Available
AI disclosure and support scope
Support agents are positioned as AI customer agents with explicit human handoff and scoped outcome definitions before public proof is used.
Zeiko SupportTrust and AI governance
Support-agent proof with clear publication boundaries
Zeiko Support separates internal QA, customer-facing claims, signed guarantees, and roadmap compliance so teams can launch support AI without overstating proof.
Available
Data handling boundaries
Support proof uses aggregate account-level rows, redacted samples, and approved customer-story fields instead of raw customer identifiers.
Available
Grounded-answer controls
Answers are checked against approved knowledge, imported helpdesk articles, support procedures, and runtime tool trace evidence.
Available
Human handoff and override
Risky support cases can create durable tickets, preserve transcript context, and route to operators before closure or public claims.
Available
Proof-pack publication gates
Public claims rely on Support QA, trust-pack controls, outcome-pricing audit rows, recommendation impact, and explicit account publication settings.
Roadmap
Compliance portfolio
External certifications and formal uptime SLA pages remain roadmap until published as signed trust artifacts.
Publication rules
Trust controls travel with every public support claim
The public benchmark program does not treat internal measurements as marketing claims until the proof pack and account publication settings agree on what can be shown.
1
Resolution and AI-resolution claims remain account and measurement-window specific.
2
Customer stories render only when account settings mark the story approved and provide an approval reference.
3
Guarantees render only from signed scope data with outcome definition, remedy, signer, and approval reference.
4
Outcome-priced usage is modeled from the persisted ledger and reported to Stripe only through explicit confirmation.
Trust artifact gates
Buyer-facing claims require named trust artifacts
Support QA can store account-scoped references, while this public page names the artifacts Zeiko requires before broader AI, retention, data-processing, data-residency, SLA, and compliance claims are buyer-ready.
AI disclosure page
AvailablePublic AI disclosure, human-handoff scope, and claim boundaries must be present before broad support-agent messaging is buyer-ready.
Provider retention policy
Account-scopedProvider-specific retention references must be saved or published per customer before Zeiko makes provider retention claims.
Signed provider-specific retention artifacts
Account-scopedSigned provider-specific retention artifacts must be attached before provider retention claims are treated as procurement-ready.
Data processing addendum
Account-scopedA DPA or approved policy reference must be attached before procurement-safe data-processing claims are treated as ready.
Data residency commitment
RoadmapData-residency commitments must name the supported regions, subprocessors, and customer-specific boundary before residency claims are buyer-ready.
Uptime and SLA evidence
RoadmapFormal uptime and SLA evidence stays caveated until a signed reporting surface is available for buyers.
Formal security certifications
RoadmapFormal security certification claims stay caveated until current certificates or auditor reports are published.
External governance artifacts
RoadmapExternal governance artifacts, audit reports, and security questionnaires remain roadmap until buyer-reviewable evidence exists.
Security compliance roadmap
RoadmapExternal certifications, data-residency commitments, and governance artifacts remain roadmap until real evidence is published.
Provider retention
Provider data boundaries are mapped before certification claims
This matrix names the main support-agent data processors and the current public-policy status. Formal certifications and uptime SLA artifacts remain separate signed evidence.
Zeiko support records
documentedPrimary account-scoped support system of record.
Stored as Zeiko account-scoped support data and governed by workspace lifecycle, export, and deletion policy.
support ticketsticket messagesticket events
Supabase
needs public policyDatabase, authentication, and account-scoped server storage.
Provider hosts Zeiko application data; buyer-facing retention wording must map Zeiko account deletion/export policy to hosted storage.
workspace identitysupport recordsagent memory
OpenAI
needs public policyModel calls for support reasoning, drafting, QA, and translation.
Runtime model calls receive bounded support context; public policy must describe model-provider data handling and training boundary.
prompt contextretrieved support snippetsdraft replies
Vercel
needs public policyApplication hosting, edge/runtime logs, and deployment delivery.
Provider may process request and runtime metadata for hosting; public policy must describe log retention and support-data redaction boundaries.
request metadataruntime logsdeployment logs
Stripe
needs public policyBilling, checkout, subscription state, and metered outcome events.
Billing events are sent only through confirmed billing flows; public policy must distinguish billing records from support transcripts.
billing identitysubscription statemetered usage events
Channel providers
customer configuredOptional customer-configured support delivery channels.
Twilio, Meta, Resend, Slack, Telegram, WhatsApp, and helpdesk retention depends on customer-enabled integrations and provider setup.
SMS messagessocial DMsemail replies
FAQ
Trust and governance questions
Does Zeiko publish broad support guarantees by default?
No. Public guarantees require a signed pilot scope, passing proof checks, a defined outcome, and a scoped remedy.
How does Zeiko keep support answers grounded?
Support agents use approved knowledge, tickets, imported articles, support procedures, runtime tool traces, and human handoff rules before customer-facing claims are published.
What remains roadmap instead of certified compliance?
External compliance certifications, formal uptime SLA reporting, data-residency commitments, and signed provider-specific retention artifacts remain roadmap or account-scoped gates until real evidence is published in this trust surface.