HugoScore hugoscore.org

Full review

Alcott CAIHL draft report

Evidence-linked HugoScore draft report for a health AI tool that affects patients.

Agency checklist · October 30, 2026

Leans expands 8 of 12 yes

Document-based draft. Independent confirmation is incomplete. Hugo has not personally verified this service.

Assessed function: Patient/family visit companion (visit.alcottai.com) with optional clinic delivery and EHR/CRM integrations. Assessed for the patient-chosen direct path.

A no can mean missing documentation or institution-dependent control; it does not by itself establish harmful behavior. These categories are not clinical safety or product-quality ratings.

  • Public documents do not establish actual product behavior.
  • The independent review screened rules and selected evidence; it did not replicate all twelve answers for this tool.
See all twelve answers and their evidence
QuestionAnswerEvidence / reviewSource and reasoning
1. Told it is AIYesdocumented
Not independently replicated
Visit Preparation: AI-guided conversations to help you organize symptoms, questions, and health concerns before appointments. Source

Terms and Privacy Policy describe AI-guided chat, AI-generated transcripts and summaries, and 'our AI' for records. Assessed for the patient/family direct app (visit.alcottai.com), not the clinic-only workflows.

2. Operator and model maker namedNonot-documented
Not independently replicated
Hosting, AI processing (Bedrock), transcription (Transcribe Medical) Source

Alcott, Inc. is named. AWS is listed for hosting and Bedrock AI processing, but no model developer is named (Bedrock hosts several vendors' models).

3. Patient chooses to use itYesdocumented
Not independently replicated
Try Alcott. It's free. Source

Patients and families can sign up directly. Clinics can also send pre-visit links (clinic-delivered path, see caveats). Direct path is patient-chosen.

4. Patient sets the taskYesdocumented
Not independently replicated
decide what you want to focus on before you show up Source

Patient page: user organizes history, uploads records and sets focus. Clinic-delivered intake would be shaped by the practice.

5. Patient can decline or stopYesdocumented
Not independently replicated
Try Alcott. It's free. Source

Self-started: Q3 is yes and no care depends on using Alcott. Source is the Q3 source. Terms 9 also allow closing the account at any time.

6. Patient sees the AI outputYesdocumented
Not independently replicated
Scribe: Record your medical visits and receive AI-generated transcripts and summaries. Source

Audio is deleted after transcription, but transcripts and summaries are kept and shown to the user.

7. Patient can correct or challengeNonot-documented
Not independently replicated
Quotation not recorded; evidence repair pending. Source

No route to correct or contest AI output is documented. Terms 6 say 'Alcott owns the software, algorithms, and AI-generated summaries', with the user holding a personal-use license.

8. Patient can get a copyYesdocumented
Not independently replicated
You can access, download, or delete your information and close your account anytime. Source

Privacy Policy section 7. Requests go to [email protected] and are answered within 30 days.

9. Patient can deleteYesdocumented
Not independently replicated
You can access, download, or delete your information and close your account anytime. Source

Closed accounts: health content deleted within 30 days. Anonymized metadata is kept indefinitely.

10. Training or sale does not override patient choiceNonot-documented
Not independently replicated
We do not sell your information or use health data for advertising. Source

No sale or advertising use is stated, but the policy is silent on model training and gives no opt-out. Anonymized metadata is used to 'improve our services'. AWS Bedrock processing is named, training terms are not.

11. Patient can run it or take it elsewhereNonot-documented
Not independently replicated
Quotation not recorded; evidence repair pending. Source

Vendor-hosted web app. Records can be downloaded, but no self-hosting or move-to-another-provider route is documented. Export alone is not enough.

12. No advertising or data businessYesdocumented
Not independently replicated
We do not sell your information or use health data for advertising. Source

Policy lists processors (AWS, Supabase, Twilio, Google, HubSpot, Keragon, Practice Better) and says it shares 'only with trusted service providers'. Provider BAA allows de-identification of PHI for provider accounts. Policy does not mention licensing de-identified data.

Initial coding: Claude Sonnet helpers, compiled by Claude Opus 5.5 (Anthropic). Selective review and corrections: GPT-6 Astra (OpenAI). Editor: Hugo Campos. Acceptance of corrections is not product verification.

Checklist rules · Independent review and limitations · Later evidence follow-up

Earlier report

Alcott: CAIHL reassessment

September 8, 2026. Published AI-assisted draft.

Alcott’s historical profile describes patient and family visit preparation with clinic delivery. Current pages could not be read in this reassessment. Patient approval of outgoing briefs, correction after delivery, and caregiver authority remain unresolved.

Scope and agency posture

Historically a hybrid patient/family visit companion with clinic delivery and EHR/CRM integrations. No current page body was recovered.

Posture: Potentially agency-expanding, mixed.

Axis: 72/100, retained provisionally. No numerical recalibration was performed.

Sources and documented findings

  • The July baseline documents patient agendas, visit summaries, family links, data rights and clinic integration, with incomplete patient review of outgoing briefs.

Patient authority

Inference from the documented workflow: Historical patient preparation and sharing support patient priorities. Patient approval of the exact clinician brief and caregiver authority remain distinct unresolved controls.

Critical capacity

Editorial assessment: Agendas and records could support reflection and action, but the current pass cannot verify correction of generated summaries or information already written to institutional systems.

Informed control

Editorial assessment: Historical named processors and deletion periods must keep their July date. The empty pages do not confirm or refute those commitments.

Assessment and limits

The specific historical hybrid assessment remains useful. Present-day uncertainty should be explicit, with no score deduction for failed extraction.

Confidence: Low for current behavior, moderate for frozen historical documentation.

Remaining uncertainty: All current public controls and changes since July remain unverified

Published documentation is evidence of stated conditions, not proof of actual implementation. Unknowns did not receive automatic negative points. Funding, sponsorship, and public code do not determine agency by themselves.

Review provenance

  • Reviewer/model: OpenAI Codex / GPT-6.
  • Method: Focused public-source reassessment using CAIHL: patient authority, critical capacity, and informed control. Existing evidence plus one focused primary-source pass and at most one targeted follow-up. No live product testing. Numerical scores remain provisional editorial placements, not a new calculation.
  • Human review: Hugo Campos authorized publication of these AI-assisted draft reassessments on September 8, 2026. This does not claim comprehensive human verification of every finding.
  • Earlier review: 2026-07-01. Historical assessment. Earlier claims are not automatically reverified by this publication.