Platform

One operating fabric for enterprise knowledge, AI, workflows, endpoints, and action.

EL-Ai-NE coordinates existing systems rather than replacing them. It provides the authority, context, orchestration, execution, verification, and evidence layer that makes AI operationally accountable.

How it works

Reasoning is separate from authority.

Models and agents reason and propose. A small set of named authorities accepts facts, grants permission, runs workflows, performs actions and records evidence. Nothing can grant itself a permission, accept its own fact or approve its own action.

  1. 01ConnectAdmitted, read-only connectors bring facts from the systems you already run. Every fact keeps its source, its rights and its age.
  2. 02AskA person or an application sends a request. Identity and policy are checked before anything is retrieved or any model runs.
  3. 03Answer locallyThe answer comes from a permitted cache or a local model, then an eligible regional model. An external provider is used only if that route has been admitted.
  4. 04Act under controlA change becomes a typed action: classified by consequence, approved by someone else, performed once and independently verified.
  5. 05Prove and measureAn evidence record and exact usage facts are written. If a source is withdrawn, everything derived from it stops being served.

The building blocks

Building blockWhat it doesWhere it runs
ManagerThe control plane: node inventory, configuration, policy distribution, dashboards and administration. It stays out of the request path.Control services
Site nodeEnforces policy, serves the cache, mediates AI requests, holds local evidence and queues work for the region.Each site
AI GatewayThe only way to reach a model. Every AI request passes the same eight stages.Site and region
Operational recordAccepted cross-system operational truth, with contradictions and unknowns kept visible. Your own systems remain the sources of record.Control services
Company Brain TrustGoverned knowledge and memory: stages, verifies, retrieves and tombstones evidence.Site, region and control
EL-FlowThe durable workflow engine. Work survives restarts, waits and disconnects.Site, region or control
Action FabricThe only admitted path for a change to another system: request, approval, execution and verification.Enforced locally
Governance ledgerSigned, append-only evidence and tombstones.Control services
StudioEvaluation, promotion, canary and suspension of models and other artifacts.Control services
Cost and OutcomeExact usage and cost facts, and accepted outcome measures.Control services

The path of a question

A request for an answer passes eight fixed AI Gateway stages. A denial, a missing authority or an unknown state stops the request at that stage. Nothing fails open.

  1. STAGE 1IdentityThe caller is authenticated. Tenant, site, role and purpose come from the verified session, never from what the caller types.
  2. STAGE 2Policy and data protectionCurrent policy decides whether the request is allowed. Sensitive content is masked, redacted or blocked.
  3. STAGE 3ClassificationThe kind of task is identified and the form of the response is chosen.
  4. STAGE 4Cache and memoryA previous answer is reused only if this caller may see it and nothing it depends on has been withdrawn.
  5. STAGE 5RetrievalA bounded set of evidence is compiled from authorized sources, with rights, freshness and lineage attached.
  6. STAGE 6ExecutionA model route is chosen: the site first, then an eligible region, then an approved external provider only if admitted.
  7. STAGE 7Response controlsThe output is checked for safety, citations and disclosure before it is delivered.
  8. STAGE 8Memory and governanceEvidence and usage are recorded. Writing anything to durable memory is a separate decision.

The path of an action

An action is anything that changes something outside EL-Ai-NE: opening a ticket, refreshing a device, changing a setting. Every action follows the same path, whoever or whatever proposed it.

  1. 01ProposeA workflow step or an agent plan produces a typed request naming the exact operation, target and parameters, chosen from an admitted catalog.
  2. 02AdmitThe target, purpose, policy, preconditions, budget and verification plan are checked, and the action is given a consequence class.
  3. 03ApproveWhere required, an approver who is not the requester approves that exact request.
  4. 04ExecuteA qualified connector receives a narrowly scoped, short-lived credential and performs the operation once.
  5. 05VerifyAn independent observer reads the resulting state. Only a confirmed end state counts as success.
  6. 06Record or reconcileReceipts join the evidence chain. If the acknowledgement is lost, the outcome is recorded as unknown and reconciled by reading the target.
Consequence classMeaningDefault handling
C0No material external effect.Proceeds within policy.
C1Low: easily reversible, low value, narrow scope.Follows the approved action lane.
C2Moderate: a meaningful effect with a bounded blast radius.Explicit human approval by default.
C3High: substantial financial, security, customer, compliance or availability impact.Explicit human approval by default, with strong authentication and separation of duty.
C4Critical: safety-critical, broadly irreversible, legally controlled or systemic.Explicit human approval by default.

