Catch is an AI agent focused solely on administrative tasks. We don’t write code, create images, or analyze data. Our single mission is to build the most advanced agent you can trust to delegate your daily workload, expertly handling scheduling, reminders, email, and travel logistics.

Genuinely proactive, Catch operates with the initiative of a human employee. It is available wherever you already are, working seamlessly over the phone, WhatsApp, and iMessage to keep your day moving forward. Because true delegation requires absolute peace of mind, Catch is built on highly secured infrastructure and is fully SOC 2 compliant, ensuring your data is always protected.

Test Catch today to see how it handles the friction in your workflow.

Set up Catch today!

This interview has been edited for length and clarity.

Nick Beadle: The big focus of what I’ve been writing and researching lately is how we can actually get people into jobs and remove barriers to their getting them. Obviously, what we call “validation”—or an employer verifying a worker has the skills—has a big part of that. Tell me about the work of GenAI Studio and how it might play into all of that.

Sophia Romee: AI has really changed the fabric of how we think about jobs. There's a lot of anxiety overall, and the way I think about it is there's really a dual crisis. It's both how we think about what will be the jobs of the future [and] how we assess and prepare students, and that is where College Board comes in.

How do we really make sure we're continuing to think about defining what authentic learning and readiness looks like at a time when traditional measures of achievement are being transformed? And this is also why the GenAI Studio, which was an AI incubator I started a little over two and a half years ago at College Board, became a thing. One of our mandates was thinking about new forms of assessments and how we think of that measurement piece.

Nick: What is an authentic education experience? How might that play into shaping who students ultimately will be in the workplace?

Sophia: To me, there are a few things to tease apart here. The first is for students themselves, there are questions [like], “What does it look like to think of the skills that I'm still going to need in the workplace?” What is authentic learning in that context, right? Where College Board has landed and our commitment is to say the future demands a mix of these foundational skills that we've always felt were important and we're assessing things like writing, right? Math, right?

And we think increasingly the durable skills matter even more than they did in years past. Things like teamwork, things like communication and really making sure we're all still exercising what I find to be uniquely human even as AI enters the [conversation].

The other thing we think a lot about, given we are our bread and butter is assessments, is just the impact on how you think of validation. Cheating has become a headline that I don't love to be honest. That has circulated in the media of like, “Oh, students are cheating. It's rampant cheating.” And there is a kernel of truth in the idea that if you are a student—I've talked to many of them—sitting there and it's 1 in the morning and you have a thing due the next day it can be very tempting to say, "OK, I might use AI for this thing instead of fully grappl[ing with the work]." One thing we're thinking about is how do you build the next generation of assessments where you need now new signals of readiness and thinking of assessing with AI as well as without AI and how we have [to have] both options and be more holistic in the ways we deliver assessment.

Nick: I read not too long ago that liberal arts majors after years of being beat up politically over and over have pretty good preparation for the AI-involved workplace because they have learning in ethics and debate and all those broad-view skills needed for AI integration.

Sophia: I mean, you're talking to an anthropology major, a former debater. I have no formal technical training [in AI] whatsoever. And if you told me five years ago even that I would be running an AI incubator and in this space I would have laughed. So through lived experience, at least, I can validate the trends that we all are seeing.

ChatGPT is a superpower if you know how to use it correctly.

Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.

Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.

Download the free guide.

Nick: If you talk to a new college grad, they already are hearing a million different things as to how they should make themselves employable, and that was already stressful. Now, they can't get through the robot hurdles. They’re applying for hundreds of jobs while they’re brushing their teeth, and crazy stories like that.

Is there a way for AI to better capture that lived experience so candidates don’t have to do so much work to maybe break through and employers can see they’re a good, authentic fit earlier on?

Sophia: I love this question because I think we spend too much time talking about all the negatives of AI, to the point of like the media fueling these headlines of cheating of societal job loss like all these things [are] just not at times I think productive. What you're getting at is [the question of] are there things that we can actually be more intentional about in using AI in supporting this transition period. I think yes, for sure.

In our internal workforce enablement work, one of the things that I think has been a real win for the org is thinking about how we really help the organization write better goals. It's such a small thing in some ways, right? But as you know, as somebody who's been in the workforce for a long time, writing a clear goal that is strategic, that is tied to priorities, that is aligned with the things you want to do—and then kind of having a back and forth dialogue with AI in whether or not that makes sense to you and how that relates to the things you're going to focus on—really helped people understand the why behind using AI as well as actually helped them translate it into impact in their day-to-day job. And I think those things are very much connected.

Extrapolating that example and connecting to what you're saying, the idea that AI will help you coach you on having a more relevant resume or help you sort of translate some things you think you have done that are kind of random into a more coherent narrative that matches the job you're applying to. [Those] are examples of great application [of AI] that should be talked about far more than they are.

I don't want to undermine the difficulties that young people are going through right now, but I do think there are applications there and they don't have to be, ‘Let's try to game the system.”

Nick: Now here's where I'm gonna ask you about all the negative stuff in the news you were just talking about with AI.

Sophia: Yeah. We should talk about all of it.

