AI & Agentic Platform Monetization

What are we actually monetizing — and what do you do about it on Monday?

1. Measuring the output of a model, and why token counts, prediction counts, and consumption metrics fail as value metrics unless structurally tied to pricing
2. Economically viable workloads — the Sora vs. Claude Code contrast
3. Why the enterprise SaaS narrative collapsed, and the tale of two software companies
4. Which frameworks to implement first, and in what order · the Week One Assessment and seven critical assessment points
5. Why you take ownership of something the business already cares about rather than something you find interesting · delivering inside a quarter · building the track record that becomes your shield when you ask for real budget

Assignment 1 Complete an initial assessment of your business against the seven critical points. This becomes the initial state for every subsequent exercise in the course.

Cases: SAP and NVIDIA.

1. The layered platform architecture — information core, operations, product, AI interface layer, customers and partners on the outside
2. The L0–L5 maturity model, and how SAP climbed it without taking the ERP system offline
3. Three-phase optimization: task → workflow → outcome · Feature → Product → Platform
4. Two platform design patterns — horizontal breadth first or vertical depth first, and why depth has monetized more reliably so far
5. The AI factory duality: you’ll use these tools to build your platform, and the gaps they deliberately leave are your opportunity
6. The Costco hot dog methodology · gap analysis · how a platform play accelerates a hardware business
7. The AI supply chain — data as raw material → AI factory → agentic platforms · three orders of optimization, each a larger market than the one beneath it · functional → reliable → affordable · where a non-technology company actually enters

Cases: Mercedes-Benz, NVIDIA, Uber.

1. Agile business → continuous transformation → continuous disruption
2. Growing the pie versus taking a bigger slice of it · the shift from coded platforms to generated platforms
3. The ecosystem triangle made concrete — how each edge creates demand on the others
4. The physical platform and the operations platform, and why the factory now extends into the car
5. Why continuously improving autonomy converts a one-time purchase into recurring revenue · flywheel acceleration
6. Pure-play simulation platforms as a monetizable layer in their own right · the simulation model taxonomy · multiple monetization — one model family reused across many domains
7. Licensing patterns, maintenance fees, and monetizing non-human workers, machines, and connections

Exercise Find the fleet-management angle in the Mercedes–NVIDIA–Uber ecosystem. Uber has never managed a fleet; it isn’t a core competency and it isn’t monetized, but it is a real cost structure. Where does a new platform partner enter, and what new flywheels does their entry create?

Cases: Tencent / WeChat and Meta.

1. Surfaces replace apps — action, commerce, distribution, and support surfaces, plus micro-surfaces like loyalty
2. Communication layers becoming workflow layers · capability enablement instead of app downloads
3. The trusted orchestrator and the new control plane · why the surfaces where your customers express intent may not be ones you own, and what that does to your moat
4. Intent-in, campaign-out: the advertiser brings goals and assets, the platform generates creative, targets, expands audiences, and runs the campaign
5. Why monetization should be quantified against customer revenue growth rather than productivity or cost savings
6. The transition from ads to commissioned sales — the setup for the outcomes economy

Exercises Evaluate Apple’s hardware and software ecosystems separately — where are Apple’s action surfaces, and if your customers reach you through Apple devices, what is your role in that ecosystem? · You run an ad agency whose value proposition is being automated for free by the companies that own the ad inventory. Your one asset is cross-platform data on ad effectiveness that Meta doesn’t have. How do you reinvent the business?

The turn. Transformation means deliberately moving pieces of the business model and operating model into the technology model — the alternative produces the feeling that everything is being torn apart.

1. Start with the workflow, not with a transformation program
2. Opportunity → use case → workflow → task
3. Deterministic versus stochastic transformation, and the two different frameworks they require — maturity models for the first, gates and balances for the second
4. Content-to-cash and the other named pipelines · introducing the outcomes economy model · pragmatic futurism

Assignment 2 Decompose one opportunity in your business into use cases and workflows. Classify each workflow as deterministic or stochastic. For the deterministic ones, sketch the maturity model. For the stochastic ones, define your gates.

The core monetization block. Case: Salesforce.

