Positioning

Questions this doc answers

What is the core hook for enterprise buyers?
What language should we avoid?
How much competitive detail belongs in public materials?

Messaging spine

Pain: Organizational AI context should not live only in Slack threads and forgotten chat sessions.
Consequence: Teams lose decisions, customer context, and compliance knowledge every time the tool resets.
Promise: Reflect is a persistent organizational memory layer — the AI tools you already use read and write the same memory under your controls.
Proof points (public): Private deploy options, Ambient Mode (toggleable), explicit visibility, auditability. Keep competitor teardown detail out of the hero and out of public architecture docs.

Ambient Mode & async diligence

Ambient Mode (default On) reduces friction: relevant memories are available in-session, and durable facts can be captured without a manual “save this” step. Toggle Off restores Suggested Memories (approval-required). Phase 1 ambient writes stay personal-only; org/team sharing remains explicit.

The async diligence bundle (this hub + markdown/PDF downloads + copy-paste prompts) lets prospects evaluate with their own AI before scheduling another call.

What not to say

Avoid calling Reflect "an AI memory tool" (app-layer semantics).
Don't lead with integration laundry lists; lead with repeated re-explanation cost.
Don't imply Ambient auto-shares to the whole org — Phase 1 is personal-only until shared.
Don't claim every write is still manual when Ambient is On; say Ambient is toggleable and audited.
Keep competitive comparisons high-level in public materials; deeper comps go under NDA.
No named companies in public or distributable material — use anonymized descriptors.
No dollar-figure pipeline claims unless a signed LOI or contract backs them.
Exploratory conversations are not pilots — only call it a pilot after a formal agreement.
Do not reference consumer PaaS brands as Reflect production hosting; Reflect cloud runs on Reflect-operated dedicated infrastructure.
Do not publish secret sauce: schemas, algorithms, internal topology, env dumps, or “how we beat X” internals on public pages.