Introduction

Companies in nearly every sector rely on stock imagery for ads, blog posts, marketing, and more. But there’s long been a significant lack of representation in stock photography, creating a widespread ripple effect in the media consumers see daily.

In a 2023 study, researchers analyzed five major stock image websites for health-related imagery and found an evident lack of diversity in the provided images. Most of the search terms they used resulted in little to no results depicting people with darker skin tones, and this was just the case for health-related images. 

The issue remains widespread, affecting businesses in many different markets. This not only means that companies can miss out on reaching their target market but also that audiences are not adequately addressed and served by companies.

Now, with AI, these concerns could amplify, and startups need to be aware of how to use AI and other media responsibly as they work to scale.

That’s where innovation and understanding the market come into play. Startup founders will need to thoroughly understand customers’ pain points, pinpoint a gap in the market, and have a clear solution to these concerns in order to succeed. 

The founder we’re spotlighting today was able to do just that and, as such, has made a significant impact on audiences that is even more far-reaching than his startup’s own client base.

A conversation with Steve Jones, Founder of pocstock

Steve Jones, founder and CEO of the stock imagery startup pocstock, saw the lack of inclusivity in the stock image market when working for 15 years at a marketing agency. The team Jones worked with had to go to great lengths to find diverse imagery on popular stock sites, which gave him the idea to create a stock image startup that represented the global majority.

The result is pocstock, a photo library representing people of all races, cultures, ability levels, ages, and body types. The library has hundreds of thousands of photos, videos, and illustrations. pocstock works with huge ad companies and brands, along with smaller businesses, to improve the diversity of the media consumers take in every day.

Jones sat down with HubSpot for Startups to discuss his journey as a founder, from fundraising and growing a startup to making a long-lasting and wide-reaching impact through his company, which is currently scaling successfully through continued innovation and strategic partnerships.

HSFS: What has your growth journey been like with pocstock?

Steve Jones: We launched a minimum viable product (MVP) of the platform in January 2020 with 1,500 images from one of our former co-founders. Currently, we have over 600K assets in our library. We started with three customers in 2020. We have over 1K now, resulting in probably 6K users at 1K companies. 

In terms of other growth, we've sourced over $1.7M in outside investments through a series of seed rounds. Right now, it's all about customer acquisition to scale our revenue. We're trying to reach a break-even point this year and then profitability going forward. 

HSFS: What was the fundraising process like for your startup?

SJ: We’ve raised $1.7M so far. We have another $225K outstanding in this round, so that'll bring us to $2M, and we're closing that in the next 45 days, hopefully. I am diluted down to about 70% of the company. I've sold off about 15% of the company to investors I met through networking and serendipity. We have an investor relations person who spent 30 years in finance and then fell in love with pocstock, helped us raise money, and left a 30-year finance career to join a startup to help us continuously raise money. 

I love people who say they like something and then take action to support it, and that's genuinely what this guy did.

HSFS: Can you share more about the role of inclusivity AI in your company?

SJ: We have two platforms. We have pocstock.com and pocstock.ai, which is coming soon. So, we approach AI from two perspectives. 

First, we work with generative AI platforms to provide them with inclusive data to round off their data sets so that if they're generating images via prompts of people of color, they have the data to support accurate representation. 

If someone prompts, “Show me an image of a Black woman from Jamaica with dreadlocks,” it won't show some woman who just happens to have dark skin. But the AI doesn't know the difference between braids and dreadlocks or the difference between some of the other cultural elements of certain outfits that people will wear that only if you're part of that culture, you would know the difference. So, our data is very granular in those ways.

AI platforms find it very valuable to have our data included in their data sets, and since we're the largest Black-owned stock company in terms of library size, it's helpful to include that amount of data. It makes an impact on the overall results. 

