Embedding the human factor into AI agent adoption

When implementing technology, the difference between training employees and helping them adopt it is often misunderstood. Training is a short-term, structured educational effort to teach technical skills, while adoption is a long-term process where employees use technology to change how they work to create more value. This distinction is crucial in the growing field of agentic AI.

Currently, 17% of businesses experiment with this technology to automate tasks, but Gartner predicts this number will rise to 60% within the next two years. Organizations aim to automate complex workflows, improve efficiency, and increase productivity, but many struggle to generate a return on investment.

A significant value gap exists, with 40% of digital transformation expenditures underperforming expectations due to adoption challenges. Employees lose a full working day each week to friction, such as re-entering the same information across multiple applications, finding workarounds, and dealing with unclear instructions. Ofir Hatsor, WalkMe's Senior Vice President of Sales for Europe, the Middle East, and Africa, attributes much of this friction to the complex, regulated business environment.

Many companies find it difficult to change daily operations, even though it is necessary to create real business value.

Identifying the value gap is a key issue, as most executives believe their organizations use 35 applications when the actual number is 661 - a visibility gap of 1,789%. Employees often use unapproved AI tools, sharing their code with others, without the visibility and control of IT leaders. This lack of official approval and control creates risks related to data privacy and governance. Organizations are slowing their agentic AI adoption until they can address these risks.

Another challenge is the actual cost of implementing agentic AI versus its real-world benefits. Many assumptions are based on boosting productivity and cutting headcount, while the real benefits may be in enhancing work quality and augmenting human activity. Additionally, organizations often lack knowledge about the number of tokens consumed and their costs, which can become very expensive.

The rapid pace of change in AI, especially among less tech-savvy employees, adds to the complexity of achieving a real return on investment.