Getting started

Set some context, let the assistant propose a structured draft, then make the call. Here is the whole loop in five minutes.

What is Elicira?

Elicira is an assistant for business analysts. It does not replace your judgement: from the context you provide, it produces a structured draft (meeting preparation, user stories, process model) that you validate and refine.

The guiding principle: filtering > generation. The AI proposes a lot; you sort and decide, you stay in control of the deliverable at every step.

Open a conversation

Everything starts in the chat. On opening, four starters cover the most common uses: summarise meeting notes, clarify a need, write user stories, frame a process. Clicking one pre-fills the composer, and you take it from there. Nothing is sent until you send it.

Worth knowing: the “Write user stories” and “Frame a process” starters explicitly ask for a deliverable. Once completed and sent, the assistant may run the matching module right away, without waiting for step 4.

The chat home screen, with its four conversation starters.
The home screen: four starters, one composer, nothing else.

Set your context

Write or paste your starting point: a need, a spec excerpt, meeting notes. The assistant answers in your interface language, and what you write feeds the modules next, together with your attachments and your project context.

If your message already is a request for a deliverable, the assistant will not just answer: it runs the module. To stay in the conversation, set the context and ask it to reframe or question it.

A first question asked to the assistant, and its answer.
The assistant reframes and proposes an angle. It asks rather than asserts.

Attach your documents

The paperclip in the composer accepts .txt, .md, .docx and .pdf files: the app extracts their text to enrich the context. Nothing is sent without your action.

A document attached to the conversation, confirmed below the composer.
The document is attached to the conversation and will serve every module.

Run a module

The shortest way: ask for it in the chat. “Prepare my framing meeting” is enough, the assistant recognises the request, runs the module and saves the deliverable while showing you its progress. That path is described in how the assistant works.

The other way, more controlled: the three chips below the composer open the modules (Meeting prep, User stories, Process). The module appears in a panel on the right, without taking you away from the conversation, and makes you review before it produces, which is what the next two steps are about.

The Meeting preparation module panel, open to the right of the chat.
The module opens beside the chat. The conversation context stays