with the company's CEO and founder, AJ Richichi
May 3, 2021, 09:01 PM UTC
Sprockets' AJ Richichi on Using AI to Make Intelligent Hiring Decisions

Welcome to BOH Banter, an interview series designed to introduce our audience of independent restaurant operators to the Verified Vendors on our Solutions marketplace. Have suggestions for who we should interview next? Pitch us: tips@backofhouse.io!

***

Every restaurant operator knows hiring is an uphill battle. It's hard to find good talent, and even harder to hang onto it. Plus, turnover is expensive. Sprockets knows this, and has built technology that aims to fight the hiring battle for managers.

The platform uses artificial intelligence to assess writing from candidates to determine how well they will match with the restaurant. Core to that mission, says founder AJ Richichi, is avoiding systemic bias in hiring.

Back of House spoke with AJ Richichi about Sprockets' inner workings, how the platform fights algorithmic bias in hiring, and why hiring technology is essential for restaurants.

This interview has been lightly edited and condensed.

Meet AJ Richichi of Sprockets

Back of House: Tell us who you are and what you do with Sprockets.

AJ Richichi: My name is AJ Richichi. I am the CEO and founder of Sprockets. My day-to-day is really around product. It's my job to ensure that we're providing the best service for our clients, as well as constantly innovating in the space and in predicting what the market is going to need. 

BOH: For people who've never heard of Sprockets before, what is Sprockets? 

AR: Sprockets is a screening tool used by restaurants to predict whether or not candidates will be successful in your particular restaurant, before you review their resume. The idea behind it is, you probably get a ton of applicants via Indeed and Craigslist and all these other platforms, and it's really a crapshoot as to whether you bring this person in. Should they start working here? Will they be successful? Will they stay long-term? Sprockets provides some data around if we feel like they're going to be a long-term employee, or if they're going to be a bad cultural fit and end up leaving in the first 90 days.

BOH: How does that work?

AR: When a candidate applies, however they're applying, they take a three-question online survey. Like, what are your short and long term goals? Or, what's your favorite movie and why? And what Sprockets does is it looks at the paragraphs that people write, and it looks at things like their word choice, their punctuation, their spacing, their use of emoticons, and their sentence structure, and, using natural language processing, is able to build really comprehensive personality assessments about each candidate, and then based off of that, map them to your restaurant to predict if that's a good fit, or a poor match.

BOH: If you're an operator, what's the onboarding process look like? You create a profile of your own on Sprockets so the technology can match up attributes that you've said are important to candidates? 

AR: Onboarding is just a 20-minute phone call. We get you set up on the system, and basically will instruct you on how to send our assessment to your top-performing people. You have great employees who have been with you for a long time — we're going to have them take that same assessment to build what we call a success profile: the aggregate mental makeup, and needs, values and personality traits of your top-performing people. So that when people come off the street, we compare them to people who have been really successful and stayed long-term, and we build some predictive analytics based off of that.

It's really important to know because I get asked this question a lot — we're dealing with very powerful tech, right? It's artificial intelligence. It's natural language processing. And to the outside world, it might seem really complex, and the engine is really complex. But the user interface is built to be so intuitive that any user of any technology prowess will be able to learn the tech in that 20-minute call, utilize it, and weaponize it to make better hiring decisions.

BOH: What's the value proposition for the operators themselves? What do they stand to gain by adopting this technology instead of running hiring the way they've been doing it?

AR: The name of the game is reducing turnover. Every operator knows the pain of hiring somebody, training them, and having them leave or give somebody bad service. So for us, we're looking at some statistics. It costs roughly $4,969 to hire, train, and replace an hourly worker. if we reduce your turnover, that's going to result in fewer cycles that you spend in hiring, better service for your customers, and overall, just fewer headaches around the hiring process that takes you away from what you're passionate about. It’s just really tough for a restaurant. And in our case studies, we were able to show that we increased 90-day retention by 43%, and we reduced screening time by 37%, putting more hours on your schedule, again, to improve the restaurant and do things that you're more passionate about.

BOH: Tell us a little bit about how Sprockets is thinking about staffing challenges changing or seeing staffing challenges changing because of the pandemic, and what Sprockets can offer to address those new needs.

AR: Bringing in a lot of people for in-person interviews can be stressful for a restaurant and the hiring manager. So when you apply more technology to the pre-screening process before somebody gets to that point, it's going to limit the amount of person-to-person contact that you need to make with somebody to do an actual interview and make an actual hire. So that's one thing: safety. The second is really preparing for this [post-COVID] mass rehiring event. We don't know if that mass rehiring is going to occur in three months, six months, or nine months. But there's going to be a day here in 2021, where people start coming out again, people with cabin fever start eating in, and we have this huge push to the restaurant scene. A lot of restaurants aren't staffed for that right now. Over probably a 90-day period, there's going to be more hires made in hospitality history than ever before. Our technology can help with making sure that the people you bring in are the right fits, stay throughout this chaotic time, and become long term solutions, not just warm bodies to address the influx of customers.

