How Nila's AI works
Constrained by design.
The Nila AI guide draws exclusively from Nila's own clinically reviewed content. It does not access the open internet. This is disclosed prominently on every chat screen.
Every piece of content the AI can draw from has been reviewed by our editorial team and tagged with an evidence grade (A, B, or C) against its primary source — study name, author, and year visible.
The AI remembers context across sessions so members don't have to repeat themselves. It does not use this context to sell products, make clinical recommendations, or push members toward any particular treatment.
When evidence is absent or limited for a population — trans menopause, neurodivergent women, racialized communities — the AI says so. It does not extrapolate from majority-population data to fill gaps.
How we grade evidence
Three grades, one standard.
Major clinical guidelines (NAMS, NICE, SOGC, BMS), large randomised controlled trials, or systematic reviews. Safe to act on with a clinician.
Solid observational studies, smaller RCTs, or consistent expert consensus. Worth considering with appropriate nuance.
Early evidence, small studies, mixed results, or expert opinion only. Useful context — not a basis for clinical decisions without specialist input.
Every clinical claim on Nila carries one of these grades, tagged to its source. Users see the grade, not just a confident sentence.
Our limits — and why they matter
What we don't do.
We do not diagnose.
Nila does not identify conditions or tell members what they have.
We do not prescribe.
No treatment recommendations, no dosage guidance, no clinical decisions.
We do not replace a clinician.
We prepare members for better clinical encounters — we don't substitute for them.
We do not extrapolate.
When the evidence base has gaps for specific populations, we name the gap rather than filling it with assumptions.
Editorial independence
Independence, not neutrality.
Nila's clinical content is written and reviewed by our editorial team and is not influenced by sponsors, advertisers, or commercial partners. Sponsored content is clearly labelled — always. Counselling, crisis, identity, and treatment pages carry no sponsorship, ever. Practitioners are listed in our directory based on verified menopause training — not on whether they pay for placement.
Data
Privacy by design.
- Member data is never sold or shared with third parties.
- Employer and plan administrator accounts receive only anonymized aggregate engagement data — never individual symptoms, activity, or clinical information.
- Members can export or delete their data at any time.
Questions about our methodology? We're happy to walk through it.