We're also building our own AI platform targeted at photographers and stock photography companies, which will be a productivity tool to make sure that the data they generate is more inclusive. So yes, it would be great for us to help solve two of the biggest challenges preventing generative AI from becoming as popular as it could be. AI is slated to add $7T to the GDP over the next five years or so. We want to make sure that it can do that and that people of color are included as producers and not just consumers of this technology.

HSFS: What can other founders and company leaders be thinking about as they incorporate AI tools when it comes to inclusivity and representation? 

SJ: I caution people. I spoke at the American Marketing Association last month; we were talking about regenerative brands, and one of the things I said during that conversation was really about how eager people are to supplant human intelligence with artificial intelligence. It's scary because, as someone who had an idea about how smart AI is before building our own AI, a lot of the platforms are generally very low in terms of accuracy. ChatGPT has released some models that are 85% to 90-something percent accurate, and that's Mercedes-class accuracy. 

People are relying on this information as if it's not indicative of the misinformation and disinformation that we have in larger society. AI only knows what it's been trained to know, and if it's taking data from our larger pool of publicly available information, it will be inaccurate just by virtue of bad data in, bad data out. And people don't think that way—they think AI is just smart because it's faster, cool, or whatever the case may be. 

My only thing is that yes, use AI as a copilot, a buddy, or an assistant to help you do things quicker, but don't rely on it to take place of your own intelligence. That would be a huge mistake. There are a lot of cool tools coming on the market that can help you write copy, newsletters, and press releases, but you still have to then go back and make it into your own voice and humanize it as much as possible. 

Before there was the wheel, before there were airplanes, before there was television, before there was the internet, these advents and technology accelerate human growth in ways we couldn't imagine. So there's nothing wrong with AI in that way; it's just that people don't see how early-stage it is because it's made available on the market by people we know and trust. People trust it, and they should not trust AI. I speak to many CMOs and marketers, and they are so eager to trust this thing, and it's just scary.

HSFS: Have you seen any situations of irresponsible use of AI by brands? 

SJ: Oh, absolutely. I wouldn't say it was irresponsible, but there was a publication that did something on Barbie. It was well-meaning, and they tried to do a Barbie with different ethnicities, and AI pulled from all of the worst stereotypes known for people of color. It militarized the Middle Eastern Barbie. It made the African Barbies into poverty or just the child soldiers. It just took all of the worst stereotypes you would find in a bad subReddit post and made it the standard. That's really the danger of it.

I wrote a LinkedIn post about an MIT student, an Asian young lady who uploaded an image to a platform called Playground AI and said, "make me look more professional." It stripped away all of her Asian features, and by the time it was done, she looked like a young white woman. It changed her lips, eyes, hair, skin tone, everything about her because that’s what AI has been trained as to what is professional. Why? Because in the larger society, we have to all change our standards of what is professional as what's set by the current professionals, and most of them are white. 

I mean, you see all of the attacks on diversity, equity, and inclusion (DE&I), but really, for AI truly to work the way it's intended, with the best of intentions, it has to solve racism. You can't have a racist AI because AI is going to be far-reaching at some point. AI will be used in court systems to determine how someone should be sentenced. 

If you look at how many Black and Brown people who've been exonerated with DNA evidence who were incarcerated during the drug war, when the U.S. painted young Black and Brown men as thugs and super predators and all those things, and locked up 2M Black people, now a lot of them are getting exonerated because that was all incorrect. Imagine AI was trained on that data, and someone new got sent to this, and AI says this group of people is inherently violent because the data says so, not because racism said so or a broken, corrupt system said so. It's because the data suggests that they are. AI would then continue to over-sentence and keep that trend going. 

It would exacerbate every negative bias that we have in society. Generating inaccurate images is one thing; that affects advertisers and marketers, and it may impact some people who it may trigger something for them, but then there are areas where it could have consequences for people's lives and healthcare, right? For Black women, healthcare has not been favorable in terms of outcomes with pregnancy, pain management, and things like that. As AI becomes something that more and more sectors of our lives—from airport security to policing—there will be so many applications and use cases for AI, and it's just scary how inaccurate the data is.

