Founders often ask us: what does it really take to build a durable AI company? In enterprise AI, we’re seeing two models emerge again and again: Oil wells (systems of record) & Pipelines (automation layers). 1️⃣ Oil wells become the source of truth for data & associated workflows. We see two entry points: Rip + Replace: When legacy systems are weighed down by tech debt, buyers are motivated to take the risk to replace with new AI-native systems. hyperexponential rebuilt insurance pricing & underwriting from ingestion to submission; Valon collapsed 25+ mortgage servicing systems into one; Vesta redesigned loan origination so tasks could run in parallel. Greenfield: When no system exists, startups capture customers early and grow with them. Rillet started as the first ERP for SMB finance teams, automating manual workflows, and has since expanded into replacing incumbents like NetSuite. Oil wells take longer to drill, but once established, they create deep, durable moats. Owning the system of record unlocks workflows no one else can build and builds switching costs. 2️⃣ Pipelines (automation/orchestration layers) -- These sit on top of existing systems and automate the “glue work” humans do between them. We’ve seen two main patterns: Fragmented Systems: When many entrenched systems coexist, pipelines unify workflows without requiring rip-and-replace. FurtherAI provides agentic workflows for insurance, automating cumbersome processes (submissions, loss runs, compliance) across multiple systems. Human Middleware: When humans are the bridge between systems, pipelines digitize that work. Concourse builds AI agents for finance teams, connecting into several financial systems so teams can query and analyze without manual effort. Sola lets customers record a workflow once and turns it into a live AI agent for tasks like invoice reconciliation. Customers don’t have to choose. Enterprises will often buy both: a new system of record in one area, lightweight automations in another. But for founders, the strategies are distinct but both can be massive. What matters is not trying to do both at once, but knowing which game you’re playing to win. New post by me & Joe Schmidt IV in comments!
Rajeev K.
11mo
Love this “Oil wells vs Pipelines” frame — especially the point about not trying to do both at once. We chose Pipelines deliberately — but we think the bar for pipelines is higher in the AI era. Most “pipeline” layers today are brittle map-scripts or UI workflow builders. Nap is an AI-native iPaaS for enterprise connectivity — built so pipelines learn, self-heal, and respect governance across fragmented systems and human middleware. What that means in practice: • Intent-level adapters that infer and evolve mappings as schemas, APIs, and policies change — without PS marathons. • Agentic orchestration that observes, retries, and remediates failures autonomously, with deterministic fallbacks when needed. • First-class governance — lineage, policy guardrails, auditability — so automation can actually go production-wide, not just pilot-wide. We’re not replacing systems of record; we make them work together intelligently. If oil wells are where truth lives, Nap is the adaptive connective tissue that keeps truth (and work) flowing — reliably enough to feel like core infrastructure. Would love to compare notes on where pipelines become durable moats in enterprise—this is exactly the game we’re playing to win.
Joydeep Rakshit
11mo
Angela Super insightful post. I'd add that many of the most successful "oil well" companies begin their journey by first building a "pipeline" solution. By starting with a lightweight automation layer, they can quickly get into the customer's workflow, prove value, and gather the data needed to eventually build the comprehensive system of record that replaces the incumbent. It's a smart, two-stage approach.
Evan Kramer
11mo
Angela Strange great post. Going beyond the theory need to consider the buyer decision. In the short term, there is a major switching cost to the oil wells as incumbents have both contracts and deep integration. This has bought them time to upgrade their AI. The pipelines right now seem to get more traction in GTM
Vijay Chakravarthy
11mo
You cannot build pipelines without having mini oil reservoirs available to you. So imho the pipeline strategy would need to encompass some level of (departmental) oil storage as well.
Arvind Bhawnani
11mo
The toughest part is picking one game and committing — oil well or pipeline, trying both early is a killer.
In a world that keeps shifting, the winner is whoever lets customers change their minds fastest. If “oil wells” give us data sovereignty and “pipelines” give us workflow velocity, what happens when the ground underneath the wells starts shifting faster than the pipelines can reroute. - Moats vs. Rivers- The real edge may belong to the startup that can turn its wells into springs - make its system-of-record modular enough that new pipes can plug in without asking permission. -The Hidden Third Pattern - We keep talking about replacing or gluing existing systems, but there’s a third vector emerging: “time machines.” Companies like Rewind or Personal.ai aren’t just storing data; they’re storing context at a point in time.The system of record isn’t the database—it’s the narrative. 3. Mirror Risk The pipeline becomes the rehearsal space; the well becomes the stage. 4. Talent Arbitrage Humans as “middleware” can specialize in the gaps the algorithms still find boring. The next billion-dollar pipeline might be one that pays bonuses to human reviewers for the weird edge cases they spot. 5. Ask the Counterfactual Zero-knowledge proofs for finance. Differential privacy for healthcare.The moat becomes trust, not data.