Nick: AI-powered applicant tracking systems [or ATS] are shutting out a lot of jobseekers right now, based on what I hear from them. Some of the tech may be underdeveloped, and I have heard stories that make me think we’re losing qualified talent. Not just because of volume, but because the systems are a little too rigid. How do we improve those systems so employers aren’t losing good talent due to rigid systems or systems with some bad, old human attitudes baked into them?

Sophia: Can I ask you what you think? I'm happy to give you my take, but I'm curious where you are with it.

Nick: I had a conversation with a tracking system vendor about their ATS at a convention. They misunderstood what I was asking about and thought I was looking for an ATS to shut out anyone without a college degree. And they told me, “Oh yes, our ATS takes out anyone without a college degree to make sure no one ever gets in the pool without a college degree.” I was asking about the opposite because I’m a big fan of skills-first hiring removing degree requirements and focusing on skills.

So I think things like that are still hanging out in the code that need to be thought through.

Sophia: What I'm hearing from you and things I have seen myself and read about is that there's a ton of bias in problematic practices embedded in some of the machine learning. This predates generative AI, as you know.

One of the first guiding principles that I wrote down when we founded the studio was, “You start with the problem and the user needs instead of the technology.” We had made this commitment, we continue to make this commitment, but we don't just give you the GenAI solution because you think that's what the answer should be. We actually first really understand what is it that we're trying to solve and why and what's the problem in that kind of vein.

Knowing that's the kind of vision I have for the things you do with AI, I think, Nick, what you're alluding to are people not focused on the right problems. Because if we're not naming these as the right problems to solve, then you are probably going to be either very ineffective in the technology you're deploying or frankly almost okay with what the outcome is, if you get my drift.

Nick: I totally do. Something I learned as an employment lawyer was that a lot of people put a lot of themselves into making their hiring decisions. By that I mean sometimes people are trying to validate their own life experiences by looking for someone as close to them as possible.

Sophia: That's right.

Nick: That gets soaked up by the machine learning as well.

Sophia Romee: Totally.

I’m thinking about how late policies seem to be in catching up to the realities of what is happening. For example, in the K-12 context, when we did our research and put out the survey, one in five school districts had no real AI policy. Because of that, it was this haphazard mixture of not knowing if it's OK to use AI [or] how to use it.

Right now, you're kind of having an educator decide what to do. Everybody's confused. I have a feeling that same analogy applies into the employer context as well because if you don't have a clear policy on what you're trying to do in terms of hiring and what you want, then, yeah, how will an individual talent manager or recruiter know what to do and what not to do?

Tell him the campaign. Viktor pulls last quarter's performance from Meta and TikTok, scrapes competitor ads, drafts the brief, posts it for review. You edit, he ships the creative requests to your designer. Inside Slack.

Get Started for Free

Nick: A lot of people are playing by ear. I mean, even if they have a policy or they find out they have a policy after the fact, hiring is only one part of their job.

Sophia: Yeah. I grew the team from 10 to 25 in one year. So, as somebody who has hired quite a lot of people, I will say it's such a time-consuming thing to do and to do it well, especially. I can imagine folks who want to try to shortcut that using AI will find good ways and bad ways to do that.

Nick: Obviously there is a lot of scary talk from folks who run the companies that are at the heart of AI. What is the best possible outcome based on what you're seeing right now in terms of AI in the workplace but also AI and education?

Sophia: I think jobs are going to be here.

Nick: That's a good start.

Sophia: What I have seen again and again in working with teams at College Board in thinking about AI workforce transformation is AI is just the final layer to help a team understand how they can transform [how teams do their work better]. Predating that piece of workforce transformation is often non-technological changes. I'm talking about things like process re-engineering. I'm talking about things like your data infrastructure needing to be different, etc.

It’s an interesting learning [experience] for me over the last few years. We might have grand visions of how AI is going to transform the workplaces of and make us head into this future but there's a ton of work to be done to really understand and realize AI potential.

I think a really great outcome would be that every company has understood the inherent in like foundational things they need to fix or improve upon or maybe reimagine to translate where they are now and really understand how AI can help them be the next, better version of themselves. But you need to do and put in the work there.

I've seen all of these companies coming out there and promising a bunch of things to you of how radically you're going to just change [because of AI]. You know, like, save all this time or increase your ROI.

I just don't think there's a magic wand here. I think it's going to take some time. At the company level or the team level, I'd say, to me, a pragmatic vision of successful AI integration is you've identified really clear areas that you want as a company to solve or change because those problems make sense.

Then you have thoughtfully integrated where those process shifts need to happen, the data shifts need to happen, and then where AI comes into the mix of that. Then at the individual employee level, you have a level of comfort in judgment on when to use AI, when not to use AI, and you understand the inherent risks that you are right now being exposed to by increasingly more advanced models.

Nick: To maybe oversimplify and put this in therapy terms, everybody’s got to do the work.

Sophia: That's exactly right. Employers have to do the work. I think employees have to do the work. You can't both just stare at each other and say, “You do it, you do it.”

I would say it's also important for everyone to understand that the evidence is still thin on whether and how AI improves learning outcomes. And I say that as the person leading an AI shop and someone who obviously is optimistic about its potential and what it can do. I think it's important for us to be honest about that.