# Elorati

> AI services and custom development studio. Custom software since 2012, AI systems since 2023. Designs, builds, and runs AI systems and custom software for operators, from one-week automations to full platforms. Three engagement models: Advisory (embedded strategy), Studio (fixed-scope custom development), Managed (build-then-manage).

Machine-readable surfaces: the full blog corpus in one file at [/llms-full.txt](https://elorati.com/llms-full.txt), Atom feed at [/feed.xml](https://elorati.com/feed.xml), sitemap at [/sitemap.xml](https://elorati.com/sitemap.xml). The Lararium subsite (open-source personal agentic stack) keeps its own corpus at [/lararium/llms-full.txt](https://elorati.com/lararium/llms-full.txt). The deslop skill (open-source AI-tell editor for prose) keeps its rulebook at [/deslop/llms-full.txt](https://elorati.com/deslop/llms-full.txt), and its page is [/deslop/](https://elorati.com/deslop/).

## Articles

- [What Actually Happens to Your Data When You Use an AI Tool](https://elorati.com/blog/what-happens-to-your-data-in-ai-tools/article.md): The question is rarely whether AI is safe. It is which account the data went through, what the vendor's terms actually say, and whether your own permissions were ever correct. A plain guide for operators, with the policies quoted.
- [Who Owns Your Software When Someone Else Builds It](https://elorati.com/blog/who-owns-your-custom-software/article.md): Paying for custom software does not automatically make it yours, and the code is the smallest part of what you need anyway. A plain checklist of what to own before, during, and after a build.
- [What an AI System Costs After It Ships](https://elorati.com/blog/ai-system-cost-after-launch/article.md): The build is the cheap part. A plain accounting of what it takes to keep an AI system running: model changes, prompt drift, API bills, monitoring, and the person who answers when it breaks.
- [Why Your Company's AI Brain Should Not Run in Real Time](https://elorati.com/blog/async-spine-sync-edges/article.md): Real-time agent organizations demo well and fail at production volume. The architecture that works is an async spine with sync edges: batch synthesis, instant answers, human gates.
- [What an 'AI Agent' Actually Is, and What It Replaces](https://elorati.com/blog/what-is-an-ai-agent/article.md): Every vendor is selling an 'AI agent' right now, and most of them mean something narrower than what gets pitched. A plain explanation of what an agent actually is, what it can replace, and what to check before you buy one.
- [How to Tell If a Task Is Worth Automating](https://elorati.com/blog/when-to-automate-a-task/article.md): Plenty of automation costs more than the work it replaces. A plain test for deciding which tasks earn the build, which ones are cheaper left alone, and why frequency matters less than most people think.
- [How to Run an AI Pilot That Actually Reaches Production](https://elorati.com/blog/running-an-ai-pilot/article.md): Most AI pilots impress everyone in the demo and then quietly die. The reason is almost never the model. A plain guide to scoping, running, and graduating a pilot so the work survives contact with real operations.
- [When to Build AI and When to Just Buy It](https://elorati.com/blog/build-vs-buy-ai/article.md): Most operators reach for a custom build when a tool would do, or buy a tool when only a build will hold. A plain framework for deciding which problems deserve a build, which deserve a license, and which deserve neither.
- [Agent Retention Technology: Tools That Actually Keep Agents](https://elorati.com/blog/agent-retention-technology/article.md): Agent turnover costs $5K–$10K per head. The right technology holds agents in seat, when implementation is taken seriously. What works, and what doesn't.
- [Data-Driven Brokerage Management: Beyond the Spreadsheet](https://elorati.com/blog/data-driven-brokerage-management/article.md): 94% of business spreadsheets contain errors. How brokerages replace gut decisions with real metrics on agent productivity, pipeline health, and profitability.
- [Why Most Real Estate Tech Fails (And What Actually Works)](https://elorati.com/blog/why-real-estate-tech-fails/article.md): CRM failure rates top 50%. Nearly half of agents rarely touch their tools. Why brokerage technology adoption fails, and what actually drives results.
- [The Tech Stack Every Brokerage Needs in 2026](https://elorati.com/blog/real-estate-brokerage-tech-stack-2026/article.md): Seven technology categories every brokerage needs in 2026. Platform comparisons, real costs, and what the integrations actually require.
- [How AI Is Changing Real Estate Operations](https://elorati.com/blog/ai-real-estate-operations/article.md): 97% of brokerages now use AI. Where it's moving real-estate operations (transaction coordination, agent onboarding, compliance review) and where it isn't.