Thesis
Today, most enterprise AI implementations are point solutions with a strategy-enforced ceiling.
The greatest opportunity for an AI-native M&A or FP&A execution is one that:
- Establishes an enterprise-wide platform and decision engine that leverages deterministic workflows, parallelizes specialized agent execution, and compounds knowledge over time
- Evaluates and evolves its compute allocation strategy to drive scale and operating leverage, and
- Leverages a build vs. buy framework that provides future optionality to use best of breed solutions throughout
Anthropic Is Building the Platform
Opportunistically building the application layer (non-core focus)
- Anthropic (Claude Code, Cowork, Claude in Excel et al) is building the infrastructure and model layer.
- Does not plan to compete with applications built on top of the Claude platform.
Priority focus is showing customers what’s possible
- Suite of Claude offerings’ primary role is to demonstrate what can be built.
- Note: Will opportunistically go vertical when have model-level insight or ecosystem signaling function
The Concept of Compute Allocation
“We’re all becoming ‘compute allocators’ and our main job is to decide what’s worth spending compute on.”
What this means in practice
- Most generated tokens (up to 99%) will not/should not end up as production code.
- These tokens are generated to confirm exactly what should be shipped.
- Problem: a lot of AI implemented into Production today is vibe coded and likely should not be in this 1%.
Production Ready: Harnesses vs. Vibe Coding
What is a harness?
- The surrounding infrastructure, memory, and operational software that transforms a raw, standalone AI model into a reliable, autonomous agent.
- While the model provides the reasoning layer, the harness provides the controls, managing external tools, context, and guardrails.
Why does it matter?
- Vibe-coded apps should largely be the 99% of tokens that never make it to Production.
- Enterprise-grade harnesses make the 1% auditable, governable, and safe to deploy at enterprise scale.
Advanced Thinking: Early Days of the Model Routing Strategy
| The Problem | Processing all employee AI requests through large frontier models is expensive and often unnecessary for routine tasks. |
| The Framework for a Solution | An intelligent routing layer that sits between customers and models. Evaluates query complexity and routes accordingly: complex tasks go to frontier models while simpler tasks can go to smaller, cheaper, and often localized models. |
| The Business Impact | Dynamic routing reduces operational and inference costs while optimizing performance across the organization. Similar to operating leverage in finance: increasing revenue while reducing variable costs and overall cost structure. |
Observations: Demand Forecasting Workflow
Bespoke workflow that automated a previously tedious, time consuming, and difficult to audit process.
- While advantageous to the business, the scaling benefits had a ceiling.
- Greater benefits required a larger, enterprise-wide AI platform initiative
Pre-dated the concept of Model Routing Strategy
- Non-optimized token consumption, which also introduces a ceiling to the operating leverage that workflows such as this can drive for the business.
Production Ready Example #1: Sigma
“You can pretend vibe coding isn’t happening and find out the hard way the first time a handbuilt agent takes an action you cannot explain to a regulator, an auditor, or a customer. Or you can put it on a foundation that lets every app and every agent inherit the governance you already have in place.” — Mike Palmer, CEO at Sigma
Production Ready Example #2: Summation
To be decision-grade, AI must:
Understand context. Connect interdependent datasets across finance and operations and model their relationships, so metrics reflect reality.
Ensure traceability. Every answer is verifiable and defensible — with sources, lineage, and auditability.
Scale analysis. Explore thousands of questions in parallel so leaders don’t miss what matters.
Summation — Bellevue, WA
Conclusion
[TO COME].