ISHN recently interviewed Clyde Calhoun, founder and CEO of Root Idea, LLC. Mr. Calhoun works closely with C-suite executives to empower them with innovative AI strategies that maximize talent.
 

Executives are walking a fine line with AI adoption. Which is the bigger risk: adopting AI too fast or too slow?

You don’t want to move too slow, but moving too fast is common. I see so many organizations that have rolled out platforms and done it because they want to say they rolled out artificial intelligence.  

These organizations haven’t thought about how AI transforms an organization. You don’t want to be doing only random AI things and pilots.

This AI is moving so rapidly you can’t wait.

I was at a meeting of 250 nonprofit leaders discussing AI and the general feeling is most organizations are playing catching up with AI.

This is because the pace of AI is advancing so fast. So, executives say, “Let’s let the dust settle. Once AI is figured out and things have calmed down, then we’ll implement.”

Also, there are so many AI apps, 100s of apps, organizations say, ”We’ll figure it out later.”

There are so many organizations that treat AI like just another tech deployment, but this is very different. AI is a transformation in how people are working. But many people misunderstand AI or don’t’ estimate how hard it is to roll out.

AI success depends on clarity, readiness and governance. Define clarity and governance.

Clarity starts with strategy. How is it that AI will be impactful to your organization? Without clarity it is hard to articulate to your organization how AI will make an impact.

Be clear with your folks in terms of what they can do with AI apps and what they can’t do.

Organizations have invited employees to test out apps in a safe environment. At Honeywell, for instance, employees have the opportunity to make recommendations for new apps that will help them in their work. And Honeywell also has a team that makes sure that the tests are safe.

Leaders have to make sure it is clear to employees that AI is welcome. Employees can have reservations. They generate something AI wrote. But they may not know if that is good thing or bad thing with leaders. If you don’t communicate an acceptance of AI, you’ll get negative stories about AI from employees. You have to say, “We welcome playing around with this technology; we welcome innovation.” This is the governance part.

What are the most common risks that derail AI initiatives?

The number one risk is data -- data security and privacy and data integration across organizations. A lot of AI goes off track here. If data is not accessible to an AI tool, you’re limited in the value you derive from AI. Organizations have siloed data bins, which limits AI. A lot of organizations don’t see a ROI on AI. They thought it could be powerful, but you must feed AI with quality data.

Employee resistance is another risk to derailing AI. About 33% of employees resist AI. This is a real phenomenon that goes to the challenge of rolling out AI as a real change management issue.

Large change initiatives start with strategy. You need an AI strategy. Have a strategy and you’re twice as likely to hit a ROI.

You need consistent clear messaging to chip away at resistance. Also, you should have AI champions advocating in your organization. And you must have resources to support deployment. Strategy, communication and champions are the three keys.

Why is the focus of AI moving from generative AI to agentic AI?

The shift taps into a greater AI capability. With generative platforms like ChatGPT, you ask a question and get an answer. Agentic AI can figure out things on its own. You don’t have to ask a question; agentic AI will take next step on its own. These agents can talk to each other. This is powerful. Agentic AI acts more autonomously.

There is a general concern about AI acting autonomously; concern about hallucinations and mistakes. In the long term there is concern about losing control of AI agentic capabilities. Right now, organizations should have AI execute on non-critical areas. It is prudent to have human oversight on the most critical areas.

Using manufacturing as an example, AI can get input from processes and AI can control manufacturing processes. AI can make some great decisions and make efficiencies, but you still need an operator to monitor just in case.

It’s like self-driving cars – you still need a person at the driving wheel.

Why is it difficult to prove economic value or get a ROI from AI?

I don’t think it’s hard to prove if you focus AI on the right things.  You can’t measure saving time on emails, but if you use AI for sales analytics, you can see that bump in revenue.

The same goes with applied AI in operations. You see a reduction in downtime or increase in operational efficiency.

Too often organizations think of AI separately than what they focus on day in day out. You should focus on what are the pain points in my organization, what are my goals?  Is there some way AI can help? Define an objective. Maybe it’s to increase revenue or improve market share. Can AI help meet these goals?  AI can take you different places depending on where you start asking questions. One question: Is it making a difference for the organization?

Why are there few (33%) AI-driven cultures in organizations?

It’s people still not seeing AI technology as a part of who they are. Owens Corning for many years saw itself as a glass manufacturer. You can’t be just a product company anymore. You must be a technology-enabled company with a culture that reflects that.

I don’t see enough organizations training people on this AI technology and having the expectations that AI is part of employees’ knowledge base. We do have expectations for leadership skills, for example, but we don’t have expectations for employees learning these very powerful AI capacities.

Why do most chief data officers and chief AI officers say their functions are not well understood?

Some tension still exists at the executive level. You already have chief technology officers or CIOs. You need clarity here. If AI is truly embedded in the business, what is role of the chief AI officer versus the president or COO? There is a lot of tension in c-suite now.

How do you resolve this? AI has to be part of the business. It can’t sit outside the organization in a silo. Leaders will have to integrate AI at all levels of the organization. It’s like when employees had to learn PC skills.  You need the capacity for AI to be absorbed across the organization.

There are different schools of thought about how long this full-scale AI absorption will take. Some say it will happen in next three years; others say ten years. I think it will take five years to get this embedded acceptance. Go back and look at how long it took to get acceptance of the Internet -- that was rapid. Acceptance of cellphones kind of blew through very fast.

Should EHS leaders be concerned about the shortage of AI-skilled talent? About the struggle their organizations have in finding AI-skilled professionals?

From the EHS perspective, pros need to have an AI strategy for their function. Too many EHS people are waiting for IT to figure out what AI platforms are to be used. EHS pros should say, “This is what we need, this is what we want.”

You’ll see all EHS pros building up their AI skillset. It will be a natural competency. In near term many organizations don’t have the necessary resources and knowledge to achieve that, so you go to outside consultants and experts.

My encouragement to everyone is there are great resources online. It is easy to get going with your AI education. Ask AI to describe itself; ask the best ways to use AI as an EHS pro; tell me more about that, give me more examples. This is a great way to learn on your own.