In a recent survey of senior leaders at large U.S. and U.K. professional services firms, 61% said they had abandoned at least one AI project in the past year because their people lacked the skills to deliver it. Deloitte’s “2026 State of AI in the Enterprise” report, based on a survey of more than 3,200 business and IT leaders across 24 countries, found that insufficient worker skills are now the single “biggest barrier to integrating AI into the business.”
There is no quick or easy solution to this problem. While it is possible to bring in new hires or contractors with the short-term capabilities you need, this approach is not sustainable in the long term as it is both expensive and creates critical dependencies. And it is equally impossible to flip a switch to develop these capabilities in-house overnight. But businesses can start the vital process of building those skills systematically. And there is no better time to begin than now. Organizations that get ahead of the pack in this critical area will build an advantage over their peers that will compound every quarter.
The Capability Stack
Organizational AI capabilities emerge from four mutually reinforcing layers of expertise.
{“blockType”:”mv-promo-block”,”data”:{“imageDesktopUrl”:”https:\/\/images.fastcompany.com\/image\/upload\/f_webp,q_auto,c_fit\/wp-cms-2\/2025\/10\/creator-faisalhoque.png”,”imageMobileUrl”:”https:\/\/images.fastcompany.com\/image\/upload\/f_webp,q_auto,c_fit\/wp-cms-2\/2025\/10\/faisal-hoque.png”,”eyebrow”:””,”headline”:”Ready to thrive at the intersection of business, technology, and humanity? “,”dek”:”Faisal Hoque’s books, podcast, and his companies give leaders the frameworks and platforms to align purpose, people, process, and tech—turning disruption into meaningful, lasting progress.”,”subhed”:””,”description”:””,”ctaText”:”Learn More”,”ctaUrl”:”https:\/\/faisalhoque.com”,”theme”:{“bg”:”#02263c”,”text”:”#ffffff”,”eyebrow”:”#9aa2aa”,”subhed”:”#ffffff”,”buttonBg”:”#ffffff”,”buttonHoverBg”:”#3b3f46″,”buttonText”:”#000000″},”imageDesktopId”:91420512,”imageMobileId”:91420514,”shareable”:false,”slug”:””,”wpCssClasses”:””}}
Technical depth. This is the specialized engineering capability that builds and maintains AI systems: machine learning engineering, data engineering, AI security, model evaluation, and related disciplines. Without sufficient technical depth, the wrong things get built and bought, and the organization creates risk that it doesn’t understand.
Domain application. This layer is where AI strategy meets business reality. It consists of the capability to apply AI within a specific business function. And it relies on people who understand not just what the technology can do, but where it creates value in a particular operational context.
General workforce fluency. This is the baseline capability that every knowledge worker needs: sufficient understanding to use AI tools productively, to recognize when outputs are unreliable, and to contribute usefully to conversations about how AI is being deployed in their area. Without this general fluency, adoption stalls, misuse spreads, and employees remain dependent on a small group of specialists.
Organizational infrastructure for learning. This is the layer that sustains the other three: the systems, incentives, and management behaviors that determine whether capability grows or erodes. It includes how learning is funded, how time for development is protected, how reskilling pathways connect to real roles, and how managers are held accountable for the capability development of their teams. Without this layer, every investment in the first three decays.
The 90-day plan that follows works through all four layers simultaneously.
The 90-Day Plan
Days 1-30: Map
The goal of this phase is to understand what you have, what you need, and where the gap between them will hurt you first.
1. Define the capability model. Use the capability stack to define what AI capability means for your organization. Be specific. What does technical depth mean in your business? Which roles require domain application? What level of AI fluency should every knowledge worker have? The shared model needs to be explicit and agreed on.
2. Identify the workforce baseline. Assess existing employees against the capability model. Use a combination of self-assessment, manager assessment, and skill validation—and treat all three with appropriate skepticism. None of these tools is perfect, but that’s okay: the goal is not a perfect picture, just a better one.
3. Map capability demand to the strategy. Take your AI strategy and the innovation portfolio it has produced, and decompose them into the specific capabilities required at each layer of the stack. This is the demand side of the equation, and it is typically missing from AI strategies altogether. Organizations approve ambitious AI portfolios and then discover, months later, that they don’t have the people to staff them. The demand map prevents that discovery from arriving as a surprise.
4. Identify the highest-leverage gaps. The gap between current state and required state will normally be large. You will not close it completely in a quarter, and attempting to do so will dilute the impact of investment across the board. Prioritize ruthlessly. Identify the handful of capability gaps that will most directly constrain the AI initiatives already in flight or about to launch. If your innovation pipeline has three experiments ready to go and two of them require data engineering capabilities that you don’t have, then that’s where the first thirty days of investment should be directed.
5. Audit how learning currently works. Map the current state of organizational learning. The infrastructure layer of the capability stack depends on it. Flag the parts of the system that will scale into the AI era and the parts that need to be rebuilt or replaced.
For a practical guide to building the AI innovation portfolio against which capability requirements should be mapped, see “How to build an AI innovation pipeline that creates real long-term value.”
Days 31-60: Build
In this phase, the organization begins closing the gaps previously identified while also laying the foundations for ongoing and systematic workforce development.
1. Launch the core technical hiring push. For the small number of roles that the organization genuinely cannot develop internally on the required timeline, run a focused external hiring effort. Be disciplined about which roles you select. Reserve external hiring for the positions where internal technical expertise of the required depth truly cannot be developed in the available window. For everything else, build from within.
