AI in the Hands of Business: Companies look to develop efficiencies, but there are cautions

Bernard Gerwel: American National uses a chatbot named Annie to assist customers.

Jarad Johnson finds AI useful for translating website content to other languages.

Gabriel Cassady is writing AI ethics guidelines for his marketing company as industry use increases.

• 
• 
• 

Most people encounter artificial intelligence multiple times a day without thinking twice: facial ID recognition on a smartphone, using Siri or Alexa to find out the weather forecast, being targeted by specific ads while scrolling through social media or finding the next best series to watch with recommendations from Netflix.

With the emergence of ChatGPT and other open-source language processing models, however, the power of producing AI-generated content has been put into the hands of everyone, not just those working in the tech sector.

McKinsey & Co., a global management consulting firm, has noted AI growing in use across business sectors. Some examples are writing code and documentation in IT, drafting and reviewing legal documents, and assisting with job interview questions. The company notes that it is not a “ready-set-go” technology, however, and that business leaders need to identify the pros and cons of how and where AI can be used.

AI language-processing models have become a valuable tool for businesses of any size. Because of its ability to automate customer support, content creation and other language-related tasks, some local and national businesses are adapting the AI to fit their customer’s needs.

“There’s a segment of the population who is getting more comfortable using these tools,” said Bernard Gerwel, senior vice president and chief innovation officer at American National Insurance Co., who has an optimistic vision of what AI can do for businesses.

At American National, for example, Gerwel said the firm with an office in Springfield and headquarters in Galveston, Texas, has been using a retrieval model chatbot, named Annie, to assist customers with insurance questions.

“If you can download a form, or get a link via a chatbot, that’s much better than waiting to call a call center or track down your agent,” Gerwel said. “It makes a better client experience.”

American National also has implemented AI in its insurance claims department, which speeds up the processing time exponentially. Gerwel said clients can take photos of their damage, whether from a fender bender or a hailstorm, and upload that directly to the claims service.

“The AI can estimate the cost of the damage, which helps us with the claims process,” he said. “From my perspective, the AI can do the busy work; it gives us the opportunity to provide a more personalized service.”

And while the company hasn’t received any specific feedback about their customers using AI, Gerwel said website use at American National has increased with overall customer satisfaction.

The insurance industry is just one of many utilizing AI with time efficiency in mind. Marketing agencies, such as Mostly Serious LLC and 2 Oddballs LLC, are using AI language processing models to handle busy work and focus more on specific client needs, which translates to cost savings as well.

“Identifying opportunities to leverage AI for our clients is a focus for Mostly Serious this year,” said Jarad Johnson, the website design and digital marketing company’s CEO. “Our primary use for AI in 2022 was to reduce cost and accelerate the timeline for translating website content to other languages.”

The efforts resulted in a custom website plug-in for clients capable of automatically translating content on sections and pages of the entire website with just a click of a button.

“The result saves our clients tens of thousands of dollars in manual content translation services,” Johnson said.

At 2 Oddballs, co-owner Gabriel Cassady said the team has been experimenting with AI to speed up response times for customers.

“We don’t have the person-power to be on call 24/7,” he said, adding that customers tend to expect almost immediate responses these days.

Implementing AI similar to ChatGPT, he said, helps meet client expectations of a quick response.

“We’re leveraging some of this technology, which is in its infancy, and we’re using natural language processing models to take in customer requests, questions and then parse that and help us understand needs, next steps, next actions,” Cassady said. “We’re still only using it internally to help us fill in knowledge gaps faster. Currently, ‘Oddbot 3000’ is nothing more than a highly advanced virtual assistant.”

He adds that the team is discussing ways a customer-facing Oddbot might work.

Businesses are anticipating higher demand for AI services as the technology continues to improve. According to a 2022 McKinsey Global Survey on AI, the use of AI has doubled since 2018, with natural-language text understanding climbing to the top of the list of AI capabilities implemented, behind robotic process automation and computer vision. Because of this, business owners using AI recommend taking the time to learn as much as possible about the technology. And with any new tech, they warn that there are obvious downsides that need to be considered.

“The systems are still not perfect. If you go to a lot of websites or call centers, you still feel like you’re talking to a machine,” Gerwel said.

And because the data from these retrieval models are input by humans, there are defined parameters. Gerwel tested American National’s virtual assistant, Annie, and asked about the cash value of his life insurance policy.

“It didn’t know how to answer it because we haven’t taught it how to do that yet,” he said.

Relying on human input is a large concern when using AI in business, according to Cassady.

“There’s still a lack of accountability, security and privacy,” Cassady said.

With AI and machine learning models being used to supplement systems such as job applicant tracking, Cassady says if the software is carelessly trained with human biases, it can be damaging to not only the individuals applying for jobs but also to the company and community. “Those are issues we take seriously, so we are taking it slow. I would caution anyone who is moving at breakneck speed, you want to be able to have accountability and talk to people and make sure they understand what they’re using.”

 Cassady said he is in the process of developing an ethics guideline for his company to follow as the use becomes more prevalent.

“We can train biases and error into AI,” he said, which is dangerous if people become so reliant on the information and forget to fact-check or question it.

In addition to biases and errors, a common concern among business owners and employees is the potential for job loss.

“There is certainly hesitation,” Johnson said. “Much like automation and robotics in other industries, the fear of AI replacing humans is an important and nuanced concern. We believe the current AI tools offer a great benefit to accelerate tasks so that our team can spend more time on strategic and creative solutions.”

Gerwel remains bullish on the tech use in his business, however.

“We’re in the early, early stages on what AI can do for us. I am in the camp that it’s going to make us more productive, and potentially could lose positions, but the greater likelihood is it will make more people have better work-life balance,” he said.