Intro
L1
·
Quiz
·
Lab
L2
·
Quiz
·
Lab
L3
·
Quiz
·
Lab
L4
·
Quiz
·
Lab
Module Test
Building an AI-First Business · Introduction

Internet-first was once a company category. AI-first is becoming one.

There's a real difference between putting AI into a company and building a company around AI. This course is about the second.

In the early 2000s, the phrase internet-first company described something specific: a business whose core operations, distribution, and product lived online by design, not as a retrofit. Amazon was internet-first; Barnes & Noble was not, even after it built a website. The difference mattered for a decade and shaped a generation of tech.

AI-first is the same kind of distinction. You can bolt AI onto an existing business — most companies are doing exactly that right now. Or you can build a business whose unit economics, product experience, workforce structure, and competitive moat all assume AI as a foundation. The second is much harder. It's also where the next generation of category-defining companies is being built.

This course is about that second path. It covers what an AI-first business actually looks like in different industries, how to design one from scratch, how to retrofit an existing company when the second path is the right answer, the economics of AI at scale, the founder profiles of current AI-first success stories, and the honest failure patterns of the AI-first companies that didn't make it.

If you finish every module, here's who you become:

  • You'll understand the structural difference between bolting AI onto a business and building a business whose model assumes AI from the start.
  • You'll be able to identify AI-native opportunities — problems where AI creates a durable competitive advantage, not just a faster workflow.
  • You'll know how to design a proprietary data strategy that compounds over time and becomes harder for competitors to replicate.
  • You'll leave with a framework for making the tech stack decisions that define AI-first companies: what to build, what to buy, and why it matters.
  • You'll recognize the hiring profiles, cultural norms, and org structures that make AI-first companies operate differently from AI-enabled ones.
  • You'll be able to diagnose the failure patterns that have ended AI-first companies — and design around them before they become your problem.
  • You'll think like a founder who builds AI into the architecture of a business, not a manager who applies AI to the edges of one.
Lesson 1 · Module 1

AI-First vs. AI-Added

The distinction that separates transformative companies from those that merely bought a subscription.
What does it actually mean to build a business around AI — and why does most "AI adoption" miss the point entirely?

In 2023, two insurance companies both announced they were "using AI." The first bolted a ChatGPT plugin onto its existing claims portal and called it an AI strategy. The second — Lemonade — had already been running a system called AI Jim since 2016: an AI that handled the entire first notice of loss, cross-referenced claims against policy data, and in one documented case, paid a claim in three seconds. Same technology era. Completely different businesses.

Defining the Spectrum

Every business today sits somewhere on a spectrum between AI-Added and AI-First. The distinction is not about how much money you spend on AI tools, how many subscriptions your team has, or whether your CEO mentions AI in earnings calls. It is about where AI sits in the architecture of your value creation.

An AI-Added business has human-designed workflows at its core, with AI sprinkled in at the edges: a Grammarly plugin here, a ChatGPT summary there, maybe a dashboard that uses machine learning to surface anomalies. These tools reduce friction. They do not reshape the fundamental economics of the business.

An AI-First business was designed — or fundamentally redesigned — so that AI sits at the center of value delivery. The workflows, the cost structure, the scalability assumptions, and often the product itself are impossible without AI at the core.

AI-AddedAI deployed as an enhancement to existing human workflows. Improves speed or quality at the margins. The business functions without it.
AI-FirstAI embedded as the primary engine of value creation. The business's core product, cost model, or delivery mechanism is architecturally dependent on AI.
Core LoopThe repeating cycle through which a business creates and delivers value. In an AI-first business, AI sits inside this loop — not outside it.

Three Documented Cases

Lemonade (Insurance, 2015–present). Lemonade was not an insurance company that added AI — it was incorporated with the explicit design that AI would handle underwriting signals, fraud detection, and claims processing. By 2020, its loss ratio and operating costs were structured differently than any traditional insurer of comparable size because the variable cost of adding a new policyholder was near zero — a structural advantage unavailable to AI-Added competitors.

