Jed Ayres published a piece in Fast Company this week that every founder building in the IT operations space should read carefully. Not as a ceiling on what is possible, but as the clearest documentation available of where the market opportunity begins.
The article captures what Digital Employee Experience technology delivers at enterprise scale. A national insurance broker with more than 1,000 retail locations cut outages by 62% across 32 states in a single year, saving close to $14 million. A financial services firm that previously learned about problems through help desk calls now catches them before the first user picks up the phone. A health system identified a misconfiguration through synthetic testing before any clinician was affected. The numbers are documented, the ROI is real, and the market problem is genuine.
What the article also reveals, without framing it this way, is that the signal detection layer has created the conditions for the next architectural shift. And that shift is where the next generation of defensible positions will be built.
Detection Proves the Market. Resolution Owns It.
The DEX category has established something important: enterprises will pay real money to know about problems faster. The $14 million case study is not an outlier. It is evidence that the pain is large, the buyers are motivated, and the infrastructure investment is justified.
That proof of market is the foundation the next layer sits on.
When a DEX platform surfaces a hung session or a corrupted registry setting, the resolution still requires institutional context: what this failure signature means in this environment, on this stack, for this class of user. In most enterprises today, that context lives with the engineers who have accumulated it over time. The DEX platform gets the signal to the right person faster. The resolution still depends on what that person knows.
This is a specific instance of a pattern that CPAG research calls the Biological Middleware Tax: the $600 billion annual cost of knowledge workers functioning as the connective layer between systems that generate signals and systems that need those signals acted upon. It is not a critique of the signal layer. It is a description of where the next addressable market begins.
The Architectural Layer the Market Is Waiting For
The founders who build durable positions in the IT operations space will do it by solving the second problem: not just surfacing the signal, but building the institutional knowledge architecture that allows an agent to traverse from alert to resolution, with human judgment preserved for the calls that genuinely require it.
That layer has a specific architectural name in the RaaS framework: the High-Fidelity Repository. It is a graph-structured knowledge architecture that stores not just what happened in an IT environment, but why, in what context, and what sequence of actions produced resolution the last time this class of failure occurred. Property graph structures allow AI reasoning engines to traverse relationships between failure types, environment configurations, user profiles, and resolution histories in a form that relational databases cannot replicate. The engineer’s institutional knowledge becomes the training set. Over time, the repository executes the resolution, and the engineer reviews the output rather than producing it.
This is the architectural difference between a platform that is paid for access to alerts and a platform that is paid for problems solved.
The Business Model Follows the Architecture
For founders, the practical implication is sequencing. The DEX market has proven that enterprises will invest in the detection layer. Founders who build the resolution layer above it are entering a market with documented demand, motivated buyers, and an incumbency gap they can occupy from day one without competing on the territory that is already well defended.
The High-Fidelity Repository build has four stages, documented in the CPAG RaaS Vendor Transition Playbook. Data Cartography in the first three months maps the three to five data domains that carry the most resolution logic: failure signatures, environment configurations, historical resolution patterns. Graph Migration from months four through nine moves one domain per quarter into a graph-structured layer. Agentic Connection from months ten through eighteen wires agents to the repository via MCP protocols, with permission tiers governing what the agent can read, recommend, execute, and escalate. The Un-Rippable Asset phase from months eighteen through thirty-six is where the moat compounds: every resolved incident adds institutional knowledge to the graph, and a customer who has run on the repository for twenty-four months has a switching cost that no per-resolution price comparison can overcome.
The business model discipline that governs this architecture is the 1-to-4 Rule: for every dollar of AI compute cost, a RaaS vendor must capture four dollars of Resolution Value to maintain the 75% gross margins that institutional investors require. A platform built on the repository layer can apply this discipline from its first contract because each resolution has a defined cost to serve and a defined outcome that is verifiable, attributable, and finite.
Founders who enter this space with a RaaS pricing model from day one hold a structural advantage that is genuinely difficult for any incumbent to replicate quickly, because the transition requires dismantling a revenue model that is currently working. The founder has no such constraint. The repository builds. The moat compounds. The switching cost grows.
The Door the DEX Market Opened
Ayres closes his piece with a framing that is exactly right: the future belongs to smarter partnership between AI and the data DEX generates. The market has validated the data layer. The AI partnership is the next chapter.
The founders who write that chapter will be the ones who asked what the partnership looks like when the resolution is also software, who built the institutional knowledge architecture that made autonomous resolution possible, and who priced on outcomes delivered from their first enterprise contract.
The DEX market did not create a ceiling. It created a floor.
The question worth building toward: when your platform can traverse from alert to resolved without routing through a human, what does that compound into over twenty-four months of customer data?
The thinking behind that question connects to a deeper architectural and governance argument at middlewayinai.com, where the question of where human judgment must be structurally preserved, rather than merely assumed, gets the analysis it deserves.