1. The agent taxonomy — conversational, proactive, ambient, autonomous, and collaborative agents
2. When the empire strikes back: an existing platform, business model, and customer expectations pulling against the new thing you’re building
3. Salesforce’s pricing journey through per-conversation, per-action, and hybrid consumption-outcomes licensing · pricing as a listening journey
4. The single pane of glass, and the shift from product catalogs to intent and outcome catalogs
5. The AI Monetization Pyramid — capabilities, autonomy, and intelligence at the base; then domain expertise; then self-improvement; then outcomes at the apex
6. Why ad-based and compute-based models are structurally misaligned with AI value creation · why different domains must carry different prices
7. Capabilities licensing — the Mayo Clinic to rural hospital pattern · value-share and outcome pricing, and what it takes to teach customers to measure the value you’re claiming a percentage of
8. The orchestration imperative: misalignment across pricing model, monetization model, technology value creation, and workflow re-orchestration — each simultaneously an opening to disrupt others and an exposure to being disrupted
9. Presenting winners and losers side by side, because C-level leaders who can’t see why one company is succeeding and another failing will not act

Exercises Where can AI create a novel marketplace that has not existed before — distinguishing creating a market from making an existing one scale more efficiently? · Analyze a workflow currently owned by Excel or Outlook. Can an agentic tool deliver the same outcome by re-orchestrating the workflow rather than replicating the features?

Zero to internal buy-in — the 70% of the problem that isn’t technical.

1. The four categories of barrier, and why the technical one is the easiest · the 70-20-10 rule
2. What winners do — data foundations, workflow re-orchestration, measurable outcomes, review processes — against what losers do: weak guardrails, no oversight, brittle production controls, no training
3. Shadow AI: a published approval process with real consequences, paired with moving fast enough that nobody needs to route around you
4. Gaming out incentives before you change a behavior · the innovation tax as the CFO-facing argument
5. The listening tour — fix what’s already broken first, because credibility is built by sticking around until an imperfect solution works
6. Mapping promoters, fence-sitters, and detractors, and watching actions rather than words
7. Internal thought leadership as an inbound funnel: build enough expertise that colleagues bring you their problems instead of the reverse

Walmart · JPMorgan Chase · Siemens + NVIDIA · Eli Lilly. Chosen deliberately to be regulated, physical, and legacy-heavy rather than technology-first.

1. Walmart — partnership as a monetization strategy; finding the gaps in a larger ecosystem and filling them; why Walmart chose to own its surface and integrate on its own terms rather than surrender the customer relationship
2. JPMorgan Chase — proof the frameworks hold in the hardest regulatory environment there is. The critical success factors stack led by an unambiguous C-suite mandate · Power Up: upskilling non-technical staff then promoting internally · security and governance as first principles that expand rather than restrict what the platform can do · closing question: are they fighting the last war?
3. Siemens + NVIDIA — the circular partnership, where each partner’s product improves through the other’s use of it · why reference environments and blueprints, not demos, build trust for high-risk platform adoption
4. Governance and trust as architecture — four agent governance archetypes: standalone, proactive, swarms, and physical-digital agents, each breaking your controls differently · the three types of drift · the least-impactful-action principle · continuous monitoring and adaptive governance
5. Eli Lilly — the AWS model: an internal best-in-class capability becomes a product sold to the market. The four-tier platform, the agentic wet lab, federated learning that turns competitors into partners and then into customers, and why a domain leader may beat a frontier lab in its own domain

Exercise Which parts of your platform will customers immediately understand, and which won’t they? Where are you asking them to imagine a re-orchestrated workflow they’ve never seen? Identify where showing beats telling, and design the reference environment you’d need to build.

Pulling the architecture back together, then the framework that integrates everything.

1. The Core-RIM framework — an intelligent core of increasing capability surrounded by a rim of irreducible complexity that stays with people
2. The AI 80/20 rule, and why the last 20% carries 80% of the work
3. Horizontal breadth and vertical depth, and how the T becomes a circle as platforms accumulate adjacent workflows
4. Roadmapping with the product layer as your workflow intelligence layer and the platform as your outcome intelligence layer — including features you cannot build yet
5. Parallel Maturity — why sequence matters (expert systems and contextual data gathering first, or everything downstream is uneconomic or infeasible), and why you no longer have time to go step by step
6. The seven parallel ladders: technology capability, business management (zombie → task-managed → intent-and-outcome-managed), visibility, data, workflow, adoption journey, and competitive ambition
7. Why technical decisions have strategic consequences and strategic decisions have technical ones — and why the two can no longer be separated

Final assignment Place your business on each ladder. Define the initial state, the justified final state, and the sequence between them. Identify which third-party tooling accelerates which steps. Attach the monetization you expect to unlock at each stage, and the pricing model that will capture it.