The assistant is the architecture.
Leysco Licia is an ERP AI assistant that queries data and drafts actions through the same typed services as the screen. Every action is entitlement-checked and audited. Early access.
AI-native ERP assistant / governed agent platform
SAP Joule · Microsoft Copilot (D365) · Salesforce Agentforce
They bolt an assistant onto a finished suite. Here the assistant and the screen call the same governed services, under one permission model for humans and AI.
propose_journal_entry → DRAFT
resolve_master_data
ingest_document
… typed tools
Every AI action is a governed tool call.
The gatewayAn MCP tool server exposes the platform as typed, permissioned tools — to Licia and to any external agent. The propose-to-draft pattern: propose_journal_entry returns a DRAFT. The agent proposes; the platform decides.
| With journal checks switched on, re-read each journal draft after it is written and confirm its total and that debits equal credits. |
| Open a draft's evidence record: the checks it passed, who asked for it, what changed and its place in the audit chain. |
Inspect the checks behind the draft.
Journal draft checksSwitch on journal checks, and each journal draft from Licia is checked before it is written, re-read afterwards and confirmed to balance. If the totals do not agree, it stops for a person to review.
Audited like a ledger, because it writes to one.
Trust planePII routing, injection defense, output-integrity checks, cost ceilings — and a hash-chain audit of every AI action. The same discipline the ledger gets, applied to the model.
| Provider | Routing | Tier |
|---|---|---|
| Claude | tiered by task | cloud |
| Ollama | self-host default | $0 local |
Your ledger never has to leave the building.
Model freedomClaude tiers routed by task — or local Ollama as a $0 self-host default. Provider-agnostic by architecture, so model choice is a configuration row, not a procurement cycle.
| Search only your company's published documents, keeping internal documents to the people permitted to see them. |
| Quarantine documents that carry hidden instructions aimed at the AI, and drop any such passage before it reaches an answer. |
| Reject any answer that cites a passage Licia did not retrieve. |
| Return a clear no-answer response when retrieval or citation checks fail. |
Answers with their sources attached.
Governed knowledgeLicia answers from your own published documents and shows the source behind every answer it gives from them.
The platform under Licia.
In this line| Document ingestion (email/WhatsApp/PDF → governed draft) | Early access | Receipt in, postable draft out. |
| Master Data Resolver | Early access | Money-safe fuzzy entity resolution. |
| Semantic Layer — governed metrics | Early access | One measure, three renderings. |
| MCP Tool Server & agent gateway | Early access | The governed surface itself. |
| Schema engine (~375 schemas → database) | Early access | Configure-not-code substrate. |
| Authorization — one chain, humans + AI | Early access | RBAC/RLS/FLS/SoD incl. 12 ai.* permissions. |
| 3-tier multi-tenancy | Early access | Tenant-bound isolation and secrets. |
| Licia standalone — bring your own ERP | Early access | Same assistant against SAP B1, ByD, BC, NetSuite. |
| AI eval harness & injection-defense tooling | Designed | Quality gates for AI features. |
| Answers from your documents | Early access | Licia's sourced answers from your published documents. |
Watch an agent get governed.
Signature · Early access · Sample dataA deterministic replay of an MCP session: the agent lists its tools, asks a governed question, proposes a journal entry — and the platform answers with a gate, not an execution.
- agent→ tools/list
- akili← typed tools · scoped by role: semantic.query, propose_journal_entry, …
- agent→ semantic.query { measure: "expenses.total", period: "2026-05", by: "cost_center" }
- akili← 12 rows · governed model · RLS applied for this principal
- agent→ propose_journal_entry { debit: "6200-FUEL", credit: "1000-PETTYCASH", amount: 18500 }
- DOPSGATEjournal checks passed · balance asserted · read back OK → returned as DRAFT je-2026-000418, not posted
- akili← audit: action hash-chained · cost metered · awaiting human approval
The team behind Leysco Akili runs 1.66M+ posted invoices on Leysco100, its sibling house. Licia comes with that team and its domain depth.
Two independent, co-equal houses. Choose by fit. The native immutable multi-book general ledger is designed.
- want to shape an agent-operated ERP from the inside, as a design partner
- need AI that can act — under entitlements, gates and audit — not just chat
- have a data-residency posture that prefers a self-host model tier
- want to move early on AI-native ERP
A chat box beside the ERP, guessing at screens → a governed caller inside it
Part of the house.
RelatedMCP Gateway
The governed tool surface.
Early accessSemantic Layer
Governed metrics under every answer.
Early accessThe Leysco Akili house
The platform Licia is woven into.
Early accessDesigned by the team behind Leysco100.