About the Framework
What is BehavesWell?
The BehavesWell Framework is a governance standard for intelligent environments — the buildings, spaces, and systems where artificial intelligence mediates the relationship between people and place. It is maintained and published by HTBW LLC as founding steward.
Purpose
The BehavesWell Framework exists to ensure that intelligent environments serve the people who inhabit them. As AI becomes embedded in the fabric of built environments — managing comfort, security, access, and experience — the governance question becomes urgent: who sets the rules, who is accountable for them, and how do occupants know they are being respected?
The framework answers those questions with a structured, evidence-based governance standard. Six responsibilities, each defining an obligation that an intelligent environment must fulfill. A rigorous evidence base drawn from real environments. A transparent evolution process that updates the standard as evidence accumulates.
Six responsibilities
Framework Specification v1.0 organises the framework into six progressive responsibilities. Each defines an obligation that an intelligent environment must fulfill. Together, they establish the conditions that support Human Trust.
01
Truth
An intelligent environment represents itself and its state accurately — to its occupants, its caretakers, and its own records.
02
Identity
An intelligent environment knows who is present and acts in accordance with that knowledge — distinguishing between occupants, caretakers, and visitors.
03
Operational Trust
An intelligent environment behaves consistently and predictably — its actions are explainable, its failures are disclosed, and its behavior does not surprise those who depend on it.
04
Stewardship
An intelligent environment is maintained with care and accountability — its caretakers are identifiable, their responsibilities are defined, and the environment is not left without responsible oversight.
05
Continuity
An intelligent environment persists the preferences, history, and context of its occupants across time — it does not forget what it has been told, and it does not reset without consent.
06
Concierge
An intelligent environment actively serves its occupants — anticipating needs, reducing friction, and acting on their behalf within the boundaries they have established.
Grounded in evidence
The framework is not theoretical. Every responsibility is tested against real environments — residential, shared, and commercial — through a structured evidence programme. Evidence records are classified as Validated, Refined, Challenged, or Emerged. Challenged and Emerged findings drive the framework's evolution cycle, ensuring the standard remains current as intelligent environments develop. Framework research began in 2022 through the Homes That Behave Well residential research programme. Framework Specification v1.0 was published in September 2026.
Explore the evidence35
Changelog entries
6
Responsibilities
3
Research site types
1.0
Current version
Governance and stewardship
The BehavesWell Framework is maintained and published by HTBW LLC, which currently serves as founding steward. Governance decisions are documented through the published evidence and changelog process. The framework does not currently certify environments, accredit organizations, or operate as an independent standards body.
Future governance expansion — including advisory boards, reviewers, contributors, or additional governance participants — may be introduced through published governance revisions. The framework gains credibility through transparency: the specification, vocabulary, evidence methodology, and full changelog are publicly available.
Current governance
The BehavesWell Framework is currently maintained by HTBW LLC as founding steward. There is no independent governance board at this time. Future advisory or governance participants will be introduced through the published governance process.
Get in touch
For research collaboration, partnership enquiries, or media requests, contact the BehavesWell Framework team.
[email protected]