Duolingo (EdTech, 2023 pivot). Duolingo's April 2023 layoffs of contract content writers were accompanied by a public statement from CEO Luis von Ahn: the company was moving to AI-generated content for its language exercises. This was not a cost-cutting measure dressed up as strategy — the company's existing content production bottleneck meant it could only support a limited number of languages and exercise types. AI removed that bottleneck structurally. By Q3 2023, Duolingo reported a 62% increase in daily active users partly attributed to the speed of new course deployments made possible by the AI content pipeline.

Klarna (Fintech, 2024). In February 2024, Klarna's CEO Sebastian Siemiatkowski announced that its AI assistant — built on OpenAI — was doing the work of 700 customer service agents in its first month, handling 2.3 million conversations. Klarna did not add AI to an existing contact center. It rebuilt the contact center around AI and resized its human workforce accordingly. This is a structural, not additive, deployment.

Why This Matters Economically

AI-Added businesses improve margins incrementally. AI-First businesses change the shape of their cost curves. When Klarna adds 100,000 new customers, its customer service cost barely moves. When a traditional bank adds 100,000 customers, it hires proportionally. That asymmetry compounds over time into a structural competitive moat.

The Adoption Illusion

A common mistake — made by executives, consultants, and investors alike — is to conflate AI usage rate with AI-first posture. A company where every employee uses Copilot daily may be profoundly AI-Added: humans still design the process, AI assists execution. A company with lower measured AI usage but where the core product literally cannot exist without a model running at its center is AI-First.

The question to ask of any organization is not "do we use AI?" but "if AI were switched off tomorrow, would our core value proposition still function?" For an AI-Added business, the answer is yes — slower, but yes. For an AI-First business, the answer is no.

The Diagnostic Question

If your AI tools went dark at 9am Monday, would your core product still ship? Would customers notice a degraded experience, or would the product simply not exist? Your honest answer locates you on the spectrum.

Quiz — Lesson 1

AI-First vs. AI-Added · 3 questions
1. Lemonade's AI Jim is best characterized as an example of which posture?
Correct. AI Jim handles the full first notice of loss — cross-referencing policy data and paying claims autonomously. The value delivery mechanism is AI-dependent, which is the defining characteristic of AI-First.
Not quite. Lemonade was architected from the start so that AI is the primary agent in claims processing — not a human-workflow enhancement. That makes it AI-First by definition.
2. A company gives all 500 employees access to Microsoft Copilot and reports 90% daily usage. Based on the framework in this lesson, this company is most likely:
Correct. High usage rate does not equal AI-First posture. If humans still design all processes and AI assists execution at the edges, the company is AI-Added regardless of how many employees use AI tools daily.
The lesson explicitly distinguishes usage rate from posture. Copilot assists humans in their existing workflows — it does not redesign the business's core value-creation loop. That is the definition of AI-Added.
3. Which diagnostic question best reveals whether a company is AI-First or AI-Added?
Exactly right. The "AI offline" test cuts through all the noise. If the answer is yes, the business is AI-Added. If no — if the core value proposition collapses without AI — the business is AI-First.
Usage counts, budgets, and executive commitments are all proxy signals that can mislead. The structural test — whether the core product survives without AI — is the definitive diagnostic.

Lab 1 — Classify & Diagnose

Apply the AI-First / AI-Added framework to real company descriptions

Your Task

You'll describe a real or hypothetical company's AI usage to the assistant. The assistant will help you classify it on the AI-First / AI-Added spectrum, identify where AI sits in the value-creation loop, and suggest what a genuine AI-First redesign might look like.

Try: "A law firm where associates use AI to draft first-pass contracts, but partners review and sign off on everything." — or describe a company you know.
AI Lab Assistant
Lesson 1 · Classification
Welcome. Describe how a company you know (or one you're building) uses AI. I'll help you classify it on the AI-First / AI-Added spectrum and identify where AI actually sits in the value-creation loop. Be as specific as you can — vague descriptions get vague analysis.
Lesson 2 · Module 1

The Architecture of an AI-First Company

What changes structurally when AI moves from the edges to the core.
How do AI-first companies actually wire themselves differently — in operations, in headcount, in how they scale?

