What Enterprises Get Wrong About AI Adoption - Crawl, Walk, Run, Fly

Summary

The speaker discusses a 'crawl, walk, run, fly' model for AI adoption, emphasizing a gradual approach to deploying AI in businesses. Instead of centralized AI Centers of Excellence, the speaker suggests empowering product teams for organic exploration. The process starts with experimentation (crawl), identifying valuable tools (walk), operationalizing AI's success (run), and ultimately building a Center of Excellence when expertise has matured (fly). The key takeaway is to focus on curiosity, remove pressure for immediate ROI, and enable teams to adopt AI at a sustainable pace.

Takeaways

• 🚀 AI adoption can be accelerated by empowering decentralized teams, especially product owners and managers, rather than relying on a centralized Center of Excellence (CoE).
• 🧠 The 'crawl, walk, run, fly' model is the best approach to AI adoption, starting with exploration and discovery before operationalizing AI at scale.
• 📉 Many businesses fail at AI adoption due to unrealistic expectations of immediate ROI, leading to frustration and decreased productivity.
• 🔍 The 'crawl' phase focuses on fostering curiosity without expecting value upfront, allowing teams to explore AI without pressure.
• 🤖 A centralized AI CoE is premature for most companies due to the immaturity of AI technology and lack of industry best practices.
• 🏃‍♂️ The 'walk' phase involves finding one specific tool or use case where AI delivers measurable value, such as better, faster, or cheaper processes.
• 📊 The 'run' phase is where businesses begin to systematically adopt AI, automating more tasks and measuring performance improvements across multiple functions.
• 👨‍💻 One practical example of the 'walk' phase is the use of AI-generated art for a novel cover, demonstrating how AI can be cheaper, faster, and market-tested better than human artists.
• 💼 The 'fly' phase is when organizations create a true AI Center of Excellence, after years of building internal expertise and operationalizing AI across the company.
• 🌐 AI's value grows as organizations transition from individual experimentation to large-scale operational use, unlocking broader productivity and innovation benefits.

Q & A

What does the speaker mean by 'accelerationist' in the context of AI?

The speaker refers to 'accelerationist' as a meme and clarifies that it means being a 'techno-optimist,' someone who looks forward to the changes that technology, especially AI, will bring.

What is the key thesis of the speaker's argument about AI adoption?

The key thesis is that creating a centralized AI Center of Excellence is not the best approach for early AI adoption. Instead, empowering product teams to explore AI in a decentralized and organic manner yields better results.

Why does the speaker believe a centralized approach to AI adoption is obstructive?

The speaker believes that a centralized approach, like an AI Center of Excellence, is obstructive because the AI industry and individual contributors are still too immature, and the best practices for AI are still being developed.

What is the 'crawl-walk-run-fly' model the speaker discusses?

The 'crawl-walk-run-fly' model is a framework for adopting AI in stages: crawl (exploration), walk (finding specific value), run (operationalizing), and fly (building a Center of Excellence).

What should companies focus on during the 'crawl' phase?

During the 'crawl' phase, companies should focus on enabling their teams to explore and play with AI without any pressure to deliver immediate value or measurable ROI. The goal is to build familiarity and curiosity.

How do companies know when they are ready to transition from the 'crawl' phase to the 'walk' phase?

Companies are ready to transition from the 'crawl' to the 'walk' phase when they become bored of exploring. Boredom indicates that the initial curiosity has been satisfied, and it’s time to start looking for specific applications of AI that add value.

What is the objective of the 'walk' phase?

In the 'walk' phase, the objective is to find one or more AI tools or applications that provide measurable value, such as improving productivity or reducing costs, and to start incorporating them into business processes.

What example does the speaker give for using AI in the 'walk' phase?

The speaker gives the example of using AI-generated art for a book cover. AI helped him brainstorm, test, and finalize cover art faster and cheaper than using human artists, demonstrating how AI can add value to specific tasks.

What is the focus of the 'run' phase in AI adoption?

The 'run' phase focuses on operationalizing AI by systematically applying the successful AI tools and methods identified during the 'walk' phase. This involves training employees and scaling AI usage to automate and optimize more business functions.

When is a company ready to build an AI Center of Excellence?

A company is ready to build an AI Center of Excellence in the 'fly' phase, once it has developed in-house expertise, learned best practices, and matured in its AI capabilities. This is when the company can centralize AI leadership and formalize its AI strategy.