One of the biggest questions currently facing the tech industry is how quickly and extensively enterprises worldwide will adopt GenAI applications and services. My in-depth research report on the topic (see: The Intelligent Path Forward: GenAI in the Enterprise for more) suggests that high-level adoption is progressing at a fairly rapid pace.
However, hidden within the broader narrative of that research – and other studies I've reviewed – is the fact that the impact and value of generative AI for individual workers remain decidedly mixed. Yes, organizations are actively developing applications and processes that leverage the impressive capabilities of large language models, but completing these applications and deploying them enterprise-wide has proven to be a significant challenge for multiple reasons.
Key challenges in GenAI adoption and the training gap
First, many enterprises are discovering that gathering the necessary in-house data to train and fine-tune models – so they reflect the unique knowledge base of their organization – is far more complex and time-consuming than initially expected.
Second, even after data collection is complete, the rapid evolution of AI models and the growing range of available options make maintaining and updating GenAI applications a difficult, ongoing process.
Most importantly, however, individual employees are not receiving the training they need to effectively use these new applications and services. One of the most surprising and concerning findings from my GenAI study is that fewer than half of the 1,010 companies surveyed offer any form of training on generative AI. Only 45% of respondents said their organizations provide introductory GenAI courses, and just 40% offer application-specific training to employees.
In real-world terms, this means most employees are left to figure out on their own how to use and maximize the potential of GenAI-powered applications. That's a significant problem because, as we are beginning to see, GenAI is not just an incremental improvement to existing workflows – it is fundamentally reinventing how work gets done. Yet despite the power and capabilities of these tools, most employees have no idea how to leverage them effectively. To put it simply, none of us are naturally born prompt engineers.
The result? Employees who attempt to use GenAI tools without proper training often have an incomplete and underwhelming experience. Even worse, a larger group of employees never even tries – or simply doesn't know where to start (see my previous column, "The rise of on-device AI is reshaping the future of PCs and smartphones" for more).
Breaking old habits
Even when training is available, another major challenge is overcoming ingrained work habits. Employees who have spent years – or even decades – using traditional productivity suites like Microsoft Office and Google Workspace often struggle to adopt new workflows.
This is likely a key reason why many enterprises, after an initial rush to invest in GenAI extensions and services for select employees, have slowed these investments – another concerning trend uncovered in my study.
On average, survey respondents reported that only about one-third of their employees currently have access to GenAI tools like Microsoft Copilot, ChatGPT, or Google's Gemini. Furthermore, they expect this figure to increase by only 3% over the next 12 months, indicating a deceleration in adoption. Without clear and consistent productivity gains – enabled only by widespread training programs – many enterprises are struggling to justify further investment in GenAI.
Another part of the problem is that the user interfaces for GenAI-powered tools need to be reimagined. Current implementations – such as text-based prompting tools or sidebar integrations in office productivity software – often feel like early-stage designs awkwardly tacked onto existing applications. These interfaces do not integrate seamlessly with traditional tools and workflows, often requiring excessive copying and pasting to be useful.
The ideal method of interacting with GenAI-powered applications is still unclear, but voice-based UIs could play a significantly larger role. However, getting people comfortable with speaking to their PCs may be more challenging than it seems.
Additionally, the rapid development of AI agents introduces new user experience challenges. While AI agents have the potential to be incredibly powerful, creating, managing, and deploying them effectively is not a straightforward task. If designed intuitively, they could drive rapid adoption. However, given the current fragmented state of GenAI applications and tools, I am not optimistic about seeing major breakthroughs in the near term.
As potentially powerful as AI agents might be, figuring out the best ways to create, manage and invoke those agents is clearly not going to be an easy task
The path forward in the enterprise
Regardless of how user interfaces evolve, the only way GenAI will have a lasting impact on employee productivity is if enterprises make substantial investments in training. Organizations need to either develop or acquire comprehensive training programs and ensure employees actively participate.
Although it may not be immediately apparent, GenAI is set to transform the way many employees perform their daily tasks. However, realizing this transformation will require an unprecedented level of workforce education.
If companies truly want to drive widespread AI adoption, they must shift their focus toward training employees on how to effectively use these tools.
Currently, too little emphasis is being placed on this critical issue. Instead, most discussions remain fixated on the latest advancements in AI models and their performance metrics. If companies truly want to drive widespread AI adoption, they must shift their focus toward training employees on how to effectively use these tools.
We also need to see vendors start spending more of their development efforts on improving the ease of use and working on the intuitiveness of their offerings. Neither of these are easy tasks, but if we're ever going to move beyond the rush to improve the technology for technology's sake story that's currently dominating the world of GenAI, this work needs to start soon.
Bob O'Donnell is the founder and chief analyst of TECHnalysis Research, LLC a technology consulting firm that provides strategic consulting and market research services to the technology industry and professional financial community. You can follow him on Twitter @bobodtech