Your organization has been "exploring AI" for eighteen months. You've sat through vendor demos. You've formed a task force. You've piloted a couple of tools. You've added "AI strategy" to your annual planning agenda.
Meanwhile, your competitor just announced they're operating with 30% fewer recruiters handling 40% more hiring volume through AI automation. A talent war just intensified because candidates now expect AI-enhanced application experiences. Your best engineers are leaving for companies where AI tools make them more productive. And the skills your workforce spent years developing are being commoditized by AI at a pace that's making your entire talent development strategy obsolete.
This isn't "let's thoughtfully plan our AI journey" territory anymore. This is "make critical decisions now or fall irretrievably behind" territory.
There are five decisions every CHRO must make in the next 90 days—not next quarter, not next planning cycle, not after you've done more research. Now. Because the window for proactive AI strategy is closing, and what replaces it is reactive crisis management from a position of competitive disadvantage.
Let's talk about what those decisions are, why they can't wait, and what happens if you keep deferring them.
Decision 1: Your AI Governance Model (Who Decides What, and How Fast)
The decision: How will your organization make decisions about AI adoption in HR—centralized control, distributed experimentation, or something in between?
Why it can't wait:
Right now, AI adoption in your organization is happening by default rather than design. Individual recruiters are using ChatGPT to write job descriptions. Managers are experimenting with AI interview tools. Your learning team is testing AI course recommendations. Each decision is being made in isolation, with no governance, no risk assessment, and no consistency.
This creates several ticking time bombs:
- Legal exposure: Someone is using an AI tool that hasn't been vetted for bias, discrimination, or compliance—and you don't even know it exists until the EEOC comes calling
- Data leakage: Employee data is being fed into AI systems with unknown privacy protections and data handling practices
- Capability fragmentation: You'll end up with seventeen different AI tools that don't integrate, creating data silos and vendor chaos
- Missed opportunities: Without coordination, you're not learning from experiments or scaling what works
The actual decision framework:
You need to decide on a spectrum:
Full centralization: All AI tools must be approved centrally before use. Slow but safe.
- Appropriate if: You're in highly regulated industry, have significant legal exposure, low risk tolerance
- Risk: You'll move so slowly that your organization adopts AI via shadow IT anyway
Controlled experimentation: Small-scale pilots allowed with light governance, scaled deployment requires approval.
- Appropriate if: You want to encourage innovation while managing risk
- Risk: Experiments proliferate without clear scale criteria, pilot fatigue sets in
Distributed with guardrails: Teams can adopt AI tools that meet specific criteria (pre-approved vendors, privacy standards, bias testing), escalate anything outside guardrails.
- Appropriate if: You need speed and want to empower teams
- Risk: Harder to maintain consistency, requires mature governance capability
What "making this decision" looks like:
By end of month one:
- Document your AI governance model in writing (one page, not a fifty-page policy)
- Assign clear decision rights (who can approve pilots, who can approve scale deployments, what requires CHRO/legal review)
- Establish decision speed targets (pilot approvals within one week, deployment decisions within 30 days maximum)
- Communicate the model broadly so people know how to move forward
What happens if you don't decide:
You'll get the worst of both worlds—slow enough that frustrated teams go around you, fast enough that ungoverned AI adoption creates legal and operational risks. Two years from now you'll be cleaning up a mess that was entirely preventable.
Decision 2: Your Skill Taxonomy and Job Architecture Approach (Rebuild or Patch)
The decision: Are you going to fundamentally rebuild how you categorize skills and define jobs for an AI era, or are you going to incrementally patch your existing framework?
Why it can't wait:
Your current skills taxonomy was probably built between 2018-2022. It categorizes capabilities like "Excel proficiency," "data analysis," "content writing," and "customer service skills" with proficiency levels from novice to expert.
AI has made those categories obsolete. When AI can generate expert-level content, analyze complex datasets, and handle customer inquiries, the skill isn't "how well can you do this task"—it's "how well can you leverage AI to do this task while applying human judgment."
Every day you operate with an outdated skills taxonomy, you're:
- Hiring for the wrong skills (recruiting "data analysts" when you actually need "people who can formulate good analytical questions and validate AI-generated analysis")
- Training for capabilities that are being commoditized (teaching Excel when you should be teaching AI-augmented analysis)
- Planning careers that won't exist (progression paths based on skill mastery that AI is making irrelevant)
- Assessing performance on wrong criteria (evaluating writing quality when you should evaluate strategic thinking and AI collaboration)
The actual decision framework:
Option A: Incremental patching
- Add "AI skills" as new category to existing taxonomy
- Update some job descriptions to mention AI tools
- Create AI training as add-on to existing programs
- Timeline: 3-6 months, lower cost, lower disruption
Option B: Fundamental rebuild
- Redesign skills taxonomy around human-AI collaboration capabilities
- Rebuild job architecture based on what humans uniquely contribute
- Reconstruct career paths for AI-augmented roles
- Timeline: 9-15 months, higher cost, significant disruption
The right answer for most organizations: Start with A, commit to B
You can't wait 15 months for a complete rebuild—your hiring, development, and planning can't be on hold that long. But incremental patching won't solve the fundamental problem.
