A common refrain I hear from organizations beginning their AI transformation is this: tell me what to do.
Of course, I do tell them what to do, with detailed steps and documents—set up their AI governance structure, assess and select an AI use case, pilot their first AI project. However, for some organizations, my answers are unsatisfactory.
This is because what they really mean is this: Tell me exactly what AI tools and use cases we should use to get the maximum benefit for the least amount of work and financial investment.
I’ve heard that a number of times, and so have all the consultants I’m connected with, but it’s the wrong question to ask, because without a rigorous understanding of their business, industry, challenges, budget, data streams, team size, level of proficiency, and appetite for risk, it’s impossible to prescribe a solution that will deliver any real value.
Regardless, even if I bypassed understanding all of that and did so anyway, my solution wouldn’t be as good or useful as the solution they would discover in their own experimentation.
To get the benefit, you need to do the work.
Wharton professor and AI expert Ethan Mollick has said this too. In no uncertain terms, he believes that if organizations desire gains that will truly cement their competitive advantage, they need R&D—and they’re largely going to have to do that R&D themselves.
“For decades,” he writes, “companies have outsourced their organizational innovation to consultants or enterprise software vendors who develop generalized approaches based on what they see across many organizations. That won’t work here, at least for a while. Nobody has special information about how to best use AI at your company, or a playbook for how to integrate it into your organization.”
He goes on to say that even the major AI companies release models without knowing how they can be best used, discovering use cases after the fact.
This need for internal R&D and experimentation perfectly exemplifies what it means to have an AI First mindset. As I wrote recently, AI First is fundamentally about curiosity, proactivity, and experimentation. It’s about embracing uncertainty while maintaining the conviction that there’s value to be uncovered.
Organizations that adopt this mindset understand that the path to AI transformation isn’t about waiting for perfect solutions to emerge from consultants or vendors – it’s about actively engaging with the technology, testing hypotheses, and learning from both successes and failures.
The simple reality for organizations who want AI transformation is that you have to do the work to figure out the work you have to do.
Target use cases that have business impact and are feasible to implement. Understand the value you’re trying to create and know what metrics you’re going to track. And perhaps most importantly, remember that “much of gen AI’s near-term value is closely tied to its ability to help people do their current jobs better.”
The path forward requires following a structured approach:
Diagnose: Begin by thoroughly understanding your organization’s current state. What are your pain points? Where are your teams spending the most time on repetitive tasks? What data do you have available? This diagnostic phase helps identify promising areas for AI implementation.
Prescribe: Based on your diagnosis, develop hypotheses about where AI could create the most value. Create a prioritized list of potential use cases, evaluating each based on both potential impact and feasibility. Remember to start small—pilot projects should be meaningful enough to matter but contained enough to manage.
Apply: Launch your experiments with clear success metrics in mind. Monitor results closely, document learnings, and be prepared to iterate. Success in one area often illuminates opportunities in others, creating a flywheel effect of continuous improvement.
The organizations that will thrive in the AI era aren’t those waiting for perfect solutions— they’re the ones actively engaging in this cycle of diagnosis, prescription, and application. Talk to your people. Brainstorm some use cases. Pilot some projects.
Even if they fail, you’ll be much closer to finding the value you’re searching for.