We have to solve that problem so that AI can have clean, ethical data to operate from, and it benefits everyone. Racism just needs to go. It's so corny. It's stupid. No person is better than another. We're all people. At the end of the day, we're one race of people, the human race, and that's it. 

HSFS: We love what you’re doing, taking direct action to combat this issue. It’s incredible. To switch gears, you’ve recently partnered with Canva. Can you share more about the impact of that partnership on your startup?

SJ: Canva is the most amazing partner ever. We work with some really good brands, but we don't work with anyone like Canva. The way it came about is on LinkedIn, there was a post from a loyal Canva customer who was posting that Canva does not have enough diversity content, and the thread went semi-viral. A handful of people just kept tagging pocstock.

It just kept bubbling from there, and I took it on. I didn't know anyone at Canva, but I'm a LinkedIn guy. I stay on LinkedIn most of the time. So I reached out to all the folks at Canva in partnerships and said, “Hey, here's what your customers are saying. Your customers say we should work together to help solve this problem.” And they said, “We absolutely should.” 

This was in October or November. By February, we were already in a global partnership contract, signed and off to the races by May. We were uploading our first batch of content into their platform, and now all of their customers worldwide have access to north of 150K assets that Canva’s customers now have access to. We'll be growing that over the years. But I know I see a lot of meetings on the books with us and the Canva team, and we're really integrated, and they're really open. They're not folks who say, “Hey, I'm about this diversity thing,” but don't put their money behind it. They’re behind it.

Steve, Canvo co-founder/CEO Melanie Perkins, and pocstock Chief Relationship Officer, DaSean Brown. (Image courtesy Steve Jones from LinkedIn)

If you're a B2B business, you have to be on LinkedIn. If you're a B2B founder, you need to spend X hours daily on LinkedIn, connecting with people, engaging with content, and all of the above. Don't assign it to your admin or any of those corny things because it won't work. No, people see you; they want to hear from you. They don't just want the generic copy-and-paste messages you'll get from an admin. But for LinkedIn, if you want to build a personal and professional brand, you need to show up and put in the work. There are no shortcuts. 

I have over 13K followers on LinkedIn, and that's a lot for me in terms of social following, but it is a really active following. I go to networking events, and people will come to take pictures with me because they follow me on LinkedIn. You know what I mean? It's like, okay, it's interesting. I didn’t know I had that cachet.

HSFS: Do you have any other advice for fellow startup founders?

SJ: Look for local resources. We partnered with Audible, which funded us to launch our office. There are a lot of resources for small business founders and startup founders; you just have to look for them. One of the biggest things for founders is access to capital. Cities have programs geared toward helping by providing technical assistance and sometimes access to capital through community development financial institutions (CDFIs). They will provide microloans. They will provide connections to investors and folks like that. 

So I'll say this to any founder looking to scale a business and maybe exit one day—not someone building a small business to pass on to their kids—but if you're building something to scale and exit, you have to raise money. You also have to realize that it's better to have 20% of a $1B company than 100% of a $1M company.

Conclusion

When it comes to creating and scaling a startup successfully, it all begins with the right idea. For Jones, that meant noticing a weakness in an industry he dealt with daily and coming up with a solution to the problem he and his colleagues were facing. Ensuring your startup’s product will solve users' problems and fill a missing piece in the market is the first step to success.

From there, it’s all about scaling strategically. As Jones explained in this interview, working with investors, being willing to share some of his ownership, and seeking out smart partnerships have all helped his startup continue to grow from just three users to over 1,000 corporate clients, each with dozens if not hundreds of employees regularly using pocstock’s library. That doesn’t even account for the thousands more people who will come across the media created from visuals sourced via the pocstock library and feel represented.

It’s clear that startups must do the research on their markets not only to find ways to improve their target customers’ lives but also to seek out the right investors and partners to power their growth sustainably.