Matt Slotnick, the CEO of Poggio Labs and author of the Substack "cloud rants, predominantly," has become a leading voice in the transition from traditional SaaS to the "Agentic Era." His insights bridge the gap between high-level market strategy and the granular mechanics of software building.

The Architecture of AI & Agents

  1. Work as an Abstraction: "There’s a new abstraction for work, and that abstraction is agents." (Source)
  2. The Loop Principle: "Everything becomes a loop.[1] And agents are great at loops." (Source)
  3. Decoupling Growth: "The total amount of work done within an organization will dramatically increase, but decoupled from cost and headcount."[1] (Source)
  4. Atomic Units of Work: "Don't try to 'solve sales' with a button. Identify the atomic units of work, make those units better, and stitch them together." (Source)
  5. Human-in-the-Loop Transition: "We are moving from 'agent in the loop' to 'human in the loop,' where humans supervise the work rather than executing every step." (Source)
  6. Infinite Scaling: "With AI, agents can act on data at infinite scale and zero marginal cost."[1] (Source)
  7. The UI as a Guardrail: "The UI is the way an organization sets the guardrails for thousands of interrelated workflows that make a business run."[2] (Source)
  8. Nouns and Verbs: "The nouns and the verbs for how you use AI don't even exist yet. We are in the market development phase." (Source)
  9. Agent Memory: "The real evolving picture of a customer relationship lives in agent memory, not just the CRM." (Source)
  10. Orchestration Value: The lion’s share of value in the next decade will go to the layer that orchestrates work acrossapplications.

Modern Software Philosophy

  1. UI is Overrated: "Application UI is overrated. As visual creatures, we value what we see, but what lies underneath is what actually matters."[1][2] (Source)
  2. Systems of Record Gravity: "Data has gravity.[3] The system of record commands immense value because it creates influence on all other systems."[3] (Source)
  3. Application as Opinion: "Enterprise applications are just opinionated frameworks for how a business process should be structured." (Source)
  4. The Best Business Model: "Shrink-wrapped software [SaaS] is the best business model on earth." (Source)
  5. Consumer-Grade Expectations: "If IT doesn’t provide modern tools, users will find them. IT will eventually find themselves in a forced procurement cycle for what their users already use." (Source)
  6. Low-Code is a Paradigm, Not a Category: "Saying 'we are a low-code company' tells you almost nothing about the value you provide." (Source)
  7. The End of 'Cheap' Growth: "When the unit economics of software growth shift, your balance sheet cannot save you." (Source)
  8. Platform Definition: "A platform is an application that provides interfaces for other applications to be built upon." (Source)
  9. Vibe Coding: The shift where the "how" of code matters less than the intent of the creator. (Source)
  10. The CRM Battleground: "The CRM isn't the battleground... at least not in the way you’re thinking. The battle is for the work that happens outside of it." (Source)

AI Sales & GTM Strategy

  1. Winning POVs: "A Point of View (POV) is more powerful than a feature list in modern enterprise sales."
  2. Research is a Waste: "Sellers waste too much time on manual research that AI can execute in 60 seconds." (Source)
  3. Amplifying People: "AI isn't a 'seat tax'; it's about amplifying the ability for people to do their jobs 30% more productively." (Source)
  4. Leading with Language: "To win, a seller must speak in the specific industry language of their customer, not generic sales-speak."
  5. Selling to Winners: "Focus on the market-leading Winners who will actually buy AI. Companies squeezing cash are often afraid of new tech." (Source)
  6. The Shadow AI Pattern: "In some companies, reps hide AI usage because they're afraid the CIO will take it away once they realize how much more efficient it makes them." (Source)
  7. AI for Sales Context: The value of AI in sales is "discovering, synthesizing, and contextualizing data" from thousands of unstructured sources. (Source)
  8. Relationship Data: AI allows you to capture the "rich and dynamic context" of a lead that manual logging misses.
  9. Outcome Centricity: In subscription software, the power rests with the customer’s ongoing success, not the initial sale. (Source)
  10. Winning the Future: "We won the last game [SaaS]; now AI helps us win the future." (Source)

Product & Team Execution

  1. Iterative Innovation: "Product innovation in AI comes from an iterative approach, not a rigid 12-month roadmap." (Source)
  2. Workflow First, Tech Second: "Team construction should be influenced by deep understanding of end-user workflows, not just technical ability." (Source)
  3. Managing Infinite Expectations: One of the hardest parts of AI is managing the "infinite expectation" users have for the technology. (Source)
  4. Data vs. Context: "Data without contextual action is just a liability."
  5. The maestros of Excel: "Today, people are the engine of business, the maestros of Excel macros... AI is the new engine." (Source)
  6. Infrastructure Costs: "Spending $1-2bn on elite talent is 'nothing' compared to $70bn on annual cloud capex." (Source)
  7. GPU over Headcount: Elite talent in the future will ask for the "fewest people reporting to me, and the most GPUs." (Source)
  8. The Logic of Applications: An application is essentially a set of "nouns and verbs" for a business. (Source)
  9. Building in Public: Building in public and engaging on social platforms is a critical modern distribution channel for founders. (Source)
  10. Open Core Risks: "Open source core is often a misaligned incentive for venture-backed SaaS." (Source)

Business Logic & Personal Philosophies

  1. Business as a Machine: "A business is just a process machine, dynamically allocating resources toward their most efficient use." (Source)
  2. Cycling & Business: Entrepreneurship is like cycling: success is about consistency, endurance, and knowing exactly when to push. (Source)
  3. Vibe Organizations: "Running an organization will start to feel like 'vibe coding'—setting the intent and letting agents execute." (Source)
  4. The Unbounded Task TAM: AI expands the TAM (Total Addressable Market) because it addresses tasks that were previously too expensive for humans to do. (Source)
  5. Cloud Rush Backlash: Hasty cloud migrations without workflow redesign often stall transformations. (Source)
  6. Data Sovereignty: In the agentic era, data control and compliance are the primary friction points for enterprise adoption.
  7. Hybrid Software: Pragmatism usually wins; hybrid models that mix local security with cloud scale are the future.
  8. Scale or Die: "In verticalized markets, you either scale and own the customer relationship or you die." (Source)
  9. Cloud Platform Incentives: Cloud incentives often drive behavior that is contrary to what the end-user actually needs. (Source)
  10. The Messy Reality: The transition from SaaS to AI is "far messier" than the marketing tropes suggest, but the destination is revolutionary. (Source)

Sources

  1. substack.com
  2. substack.com
  3. substack.com