The default ceiling for a deployment is C0. Raising it is an owner decision. The scale is an operating control and does not assign a legal risk category.

Six questions it can always answer

  1. What happened?Which request, decision, action or answer occurred.
  2. Who and under what policy?Which authenticated person or workload, tenant, site and policy version produced it.
  3. Which route, and why?Cache, local model, regional model or approved external provider.
  4. On what evidence?Which sources make the outcome safe enough for its intended use.
  5. At what cost?Against which baseline, and whether the result can be reproduced.
  6. Under which controls?Which approval, receipt, rollback, tombstone or retention rule governs it.

Platform authorities

Each material decision has one accepted owner.

Components may propose, cache, project, or display state. They cannot create a competing private truth or bypass the authority chain.

Identity and Policy

Resolve the principal, workload, tenant, site, purpose, delegation, data rights, route eligibility, consequence, approval, and budget.

Company Brain Trust

Compile permission-filtered retrieval, graph context, source assertions, approved memory, language, and workflow evidence.

EL-Flow and agents

Keep durable workflow progress, waits, retries, deadlines, approvals, compensation, and resumable multi-agent tasks outside model context.

Action Fabric

Admit typed mutations, issue short-lived credentials or endpoint leases, enforce single execution, verify the resulting state, and compensate when qualified.

Operational record and evidence

Preserve accepted operational assertions, contradictions, source lineage, governance receipts, and correction history.

Model and route control

Select qualified local, regional, or approved external models by quality, privacy, latency, capacity, reliability, and economics.

Network and endpoint truth

Join site, path, application, device, telemetry, posture, and source completeness with the business workflow.

Cost, outcome, and reporting

Keep exact cost and accepted business outcomes as separate facts, then publish reports with freshness, unknowns, lineage, and restatement state.

Agent and workflow operations

Agents exchange bounded tasks—not authority.

An EL-Ai-NE agent can interpret evidence, create a plan, call permitted models or tools, and propose an action. The durable task, permissions, side effects, and success state remain owned by the platform authorities.

TriggerOperator request, ticket, telemetry event, schedule, or API.
Agent taskIdentity-bound scope, deadline, budget, and consequence.
Context planSources, permissions, freshness, contradictions, and memory.
EL-FlowDurable steps, approvals, retries, and compensation.
Action requestTyped target, parameters, preconditions, and expected results.
Verified outcomeIndependent state observation, evidence, cost, and result.

Governance and assurance plane

Organizational governance becomes operational evidence—not detached paperwork.

Eleven bounded governance services connect inventory, impact, oversight, vendor governance, transparency, monitoring, redress, frameworks, audit, and resource impact to the existing technical authority chain.

Inventory and applicability

The inventory and applicability services establish what exists, who owns it, where it operates, which roles and obligations may apply, and when reassessment is required.

Impact and human control

The impact and oversight services record affected populations, mitigations, residual risk, intervention design, AI literacy, competence, and oversight effectiveness.

Transparency and post-market

The transparency and post-market services govern system cards, instructions, notices, monitoring, incidents, complaints, challenge, remedy, appeal, and corrections.

Evidence and improvement

The framework, audit and resource-impact services maintain crosswalks, audit, corrective action, management review, and material resource-impact evidence.

Model strategy

Current truth is retrieved. Stable behavior may be learned.

EL-Ai-NE uses deterministic logic and typed APIs first, permission-aware retrieval for current company knowledge, qualified models for reasoning, and model adaptation only when a stable behavioral gap justifies it.

1. Deterministic

Rules, schemas, validation, calculations, and APIs remain deterministic.

2. Retrieval and context

Current facts, policies, inventory, cases, and source evidence remain externally updateable.

3. Qualified reasoning

Use the smallest sufficient local, regional, or approved external model.

4. Controlled specialization

Train only with an explicit grant, immutable lineage, evaluation, canary, and rollback.

Structured Decision Intelligence Planned

A typed decision step inside the governed operating loop.