When Klarna rebuilt its customer service function around OpenAI's models in 2023, it did not simply automate existing scripts. It redesigned the entire interaction architecture: the AI had direct access to transaction data, could initiate refunds, update account states, and escalate to human agents based on sentiment signals it detected in real time. The humans who remained were no longer first-line responders — they were edge-case specialists and AI supervisors. The organizational chart changed shape. The cost per resolved ticket dropped by roughly 80 percent. The company did not become more efficient. It became a structurally different type of company.

Four Structural Differences

AI-first businesses differ from AI-added ones across four structural dimensions:

1.
Cost Curve Shape
AI-added: costs scale with volume. AI-first: marginal cost of additional output approaches zero after model and infrastructure investment.
2.
Headcount Architecture
AI-added: more customers → more people. AI-first: more customers → more data → better model → better product. Headcount decouples from revenue.
3.
Data as Infrastructure
AI-added: data is a reporting asset. AI-first: data is a production input. Every customer interaction feeds the model that serves the next customer.
4.
Speed of Iteration
AI-added: new product features require human design cycles. AI-first: capability can expand through model updates and prompt engineering without traditional development sprints.

The Flywheel Difference

The most consequential structural difference is what happens to quality over time. In a traditional business, quality is a function of the skill and experience of your workforce — it is costly to improve and easy to degrade through turnover. In an AI-first business, quality is increasingly a function of data and model iteration. Every customer interaction is, in principle, training signal.

Waymo's autonomous vehicle program illustrates this starkly. By 2023, Waymo had accumulated over 20 million miles of real-world autonomous driving data. This data is not a report — it is an infrastructure asset that makes the next mile safer than the last. A competitor entering the space cannot simply hire better drivers; it must accumulate comparable data, which requires comparable time and scale. The flywheel of an AI-first business compounds in ways that traditional competitive advantages do not.

Contrast this with a taxi company that uses AI to optimize surge pricing. That is AI-added: useful, profitable, but the pricing model does not make the rides themselves better, and the advantage does not compound through usage.

The Compounding Test

Ask: does each customer interaction make the product better for the next customer? If yes — if usage feeds improvement — you have the structural flywheel of an AI-first business. If usage just generates revenue but not capability, the flywheel is absent.

Organizational Implications

The structural shift in AI-first companies creates a new type of org chart. Anthropic, OpenAI, and Cohere are extreme examples — the majority of their headcount is researchers and engineers, not operators. But even non-AI-native companies that make the AI-first transition show predictable patterns:

AI/ML engineers move into product roles
Prompt engineers become a formal function
Data pipelines treated as core infrastructure
Traditional ops headcount compressed
Mid-level knowledge workers most displaced
QA shifts from human review to model evaluation

None of this means AI-first businesses have no humans. It means the type of human work changes. Klarna's remaining customer service staff handle disputes where emotional intelligence and legal judgment are required — tasks where the model's failure rate remains high. The humans are not gone; they are repositioned at the edges where AI still fails.

Structural Diagnostic

Map your company's core workflow. At each step, ask: is AI inside this step (AI-First signal) or does it assist a human who owns this step (AI-Added signal)? The ratio of inside to outside gives you your structural posture.

Quiz — Lesson 2

AI-First Architecture · 3 questions
4. Which of the following best describes the "flywheel" structural advantage of an AI-first business?
Correct. The AI-first flywheel is a compounding data loop: usage generates training signal, training signal improves the model, the improved model creates better product, which attracts more usage. This compounds over time in a way traditional advantages do not.
The flywheel is specifically about data compounding — usage improving capability which attracts more usage. Cost savings are a real benefit but not the structural flywheel mechanism described in the lesson.
5. Waymo's 20 million miles of real-world autonomous driving data is described in the lesson as:
Right. The lesson frames Waymo's data corpus as infrastructure — not a report, not a marketing tool — because it actively feeds model improvement that makes the core product (safe autonomous driving) better over time. A competitor cannot bypass this with capital alone.
The lesson explicitly frames Waymo's data as an infrastructure asset — one that actively improves product quality with each mile added. This is distinct from marketing or financial assets.
6. After Klarna rebuilt its customer service function around AI, its remaining human agents primarily handled:
Correct. The lesson notes that Klarna's remaining staff handled disputes "where emotional intelligence and legal judgment are required — tasks where the model's failure rate remains high." Humans moved to where AI still fails, not where AI already succeeds.
The lesson is clear: Klarna's humans were repositioned to edge cases — emotionally complex disputes and legally ambiguous situations — precisely because these are the scenarios where current AI models still have high failure rates.