What "making this decision" looks like:
By end of month one:
- Acknowledge that current taxonomy is inadequate for AI era
- Implement immediate patches for highest-priority roles (the 20% of jobs most AI-impacted)
- Commit to fundamental rebuild with timeline and resources
- Begin rebuild for one major job family as proof of concept
By end of quarter:
- Patched taxonomy operational for critical roles
- Rebuild pilot completed for one job family
- Full rebuild plan approved and resourced
What happens if you don't decide:
You'll keep recruiting, developing, and promoting people based on capabilities that are rapidly becoming irrelevant. Your talent strategies will be optimized for 2020, not 2026. Competitors who rebuild faster will attract talent you can't, deploy capabilities you don't have, and move at speeds you can't match.
Decision 3: Your Workforce AI Literacy Baseline (Universal Expectation or Optional Skill)
The decision: Is AI literacy a universal baseline expectation for all employees, or a specialized skill for specific roles?
Why it can't wait:
Right now, AI capability in your organization probably follows a power law distribution: 5% of employees are power users leveraging AI extensively, 20% are experimenting, 75% aren't engaging at all.
This distribution will determine your organization's competitiveness. If AI literacy remains concentrated in 25% of your workforce while competitors are at 80%, you're operating with a systematic capability disadvantage.
But building organization-wide AI literacy requires massive investment in training, time, and change management. If you don't decide this is mandatory and fund it accordingly, it won't happen.
The actual decision framework:
Option A: AI literacy as universal baseline
- Every employee expected to demonstrate basic AI collaboration capability
- Integrated into onboarding, performance expectations, and role requirements
- Significant training investment required
- Measured and managed like any core competency
Option B: AI literacy as role-specific requirement
- Certain roles require AI capability, others don't
- Targeted training for AI-intensive roles
- Lower investment, narrower impact
- Risk of two-tier workforce (AI-enabled and AI-excluded)
The right answer for knowledge work organizations: A (universal baseline)
You might think certain roles don't need AI. You're probably wrong. Customer service, HR operations, basic finance work, administrative support—roles you'd assume are AI-resistant actually benefit enormously from AI augmentation.
What "making this decision" looks like:
By end of month one:
- Declare AI literacy a universal expectation (or don't, but decide)
- If universal: Define what "AI literate" means for your organization (not "expert," but "can effectively use AI tools to enhance work")
- Allocate budget for universal training (this is expensive—plan for $500-1,500 per employee depending on approach)
By end of quarter:
- Pilot training program running for first cohort
- Assessment framework established (how you'll measure AI literacy)
- Integration into performance management defined (how this becomes expectation, not suggestion)
What happens if you don't decide:
AI capability will remain unevenly distributed. Your organization will have pockets of AI-enhanced productivity coexisting with large populations operating pre-AI methods. Competitors with universal AI literacy will systematically outperform you because their entire workforce is augmented, not just portions of it.
Decision 4: Your AI-Human Workforce Mix Strategy (Today's and Tomorrow's)
The decision: What percentage of work currently done by humans will be done by AI in 12 months? 24 months? And what's your plan to manage that transition?
Why it can't wait:
Most CHROs are approaching this reactively: "We'll see what AI can do and adjust workforce accordingly." This is backwards. You need to proactively decide your AI-human mix strategy, then execute toward it.
Without this decision, you'll get the worst of both worlds: slow enough that you're not capturing AI efficiency gains, fast enough that you're creating anxiety and uncertainty without clear direction.
The actual decision framework:
For each major job family, project:
Current state: X% human work, Y% AI-automatable (realistically, not theoretically)
12-month target: X% human, Y% AI, with humans redeployed to higher-value work
24-month target: X% human, Y% AI, with workforce reshaped accordingly
Example for customer service:
Current: 100% human agents handling all inquiries
12-month target:
- 65% of routine inquiries handled by AI
- Human agents focus on complex issues, relationship building, escalations
- Same headcount, radically different work mix
24-month target:
- 75% of inquiries AI-handled
- 30% reduction in customer service headcount through attrition
- Remaining agents are specialists in complex problem-solving
- Higher compensation for remaining roles (different skill requirements)
What "making this decision" looks like:
By end of month one:
- Map current human work across major job families
- Assess realistic AI-automation potential (not vendor promises, realistic capability)
- Set targets for 12-month and 24-month AI-human mix
By end of quarter:
- Detailed transition plans for highest-impact areas
- Workforce redeployment strategy (where do humans go when AI takes routine work?)
