An AI architect advises ignoring AI developments that don't directly impact your current work or sphere of influence. This is a practical filter to manage the overwhelming noise and stay productive, a strategy born from a non-traditional background without preconceived industry biases.
Empowering non-technical employees to build AI solutions forces the documentation of tacit knowledge. By aligning an AI to perform a task, they convert expert intuition into a repeatable, documented process, thus mitigating key-person risk for the organization.
The biggest value leap isn't turning non-users (L0) into users (L1), but empowering users to become 'non-technical builders' (L2). These individuals create durable, company-specific AI solutions, which provides far more leverage and real business impact than simple adoption.
For optimal team performance, there should only be one 'non-technical builder' (L2) per team. Adding more does not increase output and often creates a 'too many chefs in the kitchen' dynamic. This L2 should be supported by a technical expert (L3) for scalability and governance.
Given current locked-down, imperfect enterprise AI, a key differentiator is 'creative solutioning.' This means persistently pushing against tool limitations and governance to find transformative workarounds, rather than giving up when a tool says something isn't possible.
Contrary to 'adopt or be fired' mandates, using threats to drive AI adoption is counterproductive. This 'threat framing' slows down learning and prevents genuine engagement, turning potential advocates into resentful compliers who will not innovate with the technology.
True ROI of AI isn't found in usage metrics like token counts. It's measured by identifying entire, expensive projects (e.g., a $4M manual document conversion) and using AI to make the problem 'vanish,' completing the work in hours instead of months.
Employees who resist AI because 'it's not good enough' provide invaluable product feedback. They highlight real gaps in quality, tools, or training. Their insights are more useful than those from 'performative users' who merely go through the motions of adoption.
The best people to build internal AI tools aren't the most technically skilled but those with deep 'company DNA.' They understand the work in an 'uncanny way' and can align AI to replicate nuanced, high-quality outcomes, often outperforming distractible 'AI excited' enthusiasts.
Don't get paralyzed by speculating on AI's future impact on jobs. This grand transformation is like a river that will find its own course. Instead, focus energy on where you can make a measurable, tangible impact right now and let the future unfold.