Lab 2 — Map the Architecture

Identify the AI-first structural signals in a real business

Your Task

Walk through a company's core workflow with the assistant. Together you'll identify which steps have AI inside them vs. assisting humans, whether a data flywheel exists, and what the headcount architecture looks like.

Try: "We run an e-commerce business. AI recommends products on the homepage and our marketing team uses ChatGPT to write emails. How do we map this?" — or describe your own company's workflow.
AI Lab Assistant
Lesson 2 · Architecture Mapping
Let's map your business architecture. Walk me through your core workflow — how does your company create and deliver value, step by step? I'll help you identify where AI sits inside each step vs. merely assisting humans who own each step, and whether you have the structural conditions for an AI-first flywheel.
Lesson 3 · Module 1

The Transition Problem

Why most companies that try to become AI-first fail, and what the ones that succeed do differently.
What actually prevents a traditional business from becoming AI-first — and is the barrier technical, organizational, or something else?

IBM spent over a decade and billions of dollars trying to make Watson a commercial AI platform. The technology was not the failure point. Watson could process language. It could analyze medical records. The failure was structural: IBM sold Watson into organizations where the workflows were not redesigned around AI. Hospital systems that licensed Watson for Oncology kept their existing diagnostic processes and asked Watson to fit inside them. By 2022, IBM had sold or wound down most Watson Health divisions. The lesson was not about AI capability — it was about the impossibility of grafting AI onto unchanged organizational structure and expecting AI-first outcomes.

The Three Barriers

Companies attempting the transition from AI-added to AI-first consistently encounter three categories of resistance:

1. Data ReadinessAI-first architectures require clean, structured, accessible data as a production input. Most established companies have data in silos, governed by different teams, in formats that were never designed to feed a model. Fixing this is expensive and slow — often 12–24 months before meaningful AI-first deployment is even possible.
2. Process Lock-InExisting workflows have been optimized for human execution over years or decades. Regulatory compliance, contractual obligations, and institutional knowledge are embedded in these processes. Redesigning them for AI at the core requires dismantling things that work, which organizations resist even when the redesign would be clearly superior.
3. Incentive MisalignmentMiddle managers whose teams are displaced by AI have no career incentive to champion the transition. The people with the deepest process knowledge — the ones needed to redesign workflows — are often the same people whose roles are most threatened. This is not cynicism; it is organizational reality documented in every major transformation study.

What Successful Transitions Look Like

The companies that have successfully transitioned share a pattern. They do not attempt to become AI-first everywhere simultaneously. They identify one core workflow where AI-first architecture is structurally possible — where data is available, the process can be redesigned, and the economic incentive is large enough to override organizational resistance — and they go fully AI-first in that one place first.

JPMorgan Chase's COiN program (Contract Intelligence, 2017) is the most cited example. The bank identified commercial loan agreement review as a target: 12,000 hours of lawyer time per year reviewing standard documents. COiN reviewed those same documents in seconds. Crucially, JPMorgan did not try to make its entire legal operation AI-first. It picked the most tractable workflow, made it AI-first, demonstrated the economic case, and used that to build organizational will for the next workflow. By 2023, JPMorgan had over 300 AI use cases in production — but the path started with one.