2. Stand up the reskilling program. For the much larger population of employees who can move into AI-adjacent roles with the right investment, build a structured reskilling program tied directly to the capability model. The program should connect to real roles on the other side. Reskilling efforts fail when they become training programs with no path to a new job.
3. Drive baseline fluency across the workforce. Roll out a broad AI fluency program for the general knowledge-worker population. Tie completion to specific behavioral expectations, not just attendance.
4. Build the partner ecosystem. Identify the external partners—universities, training providers, specialist consultancies, managed service providers—that can accelerate the building of capabilities where internal investment alone cannot move fast enough. Partnerships should be structured with clear deliverables and explicit transfer-of-capability expectations. A partner that builds your capability is an investment, while a partner that performs the work without transferring the capability is a dependency-in-waiting.
5. Redesign the highest-leverage roles. Select two or three of the roles that will be most comprehensively transformed by AI in your organization. Redesign them deliberately, working with the people who do that job today. Ask practical questions. What parts of the job should AI take on? What parts should the human retain and do better? What new responsibilities emerge when routine work is automated? The redesigned role can serve as a template for the broader workforce transformation and as a concrete demonstration that capability development leads somewhere real.
6. Make managers accountable for capability development. Your middle managers are the transmission mechanism for every capability program you launch—if their teams aren’t developing, the programs aren’t working. So make your managers accountable for success. Success needs to be specific and measurable: employees reskilled into new roles, team fluency levels achieved against the capability model, learning time protected against competing demands, and internal moves into AI-critical positions. Managers who consistently develop their teams’ capabilities should be recognized and rewarded. The signal this sends through the organization is more powerful than any training program.
For more on why AI reskilling demands organizational transformation rather than individual training, see “What AI reskilling really requires.”
Days 61-90: Embed
Now it’s time to lock the changes into the operating fabric of the organization so that building workforce capabilities specific to AI becomes a permanent discipline rather than a one-off initiative that fades when the next priority arrives.
1. Operationalize capability reviews. Make capability a recurring item in talent reviews, business reviews, and board reporting. Build a capability dashboard, updated on a defined cadence, that tracks the state of each layer of the capability stack against the demand map from Phase 1. This turns a set of programs into a managed discipline, with the same rigor as that applied to financial performance or operational metrics.
2. Make learning a standing expectation. The test of whether an organization is serious about capability development is what happens when learning time collides with operational demand. In most organizations, learning loses. The fix is structural: Define the learning time expectation, make it visible, and hold managers accountable when it isn’t protected.
3. Track the flow of capability, not just the snapshot. If you only measure the stock of capability, you will miss the trends that determine whether you’re building momentum or losing ground. Track the indicators that reveal direction: internal moves into AI-critical roles, retention in those roles, reskilling throughput and placement rates, external hires converted to productive contributors, and the rate at which fluency programs change actual behavior rather than just accumulating completions.
4. Stress-test the capability with real work. Deploy the newly developed capability on an active AI initiative from your innovation pipeline and watch what happens. Where the capability holds under operational pressure, scale the playbook that produced it. Where it breaks—where the reskilled engineer can’t handle production complexity, where the fluent marketer still can’t evaluate model outputs—fix the upstream investment before you scale it.
5. Treat AI-critical roles as organizational infrastructure. Every AI-critical role in your organization is, to some degree, a new role—one that didn’t exist five years ago and may not have an established internal talent pipeline. That means every such role is a potential single point of failure. If your lead ML engineer leaves and there’s no one behind them, you don’t just have a vacancy—you have a capability collapse that can stall an entire portfolio of initiatives. Build succession depth for these roles the way you would for any other critical piece of infrastructure: Identify the successors, invest in their development, and make the pipeline visible.
6. Iterate. By day 90, the data is available. Which hires worked? Which reskilling pathways produced employees ready to do the job? Which fluency programs changed behavior rather than just generating completion certificates? Use the evidence. Reshape the next cycle based on what you’ve learned.
For a deeper look at how AI is redefining the management roles on which capability development depends, see “AI and the death (and rebirth) of middle management.”
Conclusion
This 90-day plan will not solve every capability problem. But what it will do is get you started on building the system that keeps capability growing long after the initial push. And this is more important than ever, because in the AI era, the workforce you have today is never the workforce you will need tomorrow.
{“blockType”:”mv-promo-block”,”data”:{“imageDesktopUrl”:”https:\/\/images.fastcompany.com\/image\/upload\/f_webp,q_auto,c_fit\/wp-cms-2\/2025\/10\/creator-faisalhoque.png”,”imageMobileUrl”:”https:\/\/images.fastcompany.com\/image\/upload\/f_webp,q_auto,c_fit\/wp-cms-2\/2025\/10\/faisal-hoque.png”,”eyebrow”:””,”headline”:”Ready to thrive at the intersection of business, technology, and humanity? “,”dek”:”Faisal Hoque’s books, podcast, and his companies give leaders the frameworks and platforms to align purpose, people, process, and tech—turning disruption into meaningful, lasting progress.”,”subhed”:””,”description”:””,”ctaText”:”Learn More”,”ctaUrl”:”https:\/\/faisalhoque.com”,”theme”:{“bg”:”#02263c”,”text”:”#ffffff”,”eyebrow”:”#9aa2aa”,”subhed”:”#ffffff”,”buttonBg”:”#ffffff”,”buttonHoverBg”:”#3b3f46″,”buttonText”:”#000000″},”imageDesktopId”:91420512,”imageMobileId”:91420514,”shareable”:false,”slug”:””,”wpCssClasses”:””}}
