During its insights transformation, PepsiCo learned that implementing new technology was a minor hurdle compared to changing employee behavior. The real challenge was shifting the team's mindset from simply 'doing research' to strategically 'maximizing the benefit' of the data.

Many industrial tech solutions fail because they are designed as standalone engineering fixes. True success requires embedding the technology into daily operations, like shift meetings and handovers, making it a time-saver for workers rather than an additional analytical burden to drive behavioral change.

According to MIT research, the vast majority of corporate AI pilots fail. This is not due to the technology itself, but a disconnect where executives perceive success while employees report zero actual use. The core reason is a failure to integrate the tools into existing, meaningful workflows.

The biggest resistance to adopting AI coding tools in large companies isn't security or technical limitations, but the challenge of teaching teams new workflows. Success requires not just providing the tool, but actively training people to change their daily habits to leverage it effectively.

Instead of forcing teams to adopt entirely new processes, Atlassian is integrating agentic capabilities into familiar tools like Jira. Allowing users to assign a standard work item to an AI agent minimizes disruption and friction, accelerating adoption by enhancing, rather than replacing, established workflows.

Xaira is building two parallel organizations: an AI product team and an R&D team. A key operational struggle is merging tech's rapid, months-long development cycles with biotech's methodical, decade-plus timelines. This cultural integration is a major hurdle for next-generation biopharma companies.

While the technical setup of a modern IT automation tool like Serval can take less than an hour, the real bottleneck is organizational. The majority of implementation time is spent on change management—getting stakeholders to agree on abandoning legacy processes and adopting new, more efficient workflows.

The most effective way to integrate AI is not through individual training but by empowering teams to redesign their own work processes. This team-level approach fosters agency and ensures AI is used to solve real, shared problems, which is more powerful than simply making individuals 'AI literate'.

A three-day AI sprint is effective for generating ideas and enthusiasm, but the real, harder work is the "marathon" that follows. Success requires a dedicated task force to prioritize projects and methodically integrate the new AI workflows into day-to-day operations—a crucial step where many corporate innovation efforts fail.