Stitch Fix (fashion subscription, 2011–present) provides a different pattern. Its founding premise was that AI would drive clothing selection — every box a customer receives is selected by a combination of algorithm and human stylist, with the algorithm doing the first-pass filtering from millions of SKUs. The human stylist adds the final editorial judgment. This is a hybrid AI-first architecture: AI owns the core filtering function; humans add the edge-case judgment. This was designed in from the start, not retrofitted.

The Retrofit Trap

IBM Watson's failure and dozens of similar enterprise AI deployments share a common pattern: organizations expected AI to improve outcomes without changing the processes that determined outcomes. AI-first architecture requires process redesign. There is no shortcut through this.

The Starting Point Question

For an existing business attempting the transition, the practical question is not "how do we become AI-first?" but "where should we start?" The answer involves three filters:

Filter What to Look For Red Flag
Data availability High-volume, structured, historically consistent data about the workflow Data locked in PDFs, email threads, or human memory
Economic size Large enough cost or revenue impact to justify redesign investment Marginal workflows that won't move the needle if transformed
Organizational will A senior champion who benefits from the transformation and can absorb the resistance Mid-level sponsorship with no executive backing and no budget authority
Practical Takeaway

The best starting workflow for an AI-first transition is not the most exciting one — it is the most tractable one. High volume, clean data, large economic impact, and a senior champion who won't be fired for disrupting it. JPMorgan picked loan agreement review for exactly these reasons.

Quiz — Lesson 3

The Transition Problem · 3 questions
7. IBM Watson Health's failure was primarily attributed to:
Correct. The lesson is explicit: Watson's failure was not a technology failure. Hospital systems kept existing diagnostic workflows and asked Watson to fit inside them. Without process redesign, AI-first outcomes are structurally impossible regardless of AI capability.
The lesson's analysis of Watson is specifically about organizational structure, not technology. Watson could process language and analyze medical records — the failure was that clients didn't redesign their workflows to put AI at the center.
8. JPMorgan's COiN program is cited as a successful transition example because it:
Exactly right. COiN targeted commercial loan agreement review specifically — tractable, high-volume, large economic impact. The lesson notes JPMorgan did not try to make its entire legal operation AI-first at once. One workflow, proven case, then expansion. By 2023: 300+ AI use cases in production.
The lesson explicitly notes JPMorgan "did not try to make its entire legal operation AI-first" — it picked one tractable workflow, made the economic case, and used that to build will for the next. That sequenced approach is the cited success pattern.
9. According to the three-filter framework for choosing a starting workflow, which combination signals the best AI-first starting point?
Correct. The three filters are: data availability (high-volume, structured), economic size (large enough to justify redesign), and organizational will (senior champion with authority). JPMorgan's COiN hit all three.
The framework in the lesson has three specific filters: data availability, economic size, and organizational will (specifically a senior champion). Technology maturity alone is not a filter — tractability across all three dimensions is what matters.

Lab 3 — Find Your Starting Workflow

Apply the three-filter framework to identify the best AI-first entry point

Your Task

Describe your business (or a business you're analyzing) and its major workflows. The assistant will apply the three-filter framework — data availability, economic size, organizational will — to help you identify which workflow is the best candidate for an AI-first transition.

Try: "We run a 200-person accounting firm. We do audits, tax prep, and advisory work. Most data is in Excel files and PDFs. Partners are skeptical but the managing partner is open." — or describe your actual situation.
AI Lab Assistant
Lesson 3 · Workflow Selection
Let's find your AI-first entry point. Describe your business and its major workflows — what you do, how you do it, what data you have, and roughly what the organizational dynamics look like. I'll apply the three-filter framework (data, economics, organizational will) to identify which workflow gives you the best chance of a successful AI-first transition.
Lesson 4 · Module 1

Competitive Moats in an AI-First World

What actually protects an AI-first business — and what used to work but no longer does.
If AI models are commoditizing rapidly, where does durable competitive advantage come from in an AI-first business?

In early 2023, a leaked internal Google memo titled "We Have No Moat, And Neither Does OpenAI" argued that open-source AI models were closing the capability gap with proprietary ones at a rate that would eliminate model quality as a durable advantage. Whether that specific prediction proves accurate, the underlying logic is sound: if the model is the moat, and models are becoming commodities, the moat does not hold. The companies with durable AI-first advantages are not winning on model quality alone. They are winning on data, distribution, and workflow lock-in.

