There's a version of this conversation we've been having for a year: "We need to be careful. We need to assess. We need to wait for the technology to mature." Meanwhile, enterprise AI investment just crossed $13.8 billion (a 6x increase from 2023) and the gap between AI-active and AI-inactive enterprises is becoming structural.
What You Need to Know
- The cost of AI inaction isn't just missed efficiency. It's compounding competitive disadvantage. Enterprises with AI foundations are accelerating; those without are falling further behind every quarter.
- Enterprise AI investment grew 6x from 2023 to 2024. That's capital allocation by organisations that have seen production results.
- The foundation advantage is real: enterprises that built shared AI infrastructure in 2023 are now deploying new capabilities in weeks instead of months. Latecomers still face the same 12-week foundation build.
- "Waiting for AI to mature" sounds prudent. In practice, it means waiting while competitors build data pipelines, knowledge bases, and governance frameworks that take 6-12 months to replicate.
- The minimum viable action isn't deploying AI. It's running a structured discovery to know where AI creates value in your specific business. This takes 4-6 weeks and $20-50K.
6×
increase in enterprise generative AI spending from 2023 to 2024
Source: IDC, Worldwide AI Spending Guide, 2024
92%
of companies plan to increase AI investments over the next three years
Source: McKinsey & Company, The State of AI in 2024, May 2024
The Compound Gap
Enterprise AI has a compounding dynamic that most leaders underestimate. It works like this:
Year 1: Enterprise A builds its first AI capability: claims intelligence. It takes 12 weeks because they're building the foundation simultaneously: document processing pipeline, knowledge base, integration framework, governance patterns.
Year 1, Month 6: Enterprise A deploys capability #2, fraud detection. It takes 8 weeks because it reuses 60% of the foundation. Total investment: 20 weeks for two capabilities.
Year 1, Month 9: Enterprise A deploys capability #3, customer communication AI. 5 weeks. 75% reuse. Total: 25 weeks for three capabilities.
Year 1, Month 12: Enterprise A deploys capability #4. 4 weeks. 85% reuse. Total: 29 weeks for four capabilities.
Now consider Enterprise B, which waited a year to "see how things develop." Enterprise B starts building in Year 2. They face the same 12-week foundation build that Enterprise A faced a year ago. By the time Enterprise B has one capability, Enterprise A has four, and is deploying the fifth in 3 weeks.
This is the compound gap. The delay isn't 12 months. It's a structural disadvantage that grows wider every quarter, because each new capability builds on the ones before it.
4×
faster delivery by the fourth AI capability when building on a shared foundation
Source: RIVER Group, enterprise engagement data, 2023-2024
The Five Costs of Waiting
1. Direct Competitive Disadvantage
Your competitors who are building AI capabilities are processing claims faster, serving customers better, making decisions sooner, and reducing costs. Every month you delay, the performance gap widens. In concentrated industries (like NZ insurance, professional services, and government), this is acutely visible.
2. Talent Drain
AI-capable organisations attract AI-capable people. The longer you wait, the harder it becomes to attract (or retain) team members who want to work with modern technology. This creates a reinforcing cycle: organisations without AI lose the talent needed to build AI.
3. Data Decay
Your data is most valuable when it's current and accessible. The longer you wait to build data pipelines and knowledge bases, the more institutional knowledge walks out the door, processes change undocumented, and data quality degrades. Data readiness isn't static. It erodes without investment.
4. Foundation Debt
Every month without a foundation means more future costs. Ad-hoc AI tools accumulate. Teams build workarounds. By the time you start a structured AI initiative, you'll spend the first month untangling the tools and processes that grew organically in the meantime.
5. Missed Learning
AI capability is as much about organisational learning as technology deployment. The enterprises that started in 2023 have a year of production experience: what works, what doesn't, where the edge cases are, how users adopt, how governance scales. This learning has no shortcut. You can only get it by doing.
The "Prudent Delay" Myth
The most common justification for inaction is prudence: "We're being careful. We're assessing. We don't want to move too fast."
This framing confuses two things:
- Reckless deployment (shipping AI into production without governance, testing, or oversight). This is genuinely dangerous and should be avoided.
- Structured exploration (running discovery, building data pipelines, establishing governance, deploying measured capabilities). This is the opposite of reckless.
Nobody is suggesting you deploy AI recklessly. The argument is that structured AI exploration is less risky than doing nothing, because doing nothing guarantees you fall behind while your competitors compound.
The Minimum Viable Action
If you do nothing else, run a 4-6 week AI discovery sprint. Map your highest-value opportunities. Assess your data readiness. Build a prioritised roadmap. Total cost: $20-50K. It's the single best investment you can make, even if you decide to wait another quarter to build.
How to Start (If You Haven't)
Week 1-2: Acknowledge the gap. Have an honest conversation about where your organisation stands relative to peers. Not to create panic, but to create urgency.
Week 3-6: Run an AI discovery sprint. Map your top 10 opportunities. Score them by impact, data readiness, and foundation potential.
Week 7-8: Present a roadmap to the board. Not "we need AI" but "here are the three highest-value capabilities for our business, here's the investment required, and here's the compound value from building them on a shared foundation."
Week 9+: Build. Start with the capability that has the highest combination of impact and foundation potential. Build the foundation and the first capability simultaneously.
The best time to start was 2023. The second-best time is this week.
- What if we genuinely don't have the budget for AI right now?
- A discovery sprint ($20-50K) is a fraction of most enterprise project budgets. It gives you the roadmap and business case to secure budget for the build phase. If even that isn't possible, start with internal AI literacy: workshops, experimentation with consumer tools, and identifying opportunities informally. The cost of informed inaction is much lower than the cost of uninformed inaction.
- Is it too late to start in 2024?
- No. The enterprise AI market is still early. Most organisations are in experimentation, not production. Starting a structured initiative now puts you ahead of the majority, and with the right foundation approach, you can close the gap with early movers faster than you think.
- How do we justify AI investment to a risk-averse board?
- Frame it as risk management, not innovation. "Our competitors are building AI capabilities that reduce their cost base by 20-40%. If we don't match this within 18 months, we face a structural cost disadvantage." Enterprise leaders respond to competitive risk more than they respond to opportunity cost.