AI, a two-letter abbreviation, which stands for artificial intelligence, is certainly a hot topic right now. It is stealing headlines, finding its way into our daily lives, and igniting debates about ethics and best practices. And, you’d better get used to it because it’s not going anywhere anytime soon. Local entrepreneur and startup champion Scott Meyer knows this, and he’s diving in head first.
Meyer’s entrance into the artificial intelligence world isn’t necessarily new—due to his position as an instructor for developers, he had already been developing AI courses in December—however, he is a leading member of the local charge into the new age, starting a group which meets biweekly called “Fargo AI” and, now, his own AI-based startup, BenchmarkAI.
Entrepreneurship is not novel to Meyer, who has an extensive resume. Already in his young life, Meyer has founded multiple businesses and organizations, worked as a city council member, launched an entrepreneurship center at North Dakota State, and so much more. We sat down with Meyer to discuss AI in general as well as his latest endeavor in the space.
Q: Can you explain what BenchmarkAI is, and when it started?
A: We started in mid-June. BenchmarkAl’s mission is to make public data more accessible and actionable. So, our first use case is with tax return data from nonprofits. With our tool, you can essentially search for any information about a nonprofit that they have to report publicly. That includes salaries of their key employees, how much they raise, how much they get from grants, what they pay contractors, what industry they’re in, what their mission is, who their donors are-things like that. We’ve made that information available on a searchable website.
It’s actually live at benchmark990.com. And right now, we have the 2022 tax return data in there. Currently, we’re going back and adding the additional years. One of the problems we have realized is that the IRS is really slow. The 2022 tax data is like 20% of all nonprofits right now because certain ones haven’t gotten through, or people have delayed filings or stuff like that. That’s a big thing we’re working on now. You can use our interface in two ways currently.
You can ask it a question like you would ask ChatGPT or you can go in and have a guided search. This might be useful for nonprofits looking to find people to reach out to for potential donations. I could also see it being useful for journalists. It’s really useful for finding out who is giving money to whom. Now, as we think about different types of customers probably the most profitable is actually service providers would be executive search firms, accountants, and lawyers.
Anyone who’s working for nonprofits could go in here and go into revenue and search for the nonprofits with the most revenue or for nonprofits between a certain amount. The second group would be foundations that are giving grant dollars to do due diligence on the nonprofits to see how much they made, how much they spent, and how they are paying their staff. And then there’s that large portion of nonprofits who are looking for donors. I think the other really cool part is we’re building out what are called “Action Agents.” So this is where we could actually get information and do a task for a nonprofit with Al. On our platform, we have an Al assistant who can do things like scraping a nonprofit’s website to find any email addresses that are associated with that nonprofit. And I can open up the email agent and it’ll actually write an email for me to that person. We can see a future where it could find grants and then automatically create them.
We built it out super fast. That’s the benefit of Al—you can build pretty quickly. The harder part is actually the data for us-we have to scrape the data. As you can imagine, a lot of nonprofits submit their data using PDFs. Right now, we’re exploring the nonprofit space but we’re really trying to find the right market. We could essentially use this technology with any public data. It could be used for land titles or legislation, political contributions anything that has massive amounts of data.
Q: So, there’s nothing like this that exists currently?
A: Not really, this is obviously inspired by generative Al. The difference here is we’re only looking within this data set.
Q: How’d you get the idea for this?
A: I’m working with Scott Holdman. He is really well-known in the nonprofit space. I’ve been exploring Al and am very interested in it. I spoke with him and it seemed like the nonprofit world was a really interesting use case where we could make a big impact. With our tool, you can really broaden your search. You could ask it to show you, ’10 organizations I should know about, and then go into those orgs and get the data. We knew with his knowledge we could build something interesting. His connection in terms of sales gives us a really nice way to test it with customers. That’s kind of what we’re doing now.
Q: What else do I need to know?
A: Building Benchmark990 unlocked an entirely new product. We needed a payment solution that charged our users every time they used our tool. We couldn’t find one so we built it.
It’s called Chipp.Al. We offered it to other developers and had 150 downloads the first day. To pursue the growth of this product, we raised $350,000 and will launch on September 22.
Q: What has the learning curve been like with adopting some of these new Al tools that are out there?
A: I’ve had an advantage because I was working as an instructor for developers. Back in December, I started making Al courses. So, I was doing a lot of exploration which was really early ChatGPT days. I also did quite a bit with image generation and ChatGPT. Those really got me excited. So, I’m an early learner, maybe not an early adopter. But then I just kind of tried to integrate it myself.
I’m always one that likes to learn in public. So, I started fivemin.ai, which is where I share tutorials and then do consulting for people that need training. That’s been really great because as I’ve taught other people, I’ve realized what questions people have and what the use cases that people need help with are.
Q: What advice would you give to businesses out there utilizing Al?
A: I have a YouTube video called “The Five-Min Al Method” where I explain my method for testing out integrating an Al platform. The first step is testing out the platform- don’t spend more than five minutes on it, there are so many other platforms out there. If if you don’t see any use after five minutes, you don’t have to keep playing around with it. However, if you try it, and there seems like there’s maybe interest or potential, then apply it to an actual problem. If it seems like it solves or makes that problem potentially better, then I amplify it, meaning share it with your team. I think that the third step is one that most businesses are missing. There are a lot of people on business teams that are using Al or exploring it, but they may be embarrassed to share that they are or they don’t feel like they can use it.
Al is really good at text. Al can really help a business eliminate a lot of the time they are spending creating text. ChatGPT can also do things like write a mission statement to a certain length. People worry about Al taking jobs, but nobody really wants the job of taking 5,000 words and making it 2,000 words. That’s not fun. Al is great for grant applications, emails, strategy, marketing, and communications. I do a lot with Al art and I think that’s really good for marketing and blog post kind of stuff, but works best as inspiration for actual designers. You’re probably not going to just publish what Al creates for you. The other framework that I use a lot is called the Al sandwich. So basically, it starts with a human, and then the Al creates something and then the human edits it and publishes it. I use that framework with every type of Al tool because, with almost all of them, you’re not going to be able to just ask one question and then get the thing you need. You’re typically going to have to ask, tweak it, maybe ask again, tweak it. So, you kind of need a human on both sides. But they can take out the mundane of copy editing or translating.
Q: Being someone that was on the early end of Al, are you surprised at all at how fast things have progressed here?
A: It is really fast but, also, ChatGPT-3 had been out for quite a while before they had the ChatGPT interface. So I think that it feels faster right now because ChatGPT-4 came out almost immediately, but we had ChatGPT-3 for about a year and a half. I think the time spent waiting for ChatGPT-5 will be a little bit longer. But right now, people are really catching up on the use cases that’s why it feels so fast.
I’m actually surprised that it hasn’t been integrated more into businesses because it seems so financially advantageous. Usually, when something will save you money and help you work faster, a CEO would want it implemented it’s not a discussion. I think people are hesitant and some don’t know the potential. That’s one reason why I started creating tutorials on YouTube. I get worried that when you’re in a community that’s not talking about Al whether that’s a professional community or a physical community like a city, you might get left behind. It’s harder to compete if you’re not using it. So I’d rather have Fargo companies using Al to have an advantage over a competitor. I also think sometimes people don’t know how to ask questions. They’ll ask once and get a bad answer or a boring answer and think that the software isn’t that good. You need to ask the right questions.
The Five-Min AI Method – Try It, Apply It, Amplify It
48 views * 2 months ago
Ed3 Scott Meyer
New Al tools are launching daily. How can you keep up? The secret is the Five Min Al Method: try it, apply it, and amplify it.
Scott Meyer is looking to help you and your business implement Al into your workflow. To do this, you can his a 1:1 Training or Al Consulting session with him at fivemin.ai or you can check out an number of his great YouTube tutorials and discussions at @FiveMinAi on YouTube.
BenchmarkAl is now welcoming beta users. Visit BenchmarkAl.co to sign up for a personal demo.
Q: Do you think companies should start employing a full-time person dedicated to Al?
A: I don’t think they need a full-time person. I would say it really should be an augmentation of what people in your company are already doing. Most companies have a CRM but maybe one or two people are sort of the champions of the CRM. I would say there should be somebody in every team or someone in the company that goes across teams and is thinking about how the team could work better with Al. Marketing is fairly easy to think of ideas. There’s a new feature called Code Interpreter that’s really good at crunching CSVs and financial data. Accountants and CFOs should kind of be thinking about that. It can make forecasts.
There is a lot of unknowns right now with Al, but one thing you can be sure of is that Scott Meyer is going to continue to explore!