Where Moats Actually Form

In an AI-first competitive landscape, durable moats form in four places. Model quality is not the primary one:

Moat Type How It Works Real Example
Proprietary Data Data that cannot be replicated — generated by your customers, your operations, or your unique access — and that makes your model better than anyone else's for your specific domain. Bloomberg's financial data corpus; Veeva's pharmaceutical CRM data; Waymo's miles driven
Workflow Lock-In AI embedded so deeply in customer workflows that switching requires rebuilding the workflow itself, not just changing a tool. GitHub Copilot embedded in developer IDE; Harvey AI in law firm document workflows
Network Effects + AI More users → more data → better model → better product → more users. The AI amplifies a traditional network effect into a faster flywheel. Waze traffic routing; Duolingo exercise quality improving with more learners
Brand Trust in High-Stakes Domains In domains where AI errors have serious consequences (healthcare, legal, finance), established trust and regulatory credentials create a barrier that technical quality alone cannot overcome. Epic Systems in healthcare AI; established accounting firms in AI-assisted audit

What No Longer Works as a Moat

Three traditional competitive advantages are being rapidly eroded in AI-first markets:

Access to talent alone. In traditional software, hiring the best engineers was a durable advantage. In AI-first, talent matters but is rapidly commoditized through open-source models, APIs, and tooling. A startup with three engineers and access to GPT-4 can build capabilities that would have required 50 engineers five years ago.

Feature richness. AI capabilities can be added to products so rapidly that feature gaps close within quarters, not years. The pace of iteration in AI-enabled products makes traditional product moats fragile.

Scale without data differentiation. Being large used to mean better resources and distribution. In AI-first markets, scale only matters if it generates proprietary data. A large company processing generic transactions has no AI advantage over a smaller company processing the same generic transactions. Scale + unique data = moat. Scale alone does not.

The Bloomberg Terminal Precedent

Bloomberg built a financial data terminal moat not on technology but on data and workflow lock-in. Traders integrated Bloomberg so deeply into their workflows that the $24,000/year terminal became functionally irreplaceable. Bloomberg GPT — launched in 2023 — extends this: a 50-billion parameter model trained on Bloomberg's proprietary financial corpus. The model is not the moat. The corpus is the moat. The model just makes the corpus more accessible.

Building a Moat from Scratch

For founders and builders, the implication is clear: if you are building an AI-first business, ask from day one what your proprietary data source will be. The answer drives architecture decisions. If you generate unique behavioral data through your product (like Lemonade's claim patterns), protect that data and build your model on it. If you serve a domain with no existing structured data corpus, consider whether your product can generate that corpus through operation — and whether the resulting data is defensible.

Harvey AI, the legal AI platform backed by Sequoia and OpenAI, illustrates this pattern. The company's 2023–2024 growth was not primarily a function of model quality — law firms can access the same base models. Harvey's advantage comes from fine-tuning on legal documents generated through its platform's operation, and from workflow integration that creates switching costs. The moat is in the data and the embeddedness, not the base model.

Module 1 Core Principle

AI-first is not a technology posture — it is an architectural, organizational, and strategic posture. The technology is increasingly available to everyone. What separates durable AI-first businesses is what they build on top of the technology: proprietary data, embedded workflows, and network effects that compound with every interaction.