Choice and classification

Return a permitted candidate with source references and the qualified probability of that choice.

Boolean and ordinal judgments

Keep false, unknown and abstention distinct. Ordinal labels follow a registered rubric rather than an invented score.

Abstention and review

Missing evidence, unsupported inputs or stale qualification lead to explicit abstention, a permitted fallback or human review.

See the decision architecture

Capability status

What you can use now, what is switched on per deployment, and what is planned.

This site describes EL-Ai-NE as designed in Baseline v1.0.0. The labels below say where each capability stands in the baseline build.

Available now

Switched on by default in the current baseline build. You can see it working in a proof of concept.

Included, enabled per qualification

Built into the platform and shipped switched off. An accountable owner turns it on for a deployment once it is qualified for that deployment's profile.

Planned

Specified in Baseline v1.0.0 and on the delivery path. Not yet part of a deployable build.

CapabilityStatusNote
Governed core
Sign-in with local accounts and rolesAvailable nowEnterprise single sign-on is Planned.
Manager console: fleet, configuration, health and featuresAvailable now
Site node enrollmentAvailable nowA site node refuses requests until a policy authority is configured.
AI Gateway with fixed, policy-first stagesAvailable nowEvery AI request passes the same stages.
Operational record with contradictions and unknowns kept visibleAvailable now
Signed, append-only governance ledger and audit recordAvailable now
Erasure and tombstonesAvailable nowA tombstoned item is never served again.
Emergency stopAvailable nowAsserted and cleared by a verified person, with a reason.
Usage metering by request and routeAvailable nowMoney figures need a rate card.
Database-enforced tenant separation, consequence classes, workload identities, data classification, retention bundles, signed backupsIncluded, enabled per qualificationEach engages with an owner-signed decision.
Enterprise single sign-onPlannedSpecified as the production identity path.
AI serving and knowledge
Local model serving and chatAvailable nowOne local model server.
Answer cache and context budgetAvailable now
Knowledge searchAvailable nowLoading your own documents is not yet enabled.
Evaluation suite and model registryAvailable now
AI routing console: route policy, budgets and simulatorIncluded, enabled per qualification
Durable assistant memoryIncluded, enabled per qualification
Studio evaluation, canary and promotionIncluded, enabled per qualification
Local speechIncluded, enabled per qualification
External model providersPlannedOff in every profile at this baseline.
Structured Decision Intelligence: typed decisions with abstentionPlanned
Governed Web Experiences and Persistent Agent WorkspacePlanned
Knowledge Wiki and content fabricPlanned
Workflow and action
Agent and job runs with read-only tools; recommendationsAvailable nowReliability depends on the local model you choose.
Action Center: propose, approve, run and roll backIncluded, enabled per qualificationEvery action is refused until a signed action allowlist is installed.
Connector catalog, outbound webhooks, computer useIncluded, enabled per qualification
Write connectors, change verification and active probesPlanned
Network and endpoints
Network Truth: path, intent, exposure and health questionsIncluded, enabled per qualificationRead-only.
Endpoint estate and telemetryIncluded, enabled per qualificationPartial at this baseline.
Full endpoint client fleetPlanned
Trust, governance and reporting
Trust Center, evidence bundle and privacy registersIncluded, enabled per qualification
Assurance views and governed reportingIncluded, enabled per qualification
AI governance services: inventory, impact, oversight, transparency and post-market monitoringPlanned
Financial close and chargebackPlannedUsage metering is Available now.
Distributed operation
Cross-site state convergence and regional cellsPlanned
Disconnected operation with offline leasesPlannedNo offline authority exists until a lease duration is approved.
High-availability production topologyPlannedNeeds its own acceptance evidence.
External risk and assurance services
Assurance Exchange: trust portal and questionnairesPlanned
Supplier risk and external exposurePlanned
Shadow AI discovery and Shadow AI GuardPlanned
Installation
Installation controller and signed release bundleIncluded, enabled per qualificationPartial at this baseline.

Labels describe the Baseline v1.0.0 build, not a production approval. Production use of any capability depends on qualification evidence for the accepted release and deployment profile.

Ask EL-Ai-NEPublic website guide
Ask about capabilities, governance, privacy, Brain Trust, risk, deployment, pricing, comparisons, or public resources.