Raw Content

AI Adoption: What’s Actually Working

Main Thesis

Natalia Quintero, Every’s head of consulting, interviewed over 100 companies about AI implementation. Her key finding: the problem isn’t technology—it’s clarity. Most organizations struggle not because AI tools don’t work, but because they lack clear goals and processes for using them.

The Core Challenge

“We have the tools. We have a few power users. We don’t know where to go from here,” is the recurring refrain Quintero hears. A widely cited MIT report claims 95% of generative AI pilots fail, but this reflects a clarity problem rather than a technology problem.

Three User Personas Blocking Adoption

Skeptics: Uncomfortable with new technology, initial enthusiasm fades when learning becomes necessary
The Overwhelmed: Have tools but lack bandwidth to experiment alongside existing work
Tool-Jumpers: Analyze 30 different platforms without mastering any—disguised analysis paralysis

Why AI Adoption Differs from Traditional Software

Unlike Asana, where one person’s organization benefits everyone automatically, AI tools require individual fluency. Your custom prompts and workflows don’t transfer to colleagues, creating isolated “lonely power users.”

The Success Story: The Recruiting Firm

Rather than company-wide rollouts, Quintero’s team trained 10 AI champions—a mix of junior employees and leaders with willingness to learn, permission to build, freedom to fail, and eagerness to share.

One recruiter created a GPT automating scheduling coordination across three calendars, saving each recruiter 2-10 hours per scheduling task. Critically, the tool’s peer origin made it accessible, sparking broader curiosity about automation.

What Separates Successful Companies

Leadership models usage personally: Leaders who use AI themselves understand both capabilities and limitations
Centralized testing: Dedicated teams evaluate tools, preventing organization-wide tool-jumping
Champion-based rollouts: Three to five people from different organizational levels build peer-to-peer solutions
Documentation culture: Financial institutions and engineering teams possess natural advantages—they already document processes

The Critical Insight

“You can only automate what you can clearly define.” The hard part isn’t using AI—it’s determining what you’re trying to achieve and how to accomplish it. Think of AI as training an intern: specificity about tone, process, success metrics, and pitfalls determines effectiveness.