The Boring 90% (The Models Are Ready. Your Company Isn’t.)

2026-07-12 · Source: Data Science on Medium · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Cloud Computing & IT Infrastructure · Depth: Intermediate, medium

Summary

The article, "The Boring 90%", argues that while AI models are advanced and improving, most companies and individual workflows lack the foundational infrastructure to effectively utilize them. This creates a significant gap between public AI-first claims and on-the-ground reality, where context is scattered across disparate tools and systems. Brian Balfour notes that few leaders are connected to actual AI adoption rates. The core problem isn't model intelligence but the "broken foundation" of personal and organizational data management. Shaun Clowes emphasizes that durable advantage in the AI era comes from data and business rules, not models or interfaces, suggesting 90% of effort should focus on clean, current, structured, and connected data pipelines. The author's experience building a "product management brain" confirmed that data plumbing, not intelligence, was the bottleneck.

Key takeaway

For AI Architects or MLOps Engineers tasked with deploying AI solutions, recognize that your primary challenge isn't model selection but establishing robust data foundations. You should prioritize investing 90% of your effort into cleaning, structuring, and connecting your organization's disparate data sources and codifying unwritten business rules. This foundational work, though unglamorous, creates the durable competitive advantage and ensures your AI initiatives deliver real value, rather than amplifying existing chaos.

Key insights

The primary challenge in AI adoption is not model capability but the unprepared data and organizational foundations.

Principles

• Durable AI advantage stems from data and business rules.
• Context management is 90% of AI implementation effort.
• Scaffolding around models has a short shelf life.

In practice

• Prioritize structuring and connecting scattered data.
• Codify unwritten business rules and decision logic.
• Maintain data currency and context discipline.

Topics

• AI Adoption Challenges
• Data Infrastructure
• Business Rules Management
• Organizational Readiness
• Data Governance
• MLOps

Best for: CTO, VP of Engineering/Data, Executive, Director of AI/ML, AI Architect, MLOps Engineer

Related on AIssential

• Your AI Strategy Isn’t Failing Because of the Model — It’s the Data — Data governance and quality, not AI models, are the primary bottlenecks preventing enterprise AI from scaling to production.
• Is context the new compute? How the AI race is moving to the data layer — AI competitive advantage now stems from proprietary organizational context and robust data infrastructure, not just model compute.
• The AI Reality Check: What Big Tech Learned the Hard Way in 2026 — Effective AI integration requires robust organizational capabilities, not just advanced models, to bridge the gap between demo and production.
• The Tidy House — Successful AI adoption hinges on organizational readiness and robust, clean data infrastructure, not just advanced models.
• The Data Layer Is Changing Not the Way Everyone Thinks — Reliable data foundations, not advanced models, accelerate AI initiatives by empowering existing teams.
• Three insights you may have missed from theCUBE’s coverage of Snowflake Summit 2026 — Enterprise AI's next wave demands robust data foundations, new security frameworks, and governed intelligence for successful production deployment.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Data Science on Medium.