For technology and security leaders
For CTOs, CIOs, CISOs, CDOs, enterprise architects, and digital transformation leaders at REITs and large real estate companies who must co-approve AI work alongside the business operator. Direct answers, lifted from the same brief Quoin uses with client architecture and security reviews.
The six questions
No marketing answer first. Each question is answered with the operating posture and the controls behind it. If your review needs additional detail, the relevant artifact (decision packet, control matrix, eval suite, trace policy) is named in the conversation.
01
No. Discovery and the first scaffold run on metadata, redacted samples, walkthroughs, and synthetic examples. Production credentials cross only after the Step 18 handoff and a separate implementation approval signed by your security and data owners.
02
Yes. The Operating Intelligence Platform is deployed inside your cloud, your tenant, your identity provider, your network boundary, and your audit/logging infrastructure. We do not stand up a shadow system outside of your enterprise architecture.
03
Complements. We sit beside Snowflake, Databricks, Microsoft Fabric, Yardi, MRI, RealPage, Argus, Salesforce, Workday, ServiceNow, SharePoint, Box, Power BI, Tableau, and Excel. The platform is a governed semantic layer, not a new system of record. We do not replace your warehouse, ERP, PMS, BI, or document management system.
04
No, unless explicitly approved in writing. Default posture: no training, no fine-tuning, no prompt logs that retain proprietary data, no cross-client data mixing. Where you do approve training or fine-tuning, scope is documented in the implementation decision packet with retention, deletion, and revocation rules.
05
No. The platform uses model abstraction by design. The same agent capability can route across providers (Anthropic, OpenAI, open-source, on-prem) based on cost, latency, privacy, capability, or vendor preference. Model choice is a component decision, not a strategic one.
06
RBAC/ABAC. Scoped credentials for every agent and tool. Audit and trace for every retrieval, tool call, model call, draft, and approval. Lineage and freshness on every metric. Retention and deletion policies on prompts, traces, and retrieved content. Documented incident response, revocation paths, and rollback. Sensitive data redaction. Approved vendor and subprocessor list.
Step 18 hard gate
No production credentials, no live integrations, no agent deployment until two things are true: the Operating Diagnostic decision packet is approved, and a separate implementation approval naming scope, owners, environments, access limits, budget, and timebox is signed.
What separates Section 1 from Section 2
Section 1 (Diagnostic)
Pre-build. Metadata, redacted samples, synthetic data, walkthroughs. Outputs a decision packet and managed lifecycle object. Allowed to recommend not building.
Step 18 handoff
Decision packet, lifecycle object, and method-to-build handoff contract are reviewed and accepted.
Separate implementation approval
Names scope, owners, environments, access limits, budget, timebox, and security review path. Signed by sponsor + technology + security owners.
Section 2 (Implementation)
Operating Intelligence Platform first wedge built in your environment. Agents on top under controlled pilot. Production access added stage by stage.
Data access progression
Each stage carries its own purpose and its own controls. A platform wedge can stop at any stage if evidence does not support advancing. Access is earned by the work, not assumed at the start.
| Stage | Title | Required controls |
|---|---|---|
| 01 | Metadata scaffold | No production credentials. No bulk data. |
| 02 | Redacted samples | Written approval. Data minimization. Retention rule. |
| 03 | Client-controlled environment | Security review. Identity. Audit. Network. Logging. |
| 04 | Production operating layer | RBAC/ABAC. Monitoring. Lineage. Access review. Incident path. |
| 05 | Agent layer | Tool contracts. Guardrails. Traces. Eval thresholds. Revocation. |
Full architecture and stage detail lives on /platform.
CTO/CIO approval inventory
The decision packet routes the right approvals to the right owners. The technology side typically owns or co-owns the items below; your review process determines which require delegated governance and which require executive sign-off.
01
First domain or workflow
02
Approved implementation environment
03
Data access posture and whether data may leave source systems
04
Whether redacted samples can be used in discovery
05
Whether synthetic data is sufficient for the first prototype
06
Identity provider and access model
07
Logging and retention requirements
08
Approved vendors and subprocessors
09
Network and hosting constraints
10
Security review path
11
Source owner and data owner participation
12
Incident and revocation owner
13
Change-control forum
14
Production promotion criteria
Next step
30 minutes. Bring your operating model, your environment posture, your existing stack, and the security review path that any new AI work has to clear. Leave with a sketch of how the first wedge lands inside that posture.