- Communication plan (how you'll message this to workforce)
- Budget implications (AI tool investment, workforce transition costs, productivity gains)
What happens if you don't decide:
You'll drift into an AI-human mix determined by vendor capabilities and individual manager preferences rather than strategic choice. You'll miss productivity gains because you didn't plan for workforce redeployment. You'll create anxiety because employees don't know what's being automated or what their future looks like. And you'll be making reactive workforce decisions based on what AI happened to automate rather than proactively designing for optimal outcomes.
Decision 5: Your AI Risk and Compliance Posture (Risk-Averse or Risk-Managed)
The decision: How much legal and ethical risk are you willing to accept to move quickly on AI, and how will you manage that risk?
Why it can't wait:
AI in HR creates real legal exposure: bias and discrimination risks, privacy violations, explainability requirements, regulatory compliance across multiple jurisdictions. The regulatory environment is evolving rapidly—NYC Local Law 144, EU AI Act, Illinois BIPA, and more.
Some CHROs are paralyzed by this risk and won't move until perfect compliance clarity exists. Others are charging forward and hoping legal catches up. Both approaches are wrong.
The actual decision framework:
You need to choose your risk posture on a spectrum:
Risk-averse: Move only when legal/compliance certainty exists
- Appropriate if: Highly regulated industry, low risk tolerance, significant discrimination litigation history
- Consequence: You'll be 12-24 months behind competitors, may miss window for competitive AI advantage
Risk-managed: Move quickly but with active risk mitigation
- Appropriate if: Moderate risk tolerance, ability to invest in compliance, competitive pressure to move fast
- Consequence: You'll have some exposure but you're managing it proactively
What "making this decision" looks like:
By end of month one:
- Explicitly decide your risk posture with legal counsel
- Document what level of legal uncertainty you're willing to operate in
- Establish "red lines" (things you absolutely won't do regardless of competitive pressure)
- Define risk mitigation requirements for AI adoption
By end of quarter:
- Active risk management process for AI tools (bias audits, privacy assessments, compliance reviews)
- Legal playbook for AI adoption (what approvals needed, what documentation required)
- Vendor risk requirements (what you demand from AI vendors to manage liability)
- Board-level reporting on AI risk exposure
What happens if you don't decide:
Your legal team will default to risk-averse, your business leaders will push for risk-aggressive, and you'll be caught in the middle making inconsistent decisions that create both missed opportunities (too slow) and unmanaged exposure (too fast in wrong areas). You need explicit decision about acceptable risk level, or you'll get chaos.
The 90-Day Timeline: What Happens When
This isn't theoretical. Here's the actual 90-day execution plan:
Days 1-30:
- Make all five decisions (governance model, taxonomy approach, literacy baseline, workforce mix strategy, risk posture)
- Socialize decisions with executive team, get buy-in
- Assign clear ownership for execution
- Communicate direction to HR organization
Days 31-60:
- Begin execution on highest-priority elements from each decision
- Launch pilot programs (governance process, skills taxonomy patch, AI literacy training, first AI-human mix transition)
- Establish measurement and tracking
- Address roadblocks and adjust
Days 61-90:
- Scale what's working from pilots
- Present progress to board (these aren't just HR decisions—they have strategic and financial implications)
- Refine based on learning
- Set next quarter priorities building on foundation
The Cost of Waiting
"We'll address this next quarter" is a comforting delay tactic. But the cost of that delay is real and mounting:
Every quarter you operate without AI governance, you're accumulating legal exposure and missed productivity.
Every quarter you operate with obsolete skills taxonomy, you're hiring wrong and developing ineffectively.
Every quarter your workforce remains AI-illiterate, competitors with AI-literate workforces pull further ahead.
Every quarter you avoid deciding your AI-human mix strategy, you're drifting rather than designing your workforce future.
Every quarter you don't clarify your risk posture, you're making inconsistent decisions that create both missed opportunities and unmanaged exposure.
The compounding effect of these delays is measured in competitive disadvantage you can't recover.
The Mandate
You became CHRO to build organizational capability and competitive advantage through people strategy. AI is the most significant workforce transformation in your career—probably in a generation.
You can lead this transformation proactively by making hard decisions now, or you can be dragged through it reactively by competitive pressure and crises.
Five decisions. Ninety days. No more task forces, no more "let's study it further," no more waiting for perfect clarity that won't come.
Decide. Execute. Adjust as you learn.
The mandate is clear. The timeline is now. The cost of delay is irreversible competitive disadvantage.
What's your move?
