For alarm companies
Safe AI with Customer Data
A decision guide: which AI, for which data, and the five questions to ask before it touches a customer record
Your customers trust you with the code to their front door. Every major AI company now sells a plan that keeps your data out of model training. Most exposure comes from putting the wrong data into the wrong plan.
The data decides the tier, not the brand of AI. Start with sorting the data you have.
Start with your data
- Passcodes and verbal passwords
- Alarm, gate and lock codes
- Zone lists and premises layouts
- Open and close schedules
Stays in your system of record. Never pasted into any AI tool, at any price.
- Names, addresses and phone numbers
- Emergency contacts
- Billing and account details
- Interior video (through your video platform’s own AI only)
Only in a business plan with a data processing agreement, or inside your own software’s built-in AI.
- Procedures and install standards
- Marketing and proposals
- Training material
- Public information
Any business-plan AI tool is fine. Start experimenting here.
The full guide
Everything else is in the slides
Which AI plans protect your data, what to check in the software you already use, the five questions to ask before AI touches a customer record, and three things to do this month.
Keep going
More free tools for alarm companies
All free tools for alarm companies → · AI Transformation Diagnostic →
About the author
Naomi Withers is the founder of Growth Consultant Services, where she advises founders and leadership teams on AI strategy and organizational readiness. She spent more than a decade in finance and operations leadership inside a top SDM 100-ranked security company. This guide is free and vendor-neutral.
This guide is educational and is not legal advice. Vendor terms and prices change; each figure is dated, and you should confirm current terms with the vendor. Talk with qualified counsel about your specific obligations.
Sources
Retrieved September 22, 2026 unless noted.
AI company terms and pricing 1. OpenAI, "Data Controls FAQ," help.openai.com/en/articles/7730893; "What is ChatGPT Business," help.openai.com/en/articles/8792828; "Enterprise privacy," openai.com/enterprise-privacy. 2. Anthropic, "Updates to Our Consumer Terms," August 28, 2025, anthropic.com/news/updates-to-our-consumer-terms; "Commercial Terms of Service," anthropic.com/legal/commercial-terms; pricing, claude.com/pricing. 3. Google, "Gemini Apps Privacy Hub," support.google.com/gemini/answer/13594961; "Generative AI in Google Workspace Privacy Hub," knowledge.workspace.google.com; "Gemini API Additional Terms of Service," ai.google.dev/gemini-api/terms. 4. xAI, legal FAQ, privacy policy and pricing, x.ai; security FAQ, docs.x.ai. 5. Microsoft, "Enterprise Data Protection in Microsoft 365 Copilot," learn.microsoft.com; "Privacy FAQ for Microsoft Copilot," support.microsoft.com; Microsoft 365 Copilot pricing, microsoft.com. 6. OpenAI Help Center, "Sharing feedback, evaluation and fine-tuning data, and API inputs and outputs with OpenAI."
Ratings 7. Future of Life Institute, AI Safety Index, Summer 2026, July 2026, futureoflife.org/index. 8. Incogni, "Gen AI and LLM Data Privacy Ranking 2025," June 2025, blog.incogni.com. 9. Epoch AI, "Open vs. closed model capability gap," May 2026, and training-cost estimates, epoch.ai. 10. Hugging Face, Open LLM Leaderboard retirement notice, March 2025.
Software vendors 11. HubSpot Knowledge Base, "HubSpot AI model training," updated July 10, 2026. 12. ServiceTitan Help, "Titan Intelligence data preferences." 13. Salesforce Help, "Agentforce Trust: Data Masking."
Incidents and regulation 14. FINRA, "Salesloft Drift AI supply-chain attack," 2025. 15. TechCrunch, "Your Claude shared chats and Artifacts may have ended up on Google," July 27, 2026; Forbes and Malwarebytes on Grok chat indexing, August 2025. 16. Federal Trade Commission, "AI (and other) companies: Quietly changing your terms of service could be unfair or deceptive," February 13, 2024. 17. California Privacy Protection Agency, regulations announcement, September 23, 2025. 18. National Institute of Standards and Technology, AI Risk Management Framework and NIST AI 600-1; International Organization for Standardization, ISO/IEC 42001; Cybersecurity and Infrastructure Security Agency, "AI Data Security," May 22, 2025.
Cost and build-versus-buy 19. Compute Comparison, "Fine-Tuning Llama Cost Guide: 8B, 70B, and 405B," 2026. 20. Stanford HAI, AI Index; Epoch AI, "How much does it cost to train frontier AI models?" 21. Fortune, on MIT NANDA, "The GenAI Divide," August 18, 2025. 22. Gartner, press release, August 26, 2025. 23. Carlini et al., "Extracting Training Data from Large Language Models," USENIX Security Symposium, 2021.