AI's True Impact: Transformation Beyond Automation

The Unseen Ripples: Why AI's True Impact Isn't About Automation, But Transformation

This conversation reveals a critical blind spot in how we assess AI's impact: the dangerous overemphasis on simple automation versus genuine job transformation. The easy, quantifiable approach of scoring jobs by their susceptibility to AI automation is not just flawed; it's a form of self-deception, akin to predicting the internet's impact in 1997 by focusing solely on digitizing existing paper processes. The true consequence of AI, like the internet before it, lies in its ability to fundamentally reshape what customers are buying and what constitutes value. Those who grasp this distinction--understanding the "job to be done" rather than just the tasks performed--will gain a significant advantage in navigating the coming disruption. This analysis is crucial for business leaders, strategists, and anyone whose livelihood might be affected by AI, offering a clearer lens through which to anticipate and adapt to future changes.

Why the Obvious Fix Makes Things Worse: The Illusion of Automation

The initial reaction to AI, much like the early days of the internet, is to focus on automation: identifying tasks that can be offloaded to machines. This is the "first-order" thinking, the immediate benefit that feels productive. However, as Benedict Evans and Tony Pell argue, this approach misses the forest for the trees. The real impact isn't about replacing existing jobs with AI performing the same tasks faster or cheaper. It's about how the underlying "job to be done"--what the customer is actually buying--is fundamentally altered.

Consider the example of travel agents. Early frameworks might have predicted their demise based on online booking platforms offering price transparency and liquidity. This was directionally correct; the way travel was booked changed. But it didn't account for disruptive innovations like Airbnb, which didn't just digitize hotel bookings but fundamentally redefined the concept of accommodation by leveraging underutilized physical assets. This is the "Uber test" of any framework: does it account for radical reinvention, or just linear extrapolation of existing models?

"Imagine you're clever and hard-working and you know, but you know, what would you do? Well, you'd look at retail and you'd say, I think the framework that went around was what's high touch and low touch. This is a sensible way of looking at it."

-- Benedict Evans

The transcript highlights how this linear extrapolation fails. Applying a simple "high touch" versus "low touch" framework to retail in 1997 might have predicted consumer electronics moving online faster than high fashion. This was mostly true, but it didn't predict the persistent struggle with returns and exchanges for clothing, a problem that persists decades later. Similarly, the idea that owning physical assets like hotels or taxis was the core business was upended by Airbnb and Uber, respectively. These platforms didn't just improve existing services; they created entirely new markets and business models that leveraged different forms of value.

The danger of focusing solely on automation is that it can lead to a misallocation of resources and strategic focus. Companies might invest heavily in automating tasks that are ultimately incidental to what their customers value most. This is particularly relevant with AI, where the temptation to score jobs by quantifiable automation potential--like assigning a numerical score of 72 to accountants--is strong but ultimately "ludicrous."

"It is crazy as humans that we like a good number regardless that makes us feel good and then know it could be completely wrong. Uh, but we're just like, oh, but that 72 helps me wrap my head around it. Does it? Even if it's complete bullshit."

-- Tony Pell

This numeric precision is an exercise in self-deception. The real question, as the speakers emphasize, is not if a task can be automated, but whether the automated task is the actual job the customer is paying for. For instance, a lawyer spending 32% of their day on phone calls might seem automatable by voice AI, but this ignores the core value proposition of legal counsel, which extends far beyond mere communication.

The Newspaper Paradox: When the Barrier to Entry Becomes the Product

The analogy of newspapers is particularly insightful for understanding this dynamic. A newspaper's core product is journalism--investigation, curation, recommendation, and opinion. However, their business model was intrinsically tied to a physical asset: printing and distribution. The internet, as a new distribution channel, automated away the "trucking and light manufacturing" aspect.

Initially, some might have thought this was a net positive, as it removed the costly physical component. However, this physical asset also served as a barrier to entry, protecting established players. When the internet unbundled these components, it created a catastrophe for many newspapers. The barrier to entry was gone, and the new distribution channel (the internet) was not what customers were primarily paying for; they were paying for the journalism itself.

This unbundling created new intermediaries and challenges. The rise of platforms like Substack, while lowering the barrier to entry for individual journalists, has also led to an overwhelming volume of content. The new challenge for consumers isn't finding content, but finding substance and expertise amidst the noise. This highlights a crucial downstream effect: automation of the delivery mechanism doesn't automatically translate to a better customer experience or a sustainable business model if the core value proposition is misunderstood or if the new landscape creates its own set of complex problems.

The "Hard Part" Advantage: Building Moats Beyond Code

The conversation pivots to software development and professional services, emphasizing that the "hard part" is rarely just writing the code. Companies like NationBuilder or Zillow exist not because their code is uniquely difficult to replicate, but because their success hinges on a complex interplay of factors beyond the technical. These include understanding market needs, go-to-market strategy, sales execution, pricing, and navigating existing industry dynamics.

"The problem isn't writing the code. It's everything else. It's working out what the product should actually be and what the code should be doing and how you're going to insert this into the industry and persuade people to use it and executing the sales force and getting the right go-to-market and the right pricing..."

-- Tony Pell

This is where delayed payoffs and competitive advantage emerge. While AI can automate code generation, it cannot easily replicate the years of experience, market intuition, and customer relationships that constitute the true moat for many successful companies. The businesses most vulnerable are those where the automated component was the core value proposition or the primary barrier to entry (like maintaining legacy COBOL code for banks). For others, like Doordash, the complexity lies in the marketplace dynamics, not the underlying software.

The implication is that focusing on the "hard part"--the strategic, human-centric elements that are difficult to automate--offers a more durable competitive advantage. This requires patience and a willingness to invest in areas that may not yield immediate, visible results, precisely because these are the areas where competitors are less likely to venture.

Actionable Takeaways for Navigating AI's True Impact

• Immediate Action (0-3 Months): Reframe "Automation" as "Transformation."

  • For every task currently considered for AI automation, ask: "Is this the core job my customer is paying for, or is it merely the way we deliver that job?"
  • Identify which parts of your business model are analogous to the "newspaper's printing press" (barrier to entry, not core value) versus the "journalism" (core value).
  • Challenge any internal scoring systems that focus solely on task automation percentages without considering the "job to be done."

• Short-Term Investment (3-12 Months): Map the "Hard Parts."

  • Analyze your business and identify the elements that are difficult to automate (e.g., customer relationships, strategic decision-making, complex sales cycles, unique market insights).
  • Invest in strengthening these "hard parts" where AI is less likely to provide a direct substitute.
  • Explore how AI can augment your human expertise in these difficult areas, rather than replace it.

• Medium-Term Investment (12-18 Months): Embrace the "Uber Test."

  • Continuously question your core assumptions about what customers value. Are you susceptible to a fundamental reinvention of your market, not just incremental improvements?
  • Look for opportunities where AI can enable entirely new value propositions that go beyond automating existing processes.

• Long-Term Strategy (18+ Months): Differentiate on Experience and Expertise.

  • For businesses where the "way" of doing things was automated (e.g., retail, travel), focus on enhancing the customer experience, curation, and unique value beyond mere transactional efficiency.
  • For professional services, emphasize the human element: strategic insight, judgment, and tailored advice that AI cannot fully replicate.
  • Recognize that the true competitive advantage lies not in being the cheapest or fastest, but in solving the actual job for your customer in ways that are difficult to disintermediate.