What runs, where information moves, which controls exist, what stays human, and what this site does not claim.
REVIEW RECORD
ACCOUNTABILITY BOUNDARY
Dissfam is the public and trading name released for this site. A separate incorporated entity or legal data-controller identity is not released by the current first-party claim registry. Public saved or transmitted Assessment modes stay disabled until that identity and the related privacy controls are approved.
Vercel serves the Next.js site. Assessment answers and calculations remain in browser memory and are not sent to a Dissfam API, model provider, lead system, or analytics event transport.
Boundary. Reloading, clearing, or closing the page removes the anonymous draft. There is no saved report in this mode.
The complete report, access, edit, export, share, revoke, log, expiry, code, and deletion implementation is deployed behind a server-only beta key. Production beta writes are restricted to unmistakable synthetic records in isolated aios_assessment_* MongoDB collections.
Boundary. The public Assessment has no save or email control. No real visitor answer or identity is accepted while the controller, appropriate sender, managed bot control, and public retention/deletion approval remain unavailable.
The Demo uses fictional bundled data and a deterministic reducer in the browser. It represents AI-supported work without calling a live customer system or model.
Boundary. No connection, customer operation, or realised-value claim is implied.
A visitor deliberately submits name, work email, company, role, intent, timing, budget, and problem summary to the same-origin lead endpoint. The endpoint validates and limits the payload, can write to a configured Notion database, and can forward to a configured webhook.
Boundary. Assessment answers are not added to this record. The current lead limiter is process-local and is not reused for saved reports.
Integrations and deployment controls are scoped per client. The public site contains no live business connection and the capability registry calls every named vendor capability Planned.
Boundary. Permissions, hosting, model use, retention, human gates, failures, and rollback must be agreed and verified for the actual installation.
Production is served over HTTPS by Vercel with HSTS, a per-request nonce Content Security Policy, clickjacking protection, content-type protection, and same-origin browser controls.
Public transport is encrypted with HTTPS. Anonymous Assessment answers are not stored. Private-beta email identity is application-encrypted with AES-256-GCM; access material is HMAC-signed and stored only as a keyed digest. No broader provider encryption claim is inferred.
Private reports use opaque IDs, signed bearer access, hashed 10-minute single-use codes, seven-day initial owner links, 24-hour independently revocable shares, and owner-visible content-free access logs. Public saved-report access is disabled.
Configured service credentials and report cryptographic keys are read from scoped server environment variables. No report key, access token, code, answer, email, or artifact text is logged.
Report APIs require JSON where applicable, strict payload limits, schema normalisation, origin and CSRF-style header checks, a honeypot, and MongoDB-backed shared rate limits. Public mode additionally requires a managed bot-control contract before it can enable.
The public Demo and Assessment keep consequential external sends, pricing, exceptions, and decisions behind visible human gates.
The anonymous Assessment asks only for calculation-changing ranges, priority, currency, and current system names. It asks for no name, email, company name, free-text brief, or uploaded file.
Anonymous Assessment state ends with Clear, reload, or tab close. Synthetic private-beta report records carry 30-day TTLs and owner deletion removes every report, answer, calculation, token, consent, and log record while retaining only a content-free deletion receipt. This beta default is not published as legal policy for real data. Contact retention remains separately unresolved.
A suspected issue can be reported to engineering@dissfam.com. The Assessment publication flag can close the route and an exact release can be reverted; no response-time SLA is claimed.
Production releases are bound to exact Git commits and Vercel deployments. Phase milestones and the pre-transformation baseline provide recoverable rollback points.
The public Demo and Assessment are deterministic browser software, not autonomous live agents. Client agent behaviour, model choice, permissions, fallbacks, and monitoring are installation-specific.
Public capability entries describe fail-closed stopping, preserved source state, named exception ownership, revocation, and rollback expectations. They are not proof of a tested client connection.
WORKLOAD
Occurrences × team-total or per-role effort. Team-total time is never multiplied by headcount.
CAPACITY
Baseline workload × addressable share × net time reduction × adoption. Capacity is never called cash saving.
GROSS PROFIT
Affected opportunities × deal value × absolute win-rate change × gross margin × coverage × adoption. It is withheld without the required evidence.
NEVER COMBINED
Pipeline exposure is not loss. Capacity is not cash. Released cash principal is not profit. Overlapping pains are not summed. The output intentionally has no oversized total.
Website hosting, request delivery, deployment, and the existing sitewide Insights script.
Conditional destination for the separate Contact form when configured. It receives no anonymous Assessment answer.
Isolated storage for synthetic private-beta report sessions, responses, calculations, reports, access material, consent, logs, deletion receipts, and shared rate limits. Memorial collections are not reused.
Existing infrastructure is configured for the separate memorial RSVP flow. It is not treated as an approved Opportunity Brief sender, and report delivery remains disabled.
Conditional Contact destination when configured. It receives no anonymous Assessment answer.
No model provider is called by the anonymous Assessment or illustrative Demo. Client installation model use is deployment-specific and must be disclosed before operation.
The Phase 4 assessment calculates in your browser, shows every assumption, and requires no email or call.