Many companies aren’t giving their workers the training or AI tools they need to move beyond basic functionality.
Employees are widely using AI, but they’re sticking to basic tasks because CIOs and other managers aren’t giving them the training or advanced tools they need to take full advantage of the technology.
All US workers surveyed for AI collaboration tool vendor Arthur Technologies in July say they used AI at work during the previous three months, with nearly half saying they used AI multiple times a day.
But the survey results suggest that most employees still use AI as a “junior assistant” instead of a more powerful peer or teammate, with many workers turning to AI to help draft emails or conduct basic research, according to the company.
About 56% of those surveyed say they use AI to assist with research, and another 51% for data analysis. But only 38% say they use AI for project management, 11% for customer service, and 7% to assist with sales, according to the survey. Less than half of those surveyed see AI as a true collaborative partner.
Much of the AI use described in the survey seems to be shallow, with surface-level interactions with AI, instead of the technology integrated across business processes, says Christoph Fleischmann, founder and CEO at Arthur Technologies.
“If you look at the entirety of your workday, how you work and how decisions are made, it looks very much like all of your normal work has not changed,” he says. “What you do is, every now and then, bounce away and do something with AI.”
Training lags ambitions
Fleischmann points to several problems, including a lack of training provided by employers. While 82% of respondents say their employers encourage them to use AI, 37% of employees have not received formal AI training, he notes.
Many organizations also aren’t providing AI tools to employees beyond copilot-style AI assistants, Fleischmann and other AI experts say. While in some cases, AI tools are helping employees generate new ideas for the business, organizations aren’t implementing AI in ways that can bring these ideas to life, he says.
“Companies need to update the way internal processes work not as individual tools that create more silos and fragmentation, but actually a layer accompanying your team throughout the entire lifecycle of your project or weekly work,” Fleischmann says. “You should think of it less vertically and more horizontally.”
Fleischmann sees the problem as largely a management problem, with CIOs needing to take a leading role to provide more employee AI training and to evangelize the use of the technology as part of vital business workflows.
“A lot of teams I talk to completely lack even awareness where the problem is,” he says. “The job of the CIO is to say, ‘You need to upgrade the way you work, and we need to give you the infrastructure so that AI accompanies you through all of the different tasks you need to do as a team.’”
But he also sees some employee resistance, as workers worry about the future of their jobs if they allow AI to take over too much of their roles. If employees become more productive by using AI, it’s up to management to convince them they aren’t going to lose their jobs, Fleischmann says.
“I think you will have much more job augmentation than replacement,” he says. “If your employee was at X points of productivity before AI, and this same person has a lot more productivity points now, why would you get rid of them?”
Bolted on chatbot
Other AI experts also see a problem with employees focusing on basic AI tasks, with a lack of management support holding them back.
One problem seems to be a lack of functionality in the AI tools organizations are deploying, says Zahra Timsah, cofounder and CEO at AI agent provider i-GENTIC AI.
“We’ve seen this phenomenon with the rise of agentic AI, but it’s actually not the employees who don’t move forward with advanced uses,” she says. “Rather, it’s the businesses that are failing their employees because companies believe they’ve bought into an agentic AI product, only to learn that it’s little more than a glorified dashboard with a chatbot bolted on top.”
There seems to be a trend toward “AI comfort zone syndrome,” adds Brady Lewis, senior director of AI innovation at AI-focused fractional marketing consulting provider Marketri. In some cases, organizations don’t encourage employee use of AI, leaving workers to learn it on their own, he says.
Organizations that don’t encourage workers to use AI in a variety of ways will soon be at a huge disadvantage, Lewis suggests.
“Very quickly, they’re going to fall drastically behind the organizations that have it working well and allowing their employees to actually gain those efficiencies that proper AI usage allows,” he says. “It’s kind of like if you decided that you didn’t feel like giving your employees computers when everybody else is giving their employees computers.”
Lewis recommends that CIOs and other IT managers give employees safe experimentation sandboxes. While focused adoption can backfire, CIOs and other managers should give workers hands-on experiences.
“To work with AI effectively, employees need to adopt a thought partner mindset and build related skills, such as learning to iterate with AI, knowing when and what to delegate to AI, and properly prompting with context and intent,” Lewis says. “Give people time to play without unnecessary pressure, pair AI-savvy employees with beginners, and celebrate all of the little wins, not just big efficiency gains.”
Integrate AI across the business
CIOs should also look to integrate AI across business workflows, adds Micha Kiener, CTO at intelligent business automation platform provider Flowable. As organizations embrace AI agents, there’s the potential for AI to take over many time-consuming tasks that humans once had to do, he says.
For example, an AI-driven chatbot at a health insurance company could answer questions about whether a certain medication or medical procedure is covered by a customer’s plan. To go beyond that basic use of AI, the insurance company could use a series of agents to then process the client’s insurance claim, all the way up to payment of the medical bill, he suggests.
AI needs to be woven into workflows and scaled across departments, Kiener adds, because basic AI tasks may be useful, but they don’t move the business needle.
“AI is still being treated as a sidekick for surface-level tasks like notetaking, research, or brainstorming,” he adds. “The real barrier isn’t employees’ willingness; it’s that AI hasn’t been embedded into the systems that run the business. Until AI is orchestrated inside workflows like claims management, onboarding, or case resolution, it stays a novelty rather than a driver of enterprise value.”
