The stack everyone
is running
A mid-size independent hotel or boutique group in 2026 typically operates with some version of the same architecture: a Property Management System at the core, a Channel Manager connecting to OTAs, a Revenue Management System or revenue manager using spreadsheets, a CRM that may or may not be updated, and a website with a booking engine that probably hasn't changed since 2019.
Then, over the last three years, AI tools started arriving on top. A chatbot on the website. An AI content tool for writing room descriptions. A revenue management add-on that claims to use machine learning. A guest communication tool that generates automated messages.
The stack isn't the problem. The logic underneath it is.
Why adding AI to a
broken stack amplifies it
AI tools are amplifiers. They take whatever is in the system — data quality, process logic, ownership clarity — and make more of it, faster. A well-structured hotel operation with clean data and clear workflows gets genuinely better with AI layered on top. A poorly structured operation with inconsistent data and unclear ownership gets worse, faster.
The stack isn't the problem. The logic underneath it is.
Most hotel tech stacks in 2026 have three structural problems that AI makes worse if not addressed first. First: data fragmentation. Guest records live in the PMS, booking behaviour lives in the channel manager, communication history lives in email threads, and preference data lives in no system at all. AI tools that need a complete picture of the guest are working with a partial one. Second: no operational memory. Every season starts from scratch. Staff turnover resets institutional knowledge. The hotel has years of data but no structured access to it. Third: undefined ownership. When an AI tool produces an output — a price recommendation, a guest communication draft, a demand forecast — it's not always clear who is responsible for acting on it, or what happens when it's wrong.
The fix is not a
better tool
The answer to a fragmented tech stack is not a better integration platform, though those help. It's not a new PMS, though sometimes that's the right call. The answer is a structured knowledge layer — a single, maintained representation of what the hotel knows about itself, its guests, and its operational patterns — that sits underneath the tools rather than being scattered across them.
This is what SO Labs is building with Meridian. Not another tool to add to the stack. A foundation layer that makes the existing stack more useful — by giving every tool in it access to a consistent, structured, and queryable representation of the hotel's core knowledge.
The question for a hotel group in 2026 is not "which AI tools should we buy?" It's "what does our operation actually know, where does that knowledge live, and is it in a form that AI can use?" The tech stack follows from the answer. The AI tools follow from the stack. Not the other way around.