Use Cases
Approval before sending (MAIR/NK)
Sensitive emails or texts don’t “just go out”; they pass through a brief approval process. MAIR/NK records who reviews, what was approved, and under which conditions.
Process: A draft is created (e.g., a customer reply). Based on tags/rules, the system detects that “SME” and “Legal” must be involved. Both receive a compact summary and can approve or request changes. Markers (ticket ID, source, version) are set automatically, and the decision is logged.
Outcome: Sending only after approval. Every step has a traceable record of who reviewed, what changed, and when approval was given. No chasing by phone, no grey areas.
Consistent Complaint Response (VPS)
Recurring complaints should be answered at a consistent quality level—without rewriting every time. VPS provides approved text blocks/prompts and auto-fills variables.
Process: Select the case type (e.g., “Damaged on delivery”), complete the required fields (order number, optional photo) → VPS loads approved blocks (tone, goodwill, legal clauses) via tags/versions, injects the case data, and generates the draft → optional approval via MAIR/NK → optional PII check via Protect Core → send; the versions used are recorded in VPS.
Outcome: Uniform tone, significantly less writing effort, and traceable block versions.
One Number — All Channels (Pitch Sync)
Figures and statements exist once in a central place and are referenced from there. When a value changes, email templates, slides, and website snippets update automatically.
Process: Update a KPI (e.g., MRR) with source and date. Pitch Sync identifies all artefacts where the value appears, updates those spots, and writes a change log. Texts tied to the number (claims, comparisons) are adjusted accordingly.
Outcome: No contradictions between deck, email, and website. Every number is traceable to source, timestamp, and approval.
PII Protection at Output (Protect Core)
Personal data is detected and handled before content goes out—not afterwards. This covers, among others, IBANs, phone numbers, names, and internal IDs.
Process: A text draft is checked automatically. Detected hits are masked or—depending on policy—blocked. If needed, the system requests approval from an authorised role (justification required). All decisions are logged with reasons.
Outcome: GDPR-clean outputs, traceable rejections, and a clear record of why something was redacted or released.
Market Data Normalization (Exchange Core)
Multiple data sources (APIs, files, news feeds) are harmonized and mapped to a common reference (e.g., ISIN). Outliers and gaps are surfaced instead of being silently “averaged away.”
Process: Feeds are fetched, cached, and mapped to the target schema. Plausibility checks (e.g., jumps > x%) are flagged. Ingests and versions are logged; the cleaned feed is made available internally—for queries, tests, or further steps.
Outcome: Reliable, versioned data instead of API chaos. It’s clear where each value came from and why it was accepted or discarded.
Care Notes & Handover (Doku Core)
Care and social teams document many entries every day. Doku Core structures notes (e.g., SIS/BI reference, measures, scales) and cleanly tracks versions, responsibilities, and handovers.
Process: Select client/patient and shift → record a short free text or events → Doku Core suggests phrasing/scales, checks required fields (measure, effect, specifics) → save the entry versioned → optionally generate a handover summary; flag items for follow-up.
Outcome: Consistent entries, handover export at shift change, and an auditable history (who/what/when).
Echo:me — Low-threshold support (no therapy)
Echo:me eases everyday strain: feel briefly understood, find orientation, choose one small next step. Not a diagnostic/therapy tool. Uses roles, approvals, and guardrails so nothing gets out of hand.
Process:
Start & consent: Notice “no therapy / not an emergency substitute,” obtain consent, optionally store quiet hours (e.g., 22:00–07:00) and an emergency contact.
Set roles (MAIR/NK): user, peer/support person, optional professional. Approvals control when something may be shared.
Check-in (very light): 1–2 sentences (“How are you?”) + a simple scale.
Echo: mirror in 2–3 sentences what was understood (“I hear A and B…”), name one gap, offer max. twosuggestions (e.g., short walk, breathing exercise, ping a peer).
Guardrails: crisis/alert keywords → immediate pointers to help options & “Would you like [emergency contact/peer/professional] to be notified?” (approval required). Emergency: always show the hint to dial 112.
Protection & traceability: Protect Core masks sensitive data; nothing is shared without consent. Mini-log: timestamp, chosen option, “shared yes/no” — nothing more.
Outcome: Noticeable relief from being briefly seen; small next steps instead of overload; a safe bridge to people/help; transparent control over data & sharing.
Disclaimer: Echo:me provides low-threshold support and does not replace therapy or emergency assistance (Emergency: 112).
