There’s a moment before the dot-com imploded in 2000, when the promise of Infinite Excess shined brightest. A series of commercials by Qwest, a long-defunct telecommunications & broadband service provider from that era, embodied this notion, with the video above contrasting the mediocrity of limited choice with the magic of infinite choice.
And to clear, there were a number of truly terrible concepts — Beenz, FreePC, DigiScents to name a few. But, there were many that were just too early, such as WebVan, Broadcast.com and Pets.com.
This is a central lesson of the various waves of technology, from mainframes to PCs, LANs, Internet, Mobile, Social and AI.
That each successive wave is bigger and more disruptive than the wave that preceded it.
This is in part because the Biggest Ideas move inexorably from impossible, scarce, unusable and/or mediocre, to inevitable, obvious, ubiquitous and multi-varied.
With the rise of AI, I’d argue that pretty much any idea is pursuable to functional prototype, proof-of-concept or 1.0 product.
What this means is an era of applications that are fully tweaked out to the level of Exactly What You Want (EWYW) and Exactly How You Want(EHYW) , down to the level of terminology, tracked metrics, inputs, function, flow, reporting, workflows and checklist-driven processes.
In the domain I specialize in — proptech for commercial real estate – this opens an entirely new computing surface to users, teams and enterprises, one where Claude, ChatGPT and Copilot take all of that context and enable you to use natural language to build pretty frickin incredible stylized reports, dynamic site plans & stacking plans, merchant health tracking, work order job type efficiency analysis, etc.
Everyone from senior execs to junior property managers, accounting personnel, leasing agents and asset managers and analysts is starting to use these tools as part of their daily process.
Sidebar: In the macro, there are over 1.4 billion users of these ever-improving AI engines, giving massive agency to everyone.
Garbage In, Garbage Out
There’s a saying. Garbage in, garbage out, which means that you can only do something useful with the validated, high-fidelity data that you actually have access to.
Data quality, coverage and context matter.
The way to think of that one is that each data set is an ingredient that you can create recipes for to “bake” something craveable in Claude, etc.
More and better ingredients, when married with AI, yield both better recipes but also, unleashes MORE chefs.
Our particular approach, called Datex AI Live, can leverage all of a real estate portfolio owner’s data, including Financials, Budgets, General Ledger, Rent Rolls, Collections, Occupancy, Leasing, Retail Sales, Work Orders, Investors, Loans and CRM.
The Great Thing about Industry Standards
This is why we’re so excited about the emergence of MCP, or Model Context Protocol, which is an industry standard for providing standardized data access with context to various AI engines.
When you marry standardized data access with context (a shopping center has tenants that occupy suites, and those tenants have category types, such as fast food, may be national tenants, such as Chipotle, report sales, pay rents, impact occupancy, have co-tenancy provisions, etc.), you now have a robust data model for understanding the various “fields of play” for a shopping center and by extension, a portfolio of shopping centers.
Moreover, because the above delimiters can be standardized across a portfolio, you have a scalable mechanism for tracking, segmenting and benchmarking, reporting, etc.
Keeping it real, there’s a timeless joke that the great thing about standards is there are “so many to choose from.”
So, how should one assess the adoption-worthiness of a given standard? There are three measures.
One, does it specifically solve the problem it was designed to address? Two, is there broad industry adoption of the standard in the form of actual shipping products (not just press releases)? Three, is there an ecosystem of third-parties building solutions on top of the standard?
USB-C, HTML/HTTP, Wi-Fi and SQL are all examples of widely adopted standards.
MCP meets all of these measures by virtue of the ubiquity of MCP adoption across the major chat environments, the value-add AI harnesses and the major SaaS platforms.
Two final thoughts. One, data plumbing, which is what MCP is, is only as powerful as the range of data types and the fidelity of the data that flows through those pipes, so beware that a LOT of the MCP connectors that vendors will tout will have limited data sets, so what you can DO with that data is pretty “meh.”
Two, you need security, governance and logging to limit what superset the AI has access, the subset a given user has access to, and tracking who’s accessing what.
Hit me up if you want to see some magical stuff built using Datex AI Live.
Leave a Reply