FogLifter® is designed so the same foundation that improves data visibility also protects it — from ingestion through NLQ. No LLM is trained on your data. No customer data leaves your governed environment. Access, lineage, and AI interactions are controlled and auditable at every step.
Security questions — especially around AI and data exposure — shape whether a platform makes it into the next stage of review. Here's how FogLifter answers them.
Control who can see, review, approve, and export different kinds of information — scoped by role and data domain. A finance user sees cost allocations without underlying infrastructure detail. An operations user sees asset health without provider contract terms. NLQ sessions inherit the same permissions — a conversational query can never surface data the user's role doesn't allow them to see in a dashboard.
Customer data is never used to train, fine-tune, or improve any language model. NLQ queries and results stay within the governed environment. Your data doesn't become part of a training set, doesn't flow to a third-party model provider for learning purposes, and doesn't leave the perimeter you control. FogLifter works with your preferred AI platform, following your existing enterprise security policies and data residency requirements.
Every output — dashboards, NLQ answers, exported reports — traces back to the source records that produced it. When a number is challenged, the evidence chain is already documented. This is how customers in regulated environments maintain 97.8% compliance year over year and achieve 99% security compliance across 1,100 critical applications: the proof doesn't need to be assembled after the fact — it's built into the platform.
Validation, investigation, and decision steps follow explicit paths with tracked approvals rather than opaque handoffs. The Validation View lets both parties — IT and service providers, IT and business units — comment, flag, agree, and resolve discrepancies on the same record. Every interaction is time-stamped, preserved, and available for compliance review.
Because FogLifter's ontology understands the relationships between assets, services, costs, and owners, access control is more precise than simple role-based permissions. The model itself enforces what's visible — so governance scales with the data, not against it.
This matters especially for the AI layer. NLQ doesn't send raw data to an external model for interpretation — the ontology provides the semantic context internally. The language model receives only the structured, permission-scoped query and returns answers grounded in validated records. Customer data stays within the governed perimeter. No raw records are exposed to model providers. No customer-specific information is retained beyond the session.
"The security question we hear most often is: will the AI learn from our data? The architecture answer is no — but the more important answer is that your data never needed to leave your environment in the first place. The ontology does the interpretation. The model just answers the question."
The most common security concern in enterprise AI is exposure — will a model learn from my data, retain it, or share it? FogLifter is architected to eliminate that risk.
Customer data is never used to train, fine-tune, or improve any language model — whether FogLifter's own or a third-party provider's. NLQ queries are processed, answered, and discarded. No customer-specific information is retained by the model beyond the session. Your IT asset data, financial records, service agreements, and provider details remain entirely within your control.
FogLifter integrates with the customer's chosen AI infrastructure rather than forcing a specific model provider. This means the AI layer follows your existing enterprise security policies, API governance, data residency requirements, and procurement approvals. You choose the model. FogLifter governs the data that feeds it and the answers that come back.
NLQ answers are generated from the ontology and validated data — not from unstructured web content or general-purpose model knowledge. Every answer traces back to a verified source record. If the data doesn't support the answer, the system says so rather than making one up. This is the difference between "AI-powered" and "AI grounded in evidence" — and it's what makes NLQ outputs defensible in regulated environments.
Every NLQ session inherits the user's role-based access controls. A finance user asking about cost can't inadvertently surface infrastructure security detail. An operations user asking about asset health can't access provider contract terms. The same governance model that controls dashboard visibility controls conversational access — there are no ungoverned side doors into the data.
FogLifter's governance layer is designed for regulated industries and enterprise procurement processes that require defensible evidence at every stage.
Every dashboard metric, NLQ answer, and exported report traces back through the correlation layer to the original source records. When a number is challenged, the evidence chain is already documented. This is how customers maintain 97.8% compliance year over year — the proof is built into the workflow, not assembled after the audit is announced.
FogLifter records what changed between each data refresh — new records, updated fields, resolved conflicts, and newly flagged mismatches. At one healthcare customer, this created a time-stamped audit history across 52,000 production servers, enabling 99% security compliance across 1,100 critical applications and patching management that compliance teams could verify independently.
Data exports, report generation, and NLQ responses follow configurable approval paths — so sensitive outputs don't leave the platform without the right review steps. This is built for the environments that need it most: healthcare organizations managing 70,000+ servers, financial institutions reconciling 52 currencies and 70 tax structures, and government agencies where every data movement must be justified and recorded.
Security and governance protect the foundation. The pages below cover what the platform does with the data it governs.
Cost, asset, ownership, and service analytics built on reconciled data.
How FogLifter connects to CMDB, billing, cloud, infrastructure, and ticketing systems.
Natural-language questions grounded in FogLifter's ontology and validated data.
