Four Lessons Learned From The Evolution Of AI

Eric Vermillion is the CEO of Helpshift.

It seems like new AI technology is springing up overnight and making massive leaps forward faster than the rest of the world can keep up. Companies are scrambling to invest in AI solutions. AI enthusiasts are quick to proclaim that AI is going to change everything. But I believe a lot of this hype is premature or misdirected.

AI is not a superhuman force that’s going to magically transform the world. The truth is that AI is not “new.” Companies have been developing AI tools and chatbots for years. And just like the evolution of any new technology, the development of AI has brought some successes and failures along the way.

Let’s look at some lessons learned from the evolution of AI and consider how today’s business leaders can make smarter AI investments.

1. Ignore the hype and focus on the problem.

There has been a lot of hype recently about AI, and while this hype is justified based on the amazing advances we have seen, in the end, AI is a tool for us humans to use to solve problems. We have seen the hype cycle before. For example, back in 2011, IBM’s Watson AI technology was heavily hyped, defeated Jeopardy champion Ken Jennings, and was named “Person of the Year” at the Webby Awards. Articles were written about how soon Watson would be replacing human lawyers. After the hype died down, IBM’s AI technology continued to develop and has helped serve millions of customers by being folded into other tech solutions over the years.

The lesson is that it is not about the hype but the problems that can be solved. This next generation of AI is extremely powerful. It’s pre-trained with powerful language models, has extensive real-world knowledge and comes with a very low barrier to entry. Brands that ignore the hype and instead focus on their problems will quickly see where AI can assist in making our lives easier and better. This is especially true when it comes to AI-based customer experience (CX) and AI assistants. I expect that the current AI hype cycle will die down, but AI will be infused into the background, enhancing existing processes and solving problems.

2. Approach AI from a practical perspective.

It seems like every company right now is announcing new investments and initiatives for AI. But this often feels frantic and a little desperate, like people are just trying to tack on “AI” to their company, even if it doesn’t fit.

My company has been working with AI for several years, but we try to take a holistic approach, not just implement AI for AI’s sake. Before your company invests in a big-splash AI initiative, take a step back and ask, “How can this really help solve a customer problem?”

For example, in the customer service and CX industry, lots of companies are excited about using AI for customer service. But one thing we’re discovering is that some companies don’t want “too much” AI or automation in their CX solutions. They want their customer service to sound human and authentic, not uncanny or robotic. In many instances like these, AI is used best as a “sidekick” for human workers, not a replacement. AI solutions can be massively helpful for specific problems like segmenting customers, recognizing customer requests and supporting customers with self-service for routine inquiries. But we still need the human element.

3. Treat AI like you would treat a human co-worker.

Companies are experimenting with popular generic generative AI models in a number of ways to help their workforce. However, it is important to understand that these models, although very powerful, are trained to be generalists based on a general knowledge of language and a vast set of public information. They can be good at answering questions or providing general insight, but without specialized training, guidance or process rules, they will be limited in what they can do.

Just like a newly hired human employee comes with a lifetime of experience, in order to make them effective, organizations need to take the time to train and provide processes or guides to their human employees. In the same way, generative AI should be augmented with brand-specific training, rules and processes. With proper training and guidance, AI can become a specialist with real power to solve unique problems.

4. AI is not the answer to everything.

Lots of companies are taking the attitude that AI is the “next big thing.” But AI can’t do everything. Leverage AI where it makes sense, integrate AI into non-AI areas where it adds value, but don’t expect AI to be the entire focus of your tech or business operations.

It’s OK to have menus, it’s OK to have some manual features and it’s OK to have non-AI technology working alongside the AI technology. AI doesn’t belong in every workflow or customer conversation.

For example, AI can be used for automated customer service to help analyze a customer’s intent when starting a conversation. If a customer types into the chatbot, “I want my money back,” the AI can help categorize the request and start the refund process. However, if you add a simple menu in the chat UI offering the top customer service requests to start the conversation, such as “Request Refund,” often customers simply will choose from the menu. The menu may not be “AI,” but combining AI with non-AI elements (like a menu) can lead to extremely high intent classification.

I believe in the power of AI and automation to boost productivity and help people do more of what they do best. However, business leaders must think carefully and strategically about when and how to build AI solutions into their work processes and customer conversations. Don’t get so dazzled by AI’s bright future that you lose sight of why it’s useful and how to leverage it practically.

Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?