Dave Grow, CEO of Lucid Software.
getty
AI is delivering real productivity gains, but those gains are coming with a new layer of complexity. Among users of the technology, McKinsey found that 80% report it has improved their individual productivity, and half say AI helps them make better decisions. It’s no surprise AI is having profound effects on enterprise as more teams embrace it to automate repetitive work, surface information faster and help people move from idea to execution.
However, as organizations add AI capabilities at a rapid pace, workflows are becoming more fragmented. Every new tool creates another place to work, interface to learn and handoff between systems. This tool sprawl unintentionally introduces friction drift, where the friction that AI removes in one part of a workflow reappears elsewhere, like an endless game of Whac-A-Mole. Long-term enterprise AI adoption depends on making AI work seamlessly within the environments and workflows people use every day.
Bring AI to the work, not the reverse.
Employees constantly move between AI tools and the applications where actual work happens. The context AI needs to support that work is spread across environments, including project plans and maps, process documentation, technical and architectural diagrams, institutional knowledge and team discussions. Together, these artifacts reveal how work has taken shape, why decisions were made, how processes evolved and how different pieces of a project connect.
When AI operates outside these native work environments, it lacks the context needed to understand the work or take meaningful action. Unstructured institutional knowledge, such as process logic and historical decisions buried in workplace chats, remains inaccessible. Employees end up bridging that gap themselves, pulling information from different systems and filling in the missing pieces AI needs before it can be useful.
Having a disconnect between AI tools and necessary business context creates a familiar set of challenges: fragmented information, context switching and cognitive overhead. Recent research by my company, Lucid Software, found that nearly one in five AI users say “organizing information has become more challenging in AI-enabled environments.” The result is added friction in the very workflows AI is meant to improve.
Embedding AI into the environments where teams already work can reduce this friction drift. Imagine an AI agent that can safely work from the same project plans, process maps and decision-making criteria a team uses every day. Important information stays connected to the work rather than becoming another input that employees have to gather and provide.
Prioritize interoperability.
Today’s typical tech stack no longer includes only traditional software; it also includes large language models (LLMs), AI agents and custom AI workflows. As these environments grow, connecting them can add complexity and deepen tool sprawl.
Open standards like Model Context Protocol (MCP) are already changing how AI connects to the systems where work has historically happened, reducing the need for isolated, one-off connections across tools and databases. This gives agents access to the institutional knowledge captured in working artifacts, from architectural diagrams to process docs, helping them move beyond generating outputs and take meaningful action within existing workflows.
This shift can help close the gaps where friction tends to resurface. For example, an AI agent can use a process map to generate tasks in another tool, fulfilling all the steps required and following the team’s established processes and decisions. When information moves between systems seamlessly, employees no longer need to repeatedly explain the same context to different tools, reducing the overhead that slows work down.
Build for agents with guardrails.
As workplaces become environments where both humans and agents complete tasks, governance is paramount. Deloitte found that 61% of leaders “expect most AI agents used will be generally autonomous” with human oversight.
Autonomy without governance creates new risks. Leaders have largely focused on guiding employees on the proper use of AI tools, but AI agents need clear direction, too.
Think of an agent as a brilliant stranger entering your company with incredible capability but zero institutional context. Like any new team member, it needs guidance. That starts with understanding what an agent actually needs to access and the sensitivity and quality of that information. From there, organizations can define the proper permissions and human oversight to keep data secure, maintain accountability and keep agents operating within clear boundaries.
Let the future of work flow.
As AI becomes more embedded in how work gets done, reducing friction will depend on how well it connects to the environments, information and workflows teams already rely on. When AI is integrated into the flow of work, people and agents have the context they need to move work forward. That connection is what will allow organizations to move from isolated AI wins to sustained, organization-wide impact.
Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?
