Practical Considerations for AI Adoption in Business

A few years ago, I took an AI management course at MIT and have watched the technology evolve through the hype cycle of AI benefits and risks since then.

Clearly, AI is ready for its close-up in the business world, and we should all anticipate more AI adoption in the workplace. The challenge is balancing human expertise with AI technology.

Here's the academic view: A recent MIT Sloan Management Review study emphasizes the importance of developing a strong AI strategy that focuses on creating value for both individuals and the organization as a whole. This involves identifying the unique strengths of humans and machines and designing workflows that leverage these strengths to create value.

In practice for businesses, process automation and natural language processing have proven to be valuable tools with broad potential in enterprise IT systems. Let's call this Big Case AI. Yet from a practical standpoint, Big Case AI also comes with significant risks and ROI considerations when proposed in a larger business environment. All of that must be addressed by decision-makers and investors. It's human nature for this uncertainty to delay large-scale adoption as the angles are worked out. To move quickly on a Big Case AI project, you'd better have a clear route to competitive advantage in mind. A silver lining for second-movers is that the hard work of others in this space provides an opportunity to watch and learn from early adopters.

But there is a Small Case for AI adoption that may emerge more quickly. Popular AI tools, such as ChatGPT, can be used by individual contributors to streamline and accelerate workflows in functions like marketing, legal, communications, and digital design. With low barriers to entry, individuals can experiment with these tools and quickly identify any risks or issues, allowing for the development of workarounds and personalized solutions.

AI-powered tools are becoming more prevalent and useful across a wide range of industries. For writers, Grammarly offers real-time suggestions to improve grammar, punctuation, and style. Lawyers and legal professionals can use natural language processing tools to streamline legal research and document review by quickly identifying relevant information. TensorFlow is an open-source AI library that can be used for image and speech recognition, natural language processing, and predictive modeling. Marketers can automate various tasks, including lead generation and social media management, with AI-powered HubSpot, while Canva uses AI to suggest design elements and templates for creating various types of visual content.

For everyday workers, these Small Case tools can simplify and accelerate existing workflows. As individuals experiment with AI tools, they'll learn for themselves what works and what doesn't, and make adjustments along the way. This fits with what AI thought leaders, like my former MIT professors, would call "creating a culture of learning and experimentation," a phrase that probably sounds horribly unfocused and expensive to many business leaders in industry.

What does that really look like?

In real life, as we're learning with ChatGPT and other natural language processing approaches, it's simpler than the conversation has made it. These AI tools benefit from specific guidance on which frameworks and toolkits to employ, what role the user wants the AI to play, and what outcome the user desires. The AI can then act as an efficient "leg man," to borrow a journalism phrase -- an assigned researcher who can gather specifically what the user wants.

What the AI can't be -- not yet -- is an expert on what the information means. In the academic "partnership between human and AI," expertise is the responsibility of the human. The AI becomes valuable again as a polisher of the collaborative effort in the end, as long as the final review is done by an expert human. Keep notes on what you've asked, and which sources you direct the AI to use. This will help with tranparency during your earlier experimentation.

While the adoption of AI technology can still seem daunting on a Big Case, enterprise scale, taking a Small Case approach can provide immediate benefits and a roadmap for future adoption.

NOTE: I originally thought my ChatGPT AI assistant would have an easy time writing this topic up. If there's one topic it should know, it's AI! But the output was so full of misleading mischaracterizations, I scrapped it and started over. Instead, I used ChatGPT to summarize several large studies (and fact-checked the results), had it find useful AI tools (and eliminated the suggestions that weren't right), and wrote the post by hand. Then I asked which SEO terminology and search engine queries might lead to it most efficiently. I had the AI assistant blend that language into the post (and then redid it when it made my writing sound too mechanical). In this case, I was the expert. AI was the assistant.

Have a question for GIS Group? Email us at strategy@growthinnovationstrategy.com.