Greetings everyone and welcome to the fifth edition of Barefoot Bytes, our monthly newsletter that opines on all things software, marketing, and data science. In this edition I’m going to be writing about some AI use cases that you can apply to your business today. Every board room discussion right now seems to be about leveraging AI, and how to get a return on that investment. But many don’t know where to start. These are some real world use cases that you can start working on today.
The number one use case I’m seeing deployed in production environments today is chatbots. These are powered by large language models, but trained on materials very specific to the company. Customer support is an obvious place to start. Customers can receive answers instantly, and executives can drastically reduce their customer support costs. The CEO of Dukaan, an India-based ecommerce company, laid off 90% of his customer support staff after deploying a chatbot that took a data scientist only two days to build.
Some are nervous however about putting a potentially hallucinating bot in front of their customers. I’m seeing even more interest in internal chatbots. Training a bot with your employee handbook and other HR materials would save a tremendous amount of time for HR trying to figure out if Sally’s dental plan covers Invisalign.
I have no idea why the word “robotic” is included here – it’s just process automation – but this is what the industry has agreed to call it. Process automation has been around for a really long time. The difference now is that software can actually digest and understand documents. I remember the days, as a developer, when trying to parse a PDF file just to get a single value was an arduous task fraught with problems and took many hours. These days, you can upload a PDF into a vector database, ask questions about the content, and receive responses with very high accuracy. In most cases, it will be more accurate than a human understanding and transcribing data from one document to another. So think about some of the manual business processes that you have in place today. What previously seemed impossible to automate might now be very straightforward. If your business is pushing around a lot of Word docs, it’s time to rethink what’s possible.
The internet is currently being flooded with AI-generated marketing content. It’s just too easy. I always recommend a human-in-the-loop, but the speed at which we can now generate blog posts, emails, imagery, and other content has accelerated tremendously. If your marketing team isn’t using generative AI tools today, then you’re already behind.
For sales, most of the applications are focused on personalizing outbound lead generation efforts. Think about a model that ingests relevant information about leads and accounts and can generate wildly personalized emails or LinkedIn messages. I would say that apollo.io looks to be the leader here. Other players include warmly.ai, clay.com, and kalendar.ai.
While not as approachable as the three above, there are a tremendous amount of machine learning projects being considered today. Machine learning, put as simply as possible, is training a machine on complete historical data in order to predict incomplete future data. Like whether or not a person is likely to click on an ad. The question I’m asking CEO’s on this one – what would be valuable to predict? If you have enough historical data, which is often as low as in the thousands of records, machine learning can probably help with prediction.
The four above seem to be the most prevalent today. But I’m also seeing:
Product Development
For those building products, there are so many tools that can support that effort. Github, Canva, and Autodesk have all introduced AI tools to help with development workflows.
Predictive Maintenance
For businesses that own a lot of machines, predictive maintenance can be a huge cost saver. Replace parts before they break.
Price Optimization
This is really a specific use case for machine learning, but for ecommerce and brick and mortar, optimizing pricing based on a large number of factors beats the traditional rule-based pricing hands down.
To wrap this up, when looking to leverage AI at the company-level, start thinking about chatbots, manual processes that need an overhaul, and your sales and marketing operations. Ask yourself- what would be valuable to predict? And hit me up if you want to discuss these or other AI use cases for your business.
Hunter Jensen
Barefoot Solutions, CEO
4241 Jutland Dr., Suite 300
San Diego, CA 92117