Quiz — Lesson 4

Competitive Moats in AI-First · 3 questions
10. According to the lesson, Bloomberg GPT's primary moat is:
Correct. The lesson makes this explicit: "The model is not the moat. The corpus is the moat." Bloomberg's decades of proprietary financial data — terminal data, news feeds, analytics — is what no competitor can replicate. The model is the access mechanism.
The lesson is direct on this point: "The model is not the moat. The corpus is the moat." A 50-billion parameter model can be matched; decades of proprietary financial data cannot be rapidly replicated.
11. The Google internal memo "We Have No Moat" argued that the primary risk to AI competitive advantages is:
Right. The memo's core argument was that if the model is the moat, and open-source models close the capability gap, the moat does not hold. The lesson uses this to argue that durable AI-first moats must be built on data, workflows, and network effects — not model quality alone.
The memo's specific argument was about open-source model capability closing the gap with proprietary models — making model quality an unreliable long-term moat. The lesson uses this to redirect toward more durable moat types.
12. Which of the following is identified in the lesson as a traditional competitive advantage that is being eroded in AI-first markets?
Correct. The lesson identifies feature richness as a weakened moat because "AI capabilities can be added to products so rapidly that feature gaps close within quarters, not years." The three eroded moats are: talent access alone, feature richness, and scale without data differentiation.
The lesson identifies three eroding moats: talent access alone, feature richness, and scale without data differentiation. Feature richness is specifically cited because AI-enabled iteration speed collapses the timelines that used to protect feature advantages.

Lab 4 — Design Your Moat

Identify your defensible advantage in an AI-first competitive landscape

Your Task

Describe a business you're building, running, or analyzing. The assistant will help you identify which of the four moat types (proprietary data, workflow lock-in, network effects + AI, brand trust) applies to your situation, assess the durability of your current advantages, and surface moat-building moves you may not have considered.

Try: "I'm building a hiring tool that uses AI to screen resumes. We have access to 50,000 historical hiring decisions from our beta clients. Where's my moat?" — or describe your actual business.
AI Lab Assistant
Lesson 4 · Moat Design
Let's identify your durable AI-first competitive advantage. Describe the business — what it does, who the customers are, what data you generate or have access to, and what your current differentiation claim is. I'll map your situation against the four moat types and help you figure out what's actually defensible vs. what's a temporary lead that competitors will close.

