AI Wins in Industry Software
Varun Rustomji@varun
The idea
I believe AI models will become cheap and interchangeable, so the real value flows to software built deep inside one industry, the tools that already run the daily workflow and hold the data and regulatory know-how a general model can't replicate. This basket spreads across healthcare, government and legal, construction, finance and insurance, and property and supply chains, rather than betting on a few names. It keeps a modest cash reserve to add on pullbacks, since this is a long-term shift I expect to play out over years, not a quick trade.
What it owns
- Landlords, restaurants and warehouses run daily operations on these tools, so AI features reach them through software already in use. · ToastIt serves as the checkout, payroll and payments system for a large share of US restaurants, capturing data from every order and shift., AppFolioIt runs leasing, rent collection and maintenance workflows for property managers, embedding AI assistants into their daily operations., Manhattan AssociatesLarge retailers and distributors run their warehouses and inventory on this platform, which is increasingly moving to cloud-based tools., Descartes Systems GroupIt connects shippers, carriers and customs agencies on a single network to handle filings, with data that becomes more valuable as participation grows.21%
- Builders and engineers store project designs and records in this software, where AI can save significant time on day-to-day work. · Procore TechnologiesIt sits at the center of daily project management for owners, contractors and subcontractors, accumulating years of construction project data., AutodeskAutoCAD and Revit are the standard design tools architects, engineers and builders work in every day, creating strong customer lock-in., Bentley SystemsIt makes the engineering software used to design and run infrastructure like roads, bridges, water systems and power grids tied to strict codes., ServiceTitanIt runs dispatch, quotes and payments for plumbers, HVAC and electrical contractors, and recently expanded to pest control and lawn care companies.20%
- Courts, cities and law firms rarely switch software and pay for tools that already understand their regulatory requirements. · Tyler TechnologiesCities, counties and courts run taxes, permits and court records on this software, and government buyers rarely switch after setup., Thomson ReutersIts legal research platform contains specialized content and case law that general AI models cannot replicate without years of data., IntappLaw, accounting and advisory firms use it for client intake, conflict checks and time tracking, the compliance-heavy operations of professional practices.19%
- Drug makers and hospitals operate under strict rules and manage private data, making their software extremely difficult to replace. · Veeva SystemsDrug makers run clinical trials, sales teams and regulatory filings on this platform, making it the natural entry point for AI adoption in life sciences., DoximityIt is where US doctors already network, message and provide telehealth, layering AI for clinical notes and paperwork into established habits., Waystar HoldingIt handles insurance claims and payments for hospitals and clinics, rules-heavy work where years of payer data provides competitive advantage.17%
- Banks and insurers run their core operations on industry systems built up over decades with claims and account data embedded throughout. · Guidewire SoftwareProperty and casualty insurers run policies, billing and claims on systems that take years to replace once established., Jack Henry & AssociatesIt runs core accounts and payments for community banks and credit unions, heavily regulated work that small banks prefer not to hand to general AI tools., nCinoIt is the loan software banks use from application through approval, with compliance rules built into the workflow.16%
- Cash8%