What’s the Time and Place for AI?

Up-and-coming AI tools can be used to improve efficiency and effectiveness as long as you use them in a targeted, forward-thinking way.

Sandy Smith

In just a little over a year, artificial intelligence tools like ChatGPT and Bard have taken off, offering a shortcut to unstick writer’s block or analyze massive amounts of data. While generative AI tools like these have gotten attention, AI is also being deployed in business systems, using a company’s own data to offer insights, and in search engines to help refine searches.

AI tools bring plenty of questions, but they also offer opportunities. Some may approach AI as a futuristic dream in which the most mundane tasks are handed over to a computer. Others may genuinely fear AI, worried that they may be out of a job. No matter the outlook, smart companies are exploring the possibilities of AI, if cautiously, to provide deeper insights into all aspects of business operations.

LOW-HANGING FRUIT

Generative AI tools are used to create, or generate, new content, including the written word, images and videos. It’s a rapidly expanding technology and certainly one that has drawn a lot of attention, and for good reason, says Dauphin Ewart, CEO of The Bug Master, a company that has a staff of more than 80 and services central Texas. "A lot of the uses come from individuals using generative AI to do something they already do," he said. "It might be polishing an email, creating sales messaging or writing blog content."

His company also uses AI content as a "starting point" for job descriptions or a training outline. "It’s a shortcut for work that still needs an expert. The quality of the content is pretty variable and still needs somebody to make judgement calls."

Free AI tools make anything that is input available for others to use, meaning sensitive and specific information should be avoided. But many of these same tools offer paid enterprise versions, where the data remains owned by the user.

AJ Treleven, director of operations at Sprague Pest Solutions, said his company is exploring those enterprise tools. "You can use that as a knowledge base to generate further policies and procedures, or create Q&As. We put in our procedure to a certain service and can query against it."

Company-specific training documents also can be input with the AI tool, generating quizzes or additional lesson plans.

I do believe when you COMBINE THE FACTORS OF LARGE LANGUAGE MODELS AND AI with remote sensing devices like door sensors and light sensors, we'll be ABLE TO ASSESS THE CURRENT PROBLEM and very likely a cause-and-effect relationship.

-AJ Treleven, Director of Operations, Sprague Pest Solutions

A LAYER ABOVE

Many enterprise tools, whether a CRM like Salesforce or business system like Microsoft 365, are deploying AI to enhance their services.

"Some of these less visible ways might be more impactful," Ewart said. "Financial tools like Dext or Bill.com are able to predictably understand vendors and where they might be coded. I’m not saying they’re perfect, but they are closer than some of the buzzy, fun, shiny things."

Ewart sees opportunities to help managers in particular handle massive amounts of data for more targeted actions.

"Contact center solutions are interesting, but a lot of where I’m seeing the development is around quality, using AI to grade calls and notice trends. It’s big data analysis rather than replacing an agent," Ewart said. "In my mind, if you could have someone able to manage five people, this would allow you to manage 20."

While his company has used AI chatbots, "I personally think those are further than what we’d be happy with." Ewart considers the customer interaction as a "pretty sacred transaction. That’s the last place I’m going to have us screw up."

With any new tool, there are mistakes, particularly as the tool learns. "If you miscode a bunch of financial transactions, go back and fix it," Ewart said. "But if you screw up how you handle 100 leads, you did something pretty negative. The question is, where are you willing to suffer the learning curve as well? All of these are going to have learning curves."

The real opportunity may be in deploying AI to find connection points between a variety of disparate sources. "We’ve got millions of datapoints going back 20 years," Treleven said. Sprague focuses exclusively on the commercial market, with the largest part of its customer base working in the food industry. "We’re starting to try to apply that data to give us new insight and predictive analytics. We’re moving away from preventative to predictive."

That might mean using an AI tool to assess how humidity levels could dictate breeding rates of pests and rodents. "Using that publicly available data, along with internally collected datasets, we don’t have to do a preventative treatment on every building, but only on those that are most likely to be impacted," Treleven said.

The AI tool can account for variables like weather conditions. Take winter 2024, for instance. It was a "pretty normal El Nino winter," Treleven said. But a typical El Nino hasn’t happened in years. "How do I correlate what’s happened now to the last time? It can prepare us and help us focus and train. It can start to inform the conversations we have so that we can offer a value add to our customers."

For his part, Ewart believes there is more work to be done in this area. "Sensors might give you temperature, moisture, but sensing ant activity can be very hard. Figuring out how you generate sensors that are as cost effective as sending a person, I’m not convinced that it’s actually been solved yet. Just like with generative AI, the new thing wasn’t the algorithm, it was the training and it was done by ingesting an incredible amount of data. You need a data source. Then you have to have a really good way to use those data sources to create modules that you can train. I think we’re pretty far from that."

He does see opportunities in marketing because of the large available datasets. "You can take your customer data, your closing percentages, and map those back out to the world to find where in nearby markets are your most ideal customers. You can use that data to understand what the lead costs are on those, what search volume looks like and what it’s taking to generate leads in those spaces. I am sure there are people doing some version of that right now."

THE FUTURE

With the rapid pace of adoption and the speed at which AI tools are learning, some of these points may be outdated by the time the printed magazine hits a desk. But Treleven believes AI plays a key role as the pest management industry moves from integrated pest management to progressive pest management.

"I jokingly talk about it as pest clairvoyance," Treleven said. "I do believe when you combine the factors of large language models and AI with remote sensing devices like door sensors and light sensors, we’ll be able to assess the current problem and very likely a cause-and-effect relationship. All these data points would tell us the most likely place to have an issue—and we’ll know before it happens. That level of clarity or lack of surprise is better for our team and the customers. We can all deal with something that has been planned for. It’s less of mental gymnastics to get through."

When the computer does the thinking—or at least a good portion of it—mental gymnastics can be kept to a minimum and that thought time can be spent elsewhere. It’s clear that AI has a long way to go, but the speed at which machines are learning will require virtually every company to pay attention—and to explore the options.