AI Leadership Insights: The Benefits of AI and Machine Learning

Amanda Razani: Hello, I’m Amanda Razani with Techstrong.ai and I’m excited to be here today with Jonathon Reilly. He is the co-founder and COO of Akkio. How are you doing?

Jonathon Reilly: I’m well. Glad to be here. Thanks for having me.

Amanda Razani: Glad you’re here as well. Can you talk a little bit about Akkio and the services you provide?

Jonathon Reilly: Yeah, we’re basically a next-generation AI and machine learning tool. And what we do is we make it easy for anyone who works with data to use natural language backed by GPT-4 to explore their data or do data transformations or cleanup and then build machine learning models on that data to sort of uncover the patterns that are driving key outcomes and then even deploy those models in a couple of clicks and use them in real-time decision-making. So think of us like an end-to-end business intelligence and machine learning tool.

Amanda Razani: Wonderful. So with that in mind, how is AI and machine learning being used to unlock more value in data and what benefits does this bring to various industries?

Jonathon Reilly: For a long time really taking advantage of machine learning with data has kind of lived in the data science team. These are really skilled data practitioners, usually working with very technical tools, and so they’re pretty much focused on the critical business problems in any given industry. Those are typically product problems or high-value problems, but in almost every industry and in every company I’ve ever worked at, there’s a really long tail of operational things that people do and all of those things generate data from sales, marketing support to customer success, you name it.
And we built Akkio to enable the people working in those parts of a business to start to use AI and machine learning in their daily lives to make their work with data faster and easier. And each business interestingly has a lot of really unique, really valuable data in how they operate. And a lot of the times, the way that that’s leveraged today is through backwards looking analysis where you create a graph of what’s happened in the past and then try and think about what might happen in the future. And so machine learning lets you actually get predictive and have a deeper understanding of the driving factors influencing a given outcome. But the lock there, the reason that it really hasn’t spread very far inside of businesses is because it’s really hard to understand or use. So at Akkio, we’re working really, really hard to make it easy for pretty much anyone who can use Excel to dig in and start getting ML-driven insights or using a large language model to make manipulating data faster and easier.

Amanda Razani: So it opens the doors to a lot more employees that may have not had those skills.

Jonathon Reilly: That’s right. And opening the door to more people allows a business to get more data-driven, like in everything it does, instead of in just like that very critical but small section that is typically ML-driven today.

Amanda Razani: Absolutely. So you believe this technology can accelerate the fourth industrial revolution, so can you explain that in a little bit more detail?

Jonathon Reilly: Yeah. I mean, I think there’s big sea changes in technology that periodically and when these occur, we call them industrial revolutions, but really, they’re new ways of working and I think they’re usually best identified by huge productivity changes, step changes in productivity. And so AI and ML is sort of a very hypey thing right now, but underneath all of that hype is a very serious set of tools that make people very much more productive when doing any type of work that involves working with data or even being creative.
And so today, if you’re writing copy, you should be using something like GPT-4 to make writing that copy easier, faster, and better. If you’re a software engineer writing code, you should be using copilot or something. Our engineers get twice as effective from using it. Designers, I don’t know if you’ve played around with mid-journey or stable diffusion, but you can start to generate images on the fly. And with Akkio, you can work with data really, really easily. The sort of key factor with all of these tools is they’re sort of self-serve and anybody can pick them up and use them. And so I think we’re at the very early stages of a big shift in productivity and how people work. And so I think that we will look back on that and call that a industrial scale revolution as it were.

Amanda Razani: Do you have any use case examples or journeys that you can provide as far as clients you work with or businesses that you can share their stories of how they’ve leveled up with this technology?

Jonathon Reilly: Sure, yeah. We’re building a platform, so the use ranges are really pretty broad. So I’ll give a couple of really different examples because I think they’re kind of interesting. So one of our earliest customers is this guy, Martin, and he runs a political fundraising company. They raise money for politicians by matching them with donors and they run call centers. I’m sure, you’ve gotten calls from people trying to raise political money before and not a lot of people donate, but a few do. And so that’s a productive thing for politicians to do. Martin was able to use Akkio to build a matching model that predicts the probability that any given person they call will actually donate for any given candidate, and then they can sort of stack rank the people they call by probability and start calling from the most probable to donate down the list and the dollars they raise per hour more than doubled.
And so these applications is classic ML applications, but without a tool that is sort of easy to use, that was sort of unapproachable. But we also have lots of marketing applications. So we have a big SaaS company that uses us for lifetime value prediction. So when a customer comes in and interacts with their website or browses their content, they store that data about them and they built a model that learns the patterns of which behaviors on their website are indicative, combined with the firmographic information of the customer, which behaviors are indicative of having a high lifetime value. And so they score, I think it’s like 300, 400 thousand customers a day, and then they take that score and they send it over to Google for their ad spend. So their advertising is associated with a higher spend customer, so they’re able to get really good value there.
Then we have a large range of consultants that use us to do data analysis for their customers. Everything from retail clothing, like predictions, like to large shipping companies, predicting carbon costs or shipping lane times and stuff like that, reasons things were held in customs. Pretty much anywhere you have some historic data in a business that has a pattern contained in it that’s relevant to an outcome, you can build a model really easily and leverage that to big returns and success. And I think most businesses today recognize the value of ML because they have those data scientists building models that are transformative. So the real trick is thinking about what it would mean if you were able to leverage that type of technology everywhere in the business down the long tail.