It’s gonna be very competitive to get the right people to work with your restaurant. You have to make sure you have the right people that are passionate about the work you're doing and make it so that you don't re-enter that cycle of turnover, rehire, turnover, rehire. It's a really interesting time for restaurants, but I think there's a lot of technology addressing it, and that restaurateurs just have to do their best as to picking the right tools and tying it up correctly.

BOH: What does Sprockets look like in the candidate application process? Is it just a layer that fits into the existing application workflow?

AR: It's not a social network. [You don’t] have to create a username and password. It's a simple link that can be delivered to the candidate. They just answer the three questions, and then they go on their way. If they come in for an interview in person, you would just text them the link. If they apply via Indeed, they get automatically sent via our integration. However you communicate with people, that's how you would send it, and it’s just those three questions. 

BOH: How does Sprockets address potential bias or potential algorithmic problems with weeding out the wrong people? How does the technology address that? 

AR: It's important to note that our technology is compliant by all governing bodies as it pertains to diversity — Department of Labor, the EEOC, as well as the Society for Industrial and Organizational Psychology. The way that we see the space is that hiring today, without technology, is horrifyingly discriminatory and biased. If you look at the statistics, white-sounding names get 50% more callbacks than non-white-sounding names, and the average interview includes between five and seven illegal interview questions. It’s not all on the person doing the interviewing, but somebody who's a general manager of a restaurant probably doesn't have a Society of Human Resource Management certification to do an interview. It's somebody asking, “What year did you graduate high school?” That's an illegal question. “Do you have kids?” Illegal question. A lot of things that you wouldn't think about that open your company up for liability happen during the interview process. When you talk specifically about our technology, we built the system to create a more equitable hiring process for candidates. 

The way that we look at language, we're not analyzing if somebody is smart or dumb, or if somebody comes from a family from Fairfield, Connecticut, and base the scores based off how well they use language. Our language takes cultural and socioeconomic things into consideration when building the reports and ensuring that there's a level playing field.

I'll use this as an example: If the candidate is talking to you, and they use the word “hella,” you might say something about them, or you might think something of them automatically. But the way we've worked is we've clustered “hella” up with “a lot of” and “superfluous” all to mean exactly the same thing within the data's perspective, to ensure that if somebody says any of those three things, we assign them the same score and the same values to be compared against the success profile the same way. So basically, we're listening, taking things into consideration, and leveling the playing field, and not just creating a blind hiring process, but making it more equitable. And that's really important to the core of our company. 

It's complex, and there's a lot of research and science behind it. Although our company is relatively new, we use science that's 80 years old. We just have new vehicles to use it and bring it to market in an effective way.

BOH: Can Sprockets offer value to an operator who already seems to have a pretty good hiring system?

AR: I think I can help everyone. In addition to personality screenings, we're also launching things like background checks, tax credit validation — a pre-hire suite to help you navigate the sifting of candidates in a multitude of ways. So if somebody says, “My gut is smarter than any machine would ever be,” I would say, “I understand. And if you don't want to work with us, that's okay. But we also have other products that might be able to help you with tax credits, as well as potential liability in the event something goes sideways.” 

BOH: How can Sprockets help operators dureing this "mass hiring event" that's looming on the horizon, post-pandemic?

AR: Operators can't be in every single room and in every single interview, and [in front of] every single screen. So let's say that there is a perfect human at analyzing the psyche of a human and how it maps into an individual work environment. They're not the person that's doing the first hard screening based on name and previous experience. They're not the person doing the phone interview. They might be doing in-person, but it's their gut picking hiring managers who are then using their gut to make a decision. I'd say even if you were perfect at selection, are your selections also perfect at selection? Having any additional information is a valuable tool, especially when there's so much at stake. 

BOH: Any last things that you'd like operators to know about Sprockets?

AR: They should think a little bit about the delta between their best employee and their worst employee. Oftentimes in restaurants we fall victim to the assumption that anyone with a warm body can come in and work here. Well, that's a really challenging thought process, just because all of your customers are interfacing with almost all of your employees. And as much as we as operators want to represent the company, we oftentimes don't have the opportunity in every single case. Whether you make one bad hire, or 10, or 100, over the course of the year, it can negatively impact your company. Deploying any screening tool, in my opinion, is just invaluable. I'm biased, because I love Sprockets, but I always highly recommend that if you don't work with us, consider working with other tools to help you make better hiring decisions because it is so tricky to get right.