--- name: al-396-decide-which-business-data-can-go-to-a-commercial-ai-and description: "Brokerages and other small firms now hold client, deal and financial data that agents want to feed to AI. On Oct 23, 2026 Inman reported that The Agency, a luxury real estate brokerage with more than 4,000 agents, is building an in-house AI platform that runs self-hosted open-source models alongside commercial models from companies like OpenAI and Anthropic. It pulls MLS and market data, transaction management, brokerage financials, listing syndication, marketing tools and internal communications into one conversational interface. Commercial models handle work like market analysis and forecasting, while sensitive client, transaction and financial data stays with models the company hosts itself. The platform is in alpha testing and is being rolled out in phases so problems are fixed before agents get it. You don't need your own servers to apply the same idea: the prompt below helps you sort your data and set simple rules for which AI tool may touch what. Use when the user wants help…" --- # Decide which business data can go to a commercial AI and which must stay on a model you control: sort your client, transaction and financial data by sensitivity and route each kind of task accordingly Brokerages and other small firms now hold client, deal and financial data that agents want to feed to AI. On Oct 23, 2026 Inman reported that The Agency, a luxury real estate brokerage with more than 4,000 agents, is building an in-house AI platform that runs self-hosted open-source models alongside commercial models from companies like OpenAI and Anthropic. It pulls MLS and market data, transaction management, brokerage financials, listing syndication, marketing tools and internal communications into one conversational interface. Commercial models handle work like market analysis and forecasting, while sensitive client, transaction and financial data stays with models the company hosts itself. The platform is in alpha testing and is being rolled out in phases so problems are fixed before agents get it. You don't need your own servers to apply the same idea: the prompt below helps you sort your data and set simple rules for which AI tool may touch what. ## Inputs to collect first Ask the user for anything below that they haven't already given you. Don't guess missing facts. - E.g. real estate team / brokerage / accounting practice / agency - People - E.g. ChatGPT, Claude, Gemini, Copilot, an AI built into our CRM, a… - , or suggest typical ones for my business ## Instructions Do the task below with the user's inputs in place of each [bracket]. Help me write a simple AI data policy for my business. Business: [e.g. real estate team / brokerage / accounting practice / agency], size: [people], AI tools we use or want to use: [e.g. ChatGPT, Claude, Gemini, Copilot, an AI built into our CRM, a local or self-hosted model]. Data we hold: [list: client contact details, ID documents, bank details, contracts and transaction files, commissions and company financials, MLS or market data, marketing content, internal messages]. 1) Sort each data type into three levels: Public (fine for any AI tool), Internal (only AI tools on a business plan with training turned off and a signed data agreement), and Restricted (never pasted into a commercial chatbot; only a tool we control or none). Explain each choice in one line. 2) For our 10 most common AI tasks [list, or suggest typical ones for my business], say which level of data each needs and which tool should do it, and how to do the task with less sensitive data, for example by removing names or using totals. 3) List the settings to check in each tool: data retention, training opt-out, who on the team has access, connectors and what they can read. 4) Write a one-page policy in plain language for the team, with five do's and five don'ts, and a short checklist for adding any new AI tool. 5) Suggest a phased rollout: who tries it first, what we watch for, and when everyone gets it. Mark anything that depends on local privacy law for me to confirm. ## Safety checks (share these with the user) - Anything legal, tax or compliance-related is a first draft for a qualified person to check. It is not professional advice, and clause references and deadlines must be checked against the source. - Remove names, contact details, account numbers and other personal data the task doesn't need before you paste, and check your AI tool's data and training settings. - Check names, dates, numbers and any claim against your own source material before you use the answer. AI can misread or invent details. - Say clearly which parts need a person to review or approve before anything is sent, paid, posted or filed. ## About this skill - Works with: ChatGPT, Claude, Gemini, Microsoft Copilot, self-hosted open-source models - Based on this real-world example: https://www.inman.com/2026/10/23/inside-the-agencys-in-house-ai-platform/ - Free prompt from AgentPlaybook Real Estate: https://agentic-library.netlify.app/prompt/al-396-decide-which-business-data-can-go-to-a-commercial-ai-and-whi/ - To install in an agent that supports Agent Skills, save this file as SKILL.md in a folder with the same name as `name` above (for example in your skills directory). Install steps for Claude.ai, Claude Code, Cursor and Codex: https://agentic-library.netlify.app/skills/