Module 1 Test

What AI-First Actually Means · 15 questions · Pass at 80%
1. A company's core product is an AI model that generates personalized insurance quotes in real time using behavioral and demographic data. If the AI were removed, the product would not exist. This company is best classified as:
Correct. The defining criterion is whether the core product functions without AI. In this case it does not — making it unambiguously AI-First.
If the product cannot exist without AI, it is AI-First by definition. The diagnostic question from Lesson 1 resolves this immediately.
2. Lemonade's AI Jim is documented in this module as handling which function?
Correct. AI Jim handled first notice of loss — cross-referencing claims against policy data and in one documented case paying a claim in three seconds autonomously.
The module documents AI Jim handling first notice of loss, including cross-referencing policy data and autonomous claim payment — not marketing, underwriting approval, or legal escalation.
3. In February 2024, Klarna announced its AI assistant handled 2.3 million conversations in its first month. This was equivalent to the work of approximately how many human agents?
Correct. Klarna's February 2024 announcement cited 700 full-time equivalent agents worth of work handled by the AI in its first month.
The module states the AI was doing the work of 700 customer service agents in its first month, handling 2.3 million conversations.
4. The "AI offline test" asks: if your AI tools went dark, would your core product still function? An AI-Added business would answer:
Correct. For an AI-Added business, the answer to the offline test is yes — the business functions without AI, just more slowly or with lower quality. For an AI-First business, the answer is no.
The module defines AI-Added as a business where "the business functions without it" — making the correct answer "yes, slower, but yes."
5. Which structural difference is described as a "flywheel" specific to AI-first businesses?
Correct. The AI-first flywheel is a compounding data loop: usage → data → model improvement → better product → more usage. This compounds in ways traditional competitive advantages do not.
The flywheel described in Lesson 2 is specifically about data compounding through usage. Each interaction improves the model; the model improves the product; the product attracts more interactions.
6. By 2023, Waymo had accumulated over 20 million miles of real-world autonomous driving data. The module describes this as:
Correct. The module frames it as infrastructure — actively feeding model improvement that makes the core product better over time. Not a passive asset; an active production input.
The module explicitly uses the word infrastructure — data that actively feeds model improvement, making the product better with each mile added. Not a passive or marketing asset.
7. Duolingo's 2023 shift to AI-generated content for language exercises primarily addressed which structural bottleneck?
Correct. The module notes that Duolingo's content production bottleneck meant it could only support a limited number of languages and exercise types. AI removed that structural limit — a textbook example of AI reshaping business architecture.
The module specifically identifies content production as the structural bottleneck AI removed — Duolingo could only support limited languages and exercise types before. That constraint disappeared with AI content generation.
8. IBM Watson Health's commercial failure is cited in the module primarily as evidence of which failure mode?
Correct. The module's analysis is explicit: "The technology was not the failure point." Hospital systems kept their existing workflows and asked Watson to fit inside them. Without process redesign, AI-first outcomes are structurally impossible.
The module says the technology was not the failure point. Watson could process language and analyze records. The failure was that clients didn't redesign workflows to put AI at the center — the retrofit trap.
9. JPMorgan's COiN program initially targeted commercial loan agreement review because it was:
Correct. COiN hit all three filters: data (12,000 hours of annual legal review = high-volume structured documents), economic size (significant lawyer time cost), and organizational tractability. It was the right starting point by the module's framework.
The module's three-filter framework — data availability, economic size, organizational will — explains why COiN was the right starting point. It was tractable across all three dimensions, not the most strategic or regulated workflow.
10. The module identifies "incentive misalignment" as a barrier to AI-first transition. This refers to:
Correct. The module explicitly notes that the people with the deepest process knowledge needed to redesign workflows are often the same people whose roles are most threatened. This creates structural resistance that is organizational, not technical.
The module's "incentive misalignment" refers specifically to organizational resistance: middle managers and process experts who are most needed for redesign are often most threatened by it. Their incentives run counter to the transformation.
11. Stitch Fix's founding architecture is described in the module as a "hybrid AI-first" design because:
Correct. The module describes Stitch Fix's architecture as AI doing the first-pass filtering from millions of SKUs (core function), with human stylists adding the final editorial judgment (edge-case function). AI-first where tractable, human at the boundary.
Stitch Fix's design puts AI inside the core loop (filtering millions of SKUs) and humans at the edges (final editorial judgment). That is the definition of hybrid AI-first architecture as described in the module.
12. The leaked Google memo "We Have No Moat" argued that model quality is an unreliable long-term competitive advantage because:
Correct. The memo's argument was specifically about open-source model capability catching up to proprietary models — which, if true, means "better model" is not a durable moat. The module uses this to argue for data, workflow, and network-effect moats instead.
The memo's specific claim was about open-source model capability closing the gap with proprietary models rapidly. If the gap closes, model quality as a moat collapses. The module uses this logic to argue for more durable moat types.
13. Bloomberg GPT is described in the module as an example of which moat type?
Correct. The module is explicit: "The model is not the moat. The corpus is the moat." Bloomberg's decades of proprietary financial data is what cannot be replicated. The model makes the corpus more accessible, but the data is the defensive asset.
The module makes a precise distinction: "The model is not the moat. The corpus is the moat." Bloomberg GPT's advantage is the proprietary financial data corpus — a data moat, not a network effect or workflow lock-in moat primarily.
14. Harvey AI's competitive advantage, as described in the module, comes primarily from:
Correct. The module explicitly notes that law firms can access the same base models as Harvey. Harvey's advantage is the fine-tuning data generated through operation and the workflow embeddedness that creates switching costs — data and lock-in moats, not model access.
The module notes law firms can access the same base models. Harvey's moat is the fine-tuning corpus built through operation and the deep workflow integration that makes switching costly — not model access or first-mover advantage.
15. According to the module, "scale without data differentiation" is a weakening competitive advantage because:
Correct. The module states: "Scale + unique data = moat. Scale alone does not." A large company processing generic transactions has no AI advantage over a smaller company processing the same generic transactions. Data differentiation is the operative variable, not scale per se.
The module's formula is specific: "Scale + unique data = moat. Scale alone does not." If your scale generates only generic, non-proprietary data, it creates no AI-first advantage over smaller competitors with access to the same model APIs.