The Great Reorg of Enterprise Saas in The New Ai World

How AI Is Quickly Reshaping the Enterprise Stack From the Outside in

After nearly two decades investing in and studying enterprise software, I have learned that this market rarely moves fast, but it always moves with consequence. For fifteen years, a small group of horizontal SaaS platforms formed the digital backbone of the modern enterprise. They standardized sales, HR, finance, IT, customer service, and engineering. They professionalized operations during the rise of cloud computing. And they generated some of the most durable revenue in the history of software.

Yet today, a deeper shift is underway. Not the collapse of these incumbents, but a structural reorganization catalyzed by artificial intelligence. AI introduces a fundamentally different architectural model for how software interacts with data, workflows, and human expertise, and this shift is steadily redistributing where value is created inside the enterprise.

1. The Era of the Horizontal Giants

Beginning around 2010, several companies cemented their positions as universal enterprise systems of record. Their dominance is visible in the scale of their financial performance:

Salesforce
• 38B revenue in 2024
• 9 percent annual growth
• 230B market cap
• 150,000+ enterprise customers
• More than 70 percent of Fortune 500 using at least one Salesforce cloud

ServiceNow
• 11B revenue in 2024
• 22 percent annual growth
• 175B market cap
• More than 7,700 enterprise customers
• Over 85 percent of Fortune 500 running ServiceNow in at least one function

Workday
• 8.5B–9B revenue
• 17 percent annual growth
• 60B+ market cap
• Over 65 million workers represented inside the platform

Atlassian
• 5.2B revenue
• 20 percent annual growth
• 300,000+ customers globally

SAP and Oracle (macro context)
• SAP: ~33B revenue, 180B market cap
• Oracle: ~52B revenue, 310B market cap

From 2010 to 2020, these companies represented one of the most consistent periods of growth in enterprise technology. Their models worked because they delivered uniformity, compliance, reporting, and team-wide coordination during the most dramatic corporate modernization cycle in decades.

2. The Architecture Beneath Their Success

These platforms were built around the needs of an enterprise environment that relied heavily on human operators. Their shared architecture included:

•Structured data models: Every process reduced to accounts, tickets, cases, assets, employees, SKUs, or tasks.
•Human-driven workflows: Multi-step routing, assignment, approvals, and queues that assumed a person would complete each step.
•High configuration complexity: Enterprises often spent millions on consultants to tailor these systems.
•Broad cross-industry generalization: The same CRM schema served manufacturing, biotech, telecom, and retail.
•Long implementation cycles: Six- to eighteen-month projects for HR, CRM, ITSM, or ERP modernization.
•Low adaptability to unstructured data: These systems functioned best when humans translated complexity into structured fields and forms.

This design made perfect sense at the time. The primary challenge was scaling and standardizing global operations, not automating them. Software’s job was to help humans coordinate, not to take over the work.

3. How AI Challenges the Horizontal Model

AI introduces a new set of capabilities that directly collide with the assumptions of the earlier architecture. Instead of merely supporting human decision-making, AI is increasingly capable of performing it.

•Unstructured data competency: AI can absorb call transcripts, documents, PDFs, sensor streams, medical notes, imagery, and contracts — data that traditional systems ignored or required manual transcription.
•Decision-making and reasoning: AI models can evaluate context, reconcile conflicting information, identify anomalies, and make recommendations that mimic expert judgment.
•Autonomous execution: AI agents can trigger multi-system sequences, complete forms, submit filings, route cases, process claims, and resolve issues without human intervention.
•Deep domain learning: Vertical models can internalize regulatory nuance, industry vocabulary, and workflow patterns, something horizontal systems were never designed to encode.
•Continuous operation: AI functions around the clock, not on task-based human cycles.

These capabilities do not fit neatly into platforms that were architected around human participation. Even with substantial AI R&D budgets — Salesforce alone spends ~8B annually on R&D, ServiceNow ~3B — retrofitting intelligence into structured, multi-tenant, configuration-heavy platforms introduces real constraints. The incumbents cannot optimize deeply for a single vertical without breaking the model that made them global.

4. The Vertical AI Advantage

Vertical AI companies begin by embracing the complexity horizontal platforms abstracted away. Rather than building general-purpose tools, they immerse themselves in the operational realities of a single industry. Their advantage is not speed or novelty but depth.

Key strengths of the Vertical AI architecture
•Proprietary data advantage: Industries like healthcare, logistics, finance, and defense produce unique unstructured data unavailable to generic incumbents.
•Domain-specific workflows: Vertical platforms map the real processes — not the abstracted ones — including exceptions, edge cases, and regulatory constraints.
•Operational execution: Instead of capturing tasks, these systems complete them.
•Higher ROI per workflow: Automating a multi-step underwriting or clinical workflow often produces 10–30x efficiency gains versus incremental improvements in horizontal systems.
•Built-in expertise: The software internalizes industry knowledge historically held by consultants, analysts, and operations teams.

Examples illustrate how quickly this depth translates into value:

Healthcare
• 25–30 percent of all provider administrative expense tied to documentation, billing, and prior authorization.
• Vertical AI platforms are already reducing cycle times by 40–70 percent in certain workflows.

Logistics
• Exception rates across global freight average 12–18 percent.
• AI-driven resolution reduces manual intervention by up to 80 percent.

Financial Services
• AML, KYC, and compliance checks often consume thousands of hours per institution.
• AI platforms complete portions of these workflows autonomously with higher accuracy.

Industrial & Defense
• Predictive maintenance reduces downtime by 20–40 percent.
• AI systems can monitor millions of sensor signals that humans could never feasibly process.

The deeper the domain, the more defensible the vertical platform becomes.

5. A Transition That Will Be Gradual and Permanent

Unlike prior technology shifts, this transition will not be defined by ripping out incumbent systems. Enterprises will keep Salesforce, ServiceNow, Workday, SAP, and Oracle at the core because they remain essential for structured data governance, compliance, identity, reporting, and financial controls. But the locus of operational decision-making will slowly migrate to the vertical layer.

Over time, enterprises will operate with a dual architecture:

•Horizontal platforms for system-of-record functions— data, compliance, permissions, reporting, and standardization.
•Vertical platforms for system-of-action functions— execution, intelligence, decision-making, and industry-specific workflows.

This is not a collapse. It is a reallocation. The incumbents will continue to grow, but the next major wave of enterprise value creation will concentrate in companies that embed deep domain intelligence and harness unstructured data to automate work end-to-end.

Big Desk Energy | Subscribe | Big Desk Energy

startup insights, stories, and vibes sent to your inbox every Tuesday

mail.bigdeskenergy.com/subscribe?_bhba=a912eba6-7a35-4c1b-a9cb-9721b5c72389

Venture Scout

High-quality software startups delivered straight to your inbox, every Wednesday.

www.venturescout.io/subscribe?_bhba=a912eba6-7a35-4c1b-a9cb-9721b5c72389

Shoppers are adding to cart for the holidays

Over the next year, Roku predicts that 100% of the streaming audience will see ads. For growth marketers in 2026, CTV will remain an important “safe space” as AI creates widespread disruption in the search and social channels. Plus, easier access to self-serve CTV ad buying tools and targeting options will lead to a surge in locally-targeted streaming campaigns.

Readour guideto find out why growth marketers should make sure CTV is part of their 2026 media mix.

Shoppers are adding to cart for the holidays

Share Trace Cohen's Vertical Ai Investor Newsletter