Security and privacy, in plain words.
Here is exactly what we do, who we rely on, and how your data is handled. Security reviews and questionnaires are welcome.
MaintenanceTech
How the product handles calls, data and money.
Every AI call says it's an AI — and that it's recorded.
It's the first sentence, it can't be talked over, and afterwards the system checks it was actually spoken.
Recordings have a retention limit.
Both sides of a call are recorded. Audio is kept up to 90 days and a company can choose less; emergency jobs and legal holds are kept longer. Transcripts and decisions are kept longer than audio.
Personal details are stripped from the record your team reads.
Names, numbers and addresses are redacted before the stored record is written. Our service providers process a live call only to run it.
Tenants can ask for their record — or ask for it to be erased.
Export and erasure requests are supported.
One company can't see another's data.
Every screen checks which company a person belongs to, and on many screens a second lock sits inside the stored records themselves.
The AI never moves money.
MaintenanceTech records quotes, costs and decisions. Anything at or over an owner's limit waits for a person, and money steps ask for a PIN or Face ID.
Everything can be paused.
Any job can be taken away from the AI, and an account admin can pause all automatic actions with one switch.
Every decision is on the record — and it's yours.
Decisions are saved with who made them and why, and a company can download its records at any time.
Access codes are never taken on a recorded line.
Entry codes and similar details are kept out of recorded calls by design.
A vendor who says stop, stops.
A do-not-call request is honoured across every company on the platform, not just the one that called.
Where MaintenanceTech runs depends on which offering you use.
MaintenanceTech for Owner-Operators runs as a hosted service. MaintenanceTech for Property Managers has an Enterprise Edition that can be self-hosted on infrastructure you control. Private AI systems we design and build for clients are a separate engagement again.
Who we rely on, and what we build.
Some workloads run on established AI and telecommunications providers, chosen per workload. Others run on models we build ourselves — including models built around a particular customer’s data and the way they work. Where an outside provider is involved, their security programs and certifications cover their own services, not ours. Details are available to customers and prospective customers on request.
The first model we’ve named is HueMann Mini, a realtime voice model in development.
This website
- AI Opportunity Assessment answers are processed to write your brief and not stored by us.
- Contact messages are delivered to our team by email and used only to reply.
- Details are in the privacy notice.
Send us your security questions.
We'll answer them directly. Email hello@k3yhold.com or use the contact form.