Frequently asked questions

They serve executives, operations leaders, product teams and technical teams. The starting point depends on the role: prioritizing an investment, reducing repetitive work or preparing a system for production.

Start with a concrete problem: a slow task, information that is hard to find or a decision lacking useful support. Framing clarifies users, data, constraints and success criteria before choosing technology.

Compare expected value, feasibility, risks and data readiness. Framing can produce prioritized use cases and a roadmap; benefits remain hypotheses to validate, not achieved results.

Yes, a simpler non-AI option is preferable when it better fits the problem. Clarifying a process or using deterministic rules may be more appropriate than adding a model to an already well-defined task.

Neuralys provides AI strategy, workflow automation, custom development, technical advisory, training and responsible governance. Depending on the need, the work may focus on clarifying a decision, building a system or equipping a team.

Candidates include request triage, document processing, data entry and validation, summaries and administrative follow-up. Steps, exceptions and human approvals need to be defined for the selected process.

Custom development can connect existing software, APIs, databases and workflows. Connections, access rights and technical limits are reviewed during framing; this is not a promise of immediate compatibility with every tool.

Technical advisory is useful for reviewing architecture, evaluating a prototype, selecting tools or planning an integration. It can also help a team identify implementation risks and issues to address before production.

It tests the feasibility of an approach through a working prototype and evaluation. The result should inform whether to continue, adjust or stop, rather than present a demonstration as an operationally ready system.

Readiness includes data quality, evaluations, access, deployment, monitoring and maintenance. Responsibilities and human review must be explicit; a readiness review is not a guarantee against failures.

The guiding flow is Understand, Shape, Build, Transfer and Evolve. Scope, success criteria, constraints and the delivery model are clarified before implementation, with validation suited to the work.

Handoff includes documenting decisions, sharing knowledge and clarifying operational ownership. Maintenance and evolution need to be defined before operational use rather than left implicit.

Yes. Local AI, JUNO and Virtual Manager are available on demand and can be used separately. No solution is the parent or mandatory prerequisite of another.

Yes, they can optionally complement one another according to the need. For example, Local AI can serve as the inference engine for JUNO or Virtual Manager, but neither requires it.

Local AI uses models on your own infrastructure in an environment suited to your uses. Work includes model and infrastructure selection, access controls, data-flow review, evaluation and operational handoff.

Configurable uses include asking questions, summarizing documents, drafting, rephrasing and translating. They depend on the selected model and application, and important answers need checking before use.

No, local hosting alone cannot guarantee that. Connectors, external services, telemetry, backups and support arrangements must be reviewed and configured within the agreed scope.

JUNO is an enterprise operational digital twin: a virtual company of AI agents with defined tasks and permissions. It can assist internal operations and repetitive processes within agreed action and approval boundaries, without unrestricted financial or management authority.

It organizes agreed information, procedures and operational documents to support navigation and sourced answers. It respects access rights and does not imply automatically ingesting all company information.

JUNO can prepare and issue invoices, integrate the client’s accounting software, reconcile receipts and send reminders under agreed rules. Payments use the client’s tools; Neuralys does not take custody of funds.

JUNO can monitor agreed technology, legislative and funding-program sources and prepare updates for review. This monitoring is not exhaustive and does not guarantee legal compliance or funding eligibility.

It gathers useful information, prepares meetings and documents, and helps track priorities and commitments. It supports decision preparation by distinguishing facts, assumptions and information gaps; people retain decisions.

It can prefill forms and grant or financing applications from authorized information. Missing information is flagged, not invented. Human review is required before submission or signature; autonomous submission or signing is not implied.

People retain management decisions and final commitments. Connected tools, permitted actions and approval points are explicitly defined for each engagement rather than granting agents general autonomy.

The catalog contains six courses with 22 modules, grouped as 5, 4, 4, 3, 3 and 3 modules. It covers the general public, managers, AI opportunities, internal policy, agentic programming and cybersecurity in the age of agents.

No, each course is available on demand, individually or in a custom series, without a mandatory pathway. Audience, objectives, modules and exercises are agreed for the engagement; the catalog does not establish a separate sales rule for each module.

The general-public course covers requests, verification, limitations, privacy and everyday uses. The management course addresses risks, oversight, management uses and value assessment. AI opportunities explores HR, accounting, project management and administration.

Yes, it addresses permitted uses, responsibilities, professional confidentiality, data flows, access and retention. It helps formulate questions and rules for validation but does not replace legal advice.

Agentic programming covers work organization, version control, testing, code review and permissions. Cybersecurity addresses confidential data, prompt injection and agent isolation. Objectives include reviewing residual risks, not promising absolute security.

The catalog does not promise standard prices or durations, or certification. Commercial arrangements, format and teaching scope are agreed for your audience and objectives. Described learning outcomes are teaching goals, not a guarantee of mastery or results.