The nine questions
The same nine questions, answered twice
Numbers in brackets point to the numbered sources at the foot of the page. The line under each question says what a complete answer contains.
1 · Where must the data live for questions and agents to work?
A complete answer names the systems the data has to be copied into, and the systems that are read where they stand.
MicrosoftFabric brings data together in OneLake: shortcuts read Amazon S3, Google Cloud Storage and Dataverse in place, while mirroring copies Oracle, SAP Datasphere, Snowflake and SQL Server data into OneLake for Copilot and Fabric agents to use. Ontology bindings work only on managed lakehouse tables held there [1][2].
Lima JunoYour rows stay in the systems that hold them. Juno registers tables through the Lima data fabric and reads them in place, so a warehouse sits beside SAP, Oracle and planning data in one answer. 2 · How far does it reach into SAP, Oracle and planning systems?
A complete answer names the interface used for each system and says which of them are live today.
MicrosoftMirroring lands a copy of Oracle and SAP Datasphere data inside OneLake for Fabric agents to query; separately, a premium Copilot Studio connector calls SAP's own interface functions directly, reads tables, and supports commit and rollback through an on-premises gateway with each user's own sign-in [2][3].
Lima JunoMore than 40 connectors cover SAP ECC and S/4HANA, Oracle Fusion and E-Business Suite, planning suites, warehouse systems, lakes, databases, files and REST services. One question can join several of them, and the one agent in production, the PO cancellation agent, runs on federated ERP data the same way. 3 · How is the shared meaning, the semantic model or ontology, built, and by whom?
A complete answer says what arrives ready, what your team builds, and how long that modelling work takes.
MicrosoftA Power BI semantic model comes first, and Microsoft's guidance warns that Copilot may misread data left unprepared. Fabric IQ can generate a draft ontology automatically from that semantic model, though keys, time bindings and relationships still need a person's hand, and the ontology stayed in preview through mid-2026 [4][5].
Lima JunoLima’s supply chain ontology arrives filled in: materials, plants, suppliers, orders and stock. Juno proposes a mapping for each registered column with a confidence and a reason; your team confirms, adjusts and extends it. 4 · How do business users ask questions, and is the working shown?
A complete answer shows the question, the query that ran, the sources read and the access rules applied.
MicrosoftPeople ask questions inside Power BI or a Fabric data agent, which routes each one to the matching query language across up to five sources. Builders can inspect the generated query while tuning it, though everyday answers cap at 25 rows by 25 columns and stay in English at this stage [4][6].
Lima JunoPeople ask ChatLima in plain language, in its own screen, in Microsoft Teams or through the Power BI connector. Each answer travels with its working: the ontology terms, the systems read, the seconds taken and the query. 5 · How are agents and rules defined, and by whom?
A complete answer says who writes a rule, in which language, and how it is reviewed before it runs.
MicrosoftCopilot Studio lets people describe an agent in plain language on a low-code canvas, with templates and an Agent Store for common patterns; Microsoft Foundry serves coding teams building hosted, multi-agent workflows with more control over each step [7][8].
Lima JunoA planner writes the policy in plain language, and the Planner Agent turns it into a reviewable planning table of tasks, data and rules. The PO cancellation agent, built this way, runs in production; Lima’s agent blueprint designs 960 further roles across five chains, the other supply chain task agents among them, yet to be built. 6 · How is autonomy set, and how do approvals work?
A complete answer gives the levels available, who approves what at each one, and how a level is raised and lowered.
MicrosoftCopilot Studio applies data policies across knowledge sources, connectors and triggers, together with credit caps and environment routing; a graded, numbered autonomy setting is undocumented on the pages reviewed, and approval steps for a given agent action stay for the customer to configure [8][9].
Lima JunoEach agent carries a level you set, A0 to A4. At A2 a named person approves every action; at A3 the agent acts inside value, budget and confidence limits and sends exceptions to people. The level is one of eight settings on an agent, each set by a named person. 7 · How do actions reach the ERP, and is each write confirmed?
A complete answer names the interface for each ERP and says where the acknowledgement of a write is kept.
MicrosoftThe SAP ERP connector calls BAPI and RFC functions and supports commit and rollback through an on-premises gateway, so an approved change reaches SAP directly; the Fabric data agent stays read-only, and operations agents trigger Power Automate or Activator flows for any downstream action [3][10].
Lima JunoWrite-back uses each ERP’s own supported interfaces: released APIs and BAPIs for SAP, REST resources and interface tables for Oracle. It starts per application after a confirmed test write, and every write keeps the acknowledgement. The PO cancellation agent in production writes back this way, under human approval. 8 · How does it learn from people’s decisions, and what is audited?
A complete answer says what is recorded for each decision and how that record shapes the next recommendation.
MicrosoftMicrosoft Foundry lists a procedural-memory feature in private preview, aimed at learning across runs; Purview and Sentinel record audit logs, and Agent 365 adds a registry with analytics and lifecycle management for every deployed agent, generally available from May 2026 [10][11].
Lima JunoEvery accept, edit and override is stored with the rule version, the facts at that moment and the outcome, so the next recommendation improves. The audit trail holds signal, reasoning, limits, approval, action and acknowledgement for each run; in production, 100% of the cancellation agent’s actions trace to a rule. 9 · What is the pricing unit, is there a trial, and where do analysts place it?
A complete answer names the unit of charge, the route to try it, and the analyst reports that cover the category.
MicrosoftPower BI Pro costs $14 and Premium Per User $24 per user monthly, though Copilot itself needs a paid Fabric capacity metered in compute seconds; Copilot Studio sells credit packs at $200 for 25,000 credits, with a free Azure trial credit. Gartner names Microsoft a Leader in its 2026 Analytics and Business Intelligence and Conversational AI Platform quadrants [6][12][13].
Lima JunoJuno is licensed with the Lima platform and priced per engagement. The token manager budgets and meters every model call, per agent and per team. Twenty use cases are worked through on this site, one page each, and the analyst guide sets out the markets these products sit in.