Amanda Razani: Absolutely. So from your experience, what are some of the roadblocks or the struggles companies have when they’re trying to implement this or other key technologies into their companies?

Jonathon Reilly: Yeah, I think there’s definitely a data journey that you have to go through as an organization, and it starts with understanding that most things are running on data-driven systems these days and leveraging those systems is really important to being competitive in your market. That starts with pulling the data together. And so we work with companies that are ways down that data journey. They typically have their data gathered in a centralized location like Snowflake or BigQuery, but we also have integrations with CRMs too, so it kind of depends.
You have gotten started down the path of actually capturing data that’s relevant to what’s going on and started to use it in some basic ways. So sometimes, we find people who come in and they’re excited about the technology and what it could do, but they don’t really have the data capture infrastructure or the organization in place to take advantage of it. But very many companies have been through that journey in the last 10 years or so and really come up the curve on it. So more and more we’re talking to people who are actually ready to start using it throughout the organization. And I think in the next five years, if you’re not doing it, it’s going to be a big anchor on your competitiveness. So I suspect everybody will be using ML in any data-driven workflow in the next five years or you’re not working with the right tools.

Amanda Razani: So considering current concerns about privacy and ethics, what measures are being taken to have responsible and transparent use of AI and machine learning in business?

Jonathon Reilly: Yeah. So it’s a really interesting question. The patterns in your data are like what you use with AI and ML in order to make predictions or take actions. In a way, it’s not so different than looking at data and making a decision based on that data. It’s up to the user to really think about doing the right thing with their data. And of course there’s laws and regulations surrounding that as well. And so what we try to do is we try to surface the driving factors or the things that the model is relying on to make its decisions as clearly as possible so that you can look at them and say, “Yes, that’s something that I can, while I’m legally allowed to rely on when making a decision.” Let’s say you were working in an HR capacity and you were screening resumes. There’s some regulations around what you can discriminate on or not, right?
And if your data happens to contain information that you shouldn’t be making decisions on, but the outcome is driven by, we surface that for you so you can see it and then ignore it from your model and not use it in the decision making process. But any tool, I think there’s a hybrid responsibility here, one, on the tooling vendors to make sure that they’re surfacing what’s happening and why it’s happening in the tool, so it’s really clear to the users. And then two, on both the users and I guess the industry and society to come up with regulatory frameworks around what’s acceptable and ethical and what’s not. I think most people want to act in the right ways when working with data. And in fact, because we’re making a platform, we don’t get a see so much how they’re using the platform or their data unless they choose to share it with us. But we do go out of our way to try and make it really clear like this is what’s happening with this model and why, so that you can make the right decision for you and your industry and your problem.

Amanda Razani: So last question. In two years from now, as rapidly as this technology is advancing, where do you see the enterprise regarding this technology?

Jonathon Reilly: I think almost everybody who works with data is about to get wildly more efficient. One of the things I think is really interesting to think about is a lot of these generative models like OpenAI’s GPT-4, they’re like a level playing field, right? Every company has access to the same model that is trained on the same set of data so they can leverage it to get equally as efficient. And the competitive playground there is how quickly you can come up the curve on using these tools to make your employees more efficient.
The competitive advantage in two years is your custom business data that nobody else has. And so the ability for companies to leverage that into competitive advantage I think is going to be a key differentiator two years out. I mean, today, it’s just get people using AI and ML and get them more efficient. But I suspect that companies who are able to gather data relevant to their outcomes and then leverage it in their decision-making, two years from now, will be really well positioned to be competitive in the industry and they’ll have a unique differentiated advantage. So at the end of the day, I guess, I would sort of sum that up into your data is probably one of the most valuable things you’re generating as a business, and it’s got a lot of gold locked in there, and you need to find ways to extract that because it’s unique and it’s a competitive differentiation.

Amanda Razani: Harnessing that data is definitely key for most businesses. I want to thank you for coming on our show today and giving your insights.

Jonathon Reilly: Yeah, thanks for having me. Happy to help and always happy to answer questions.

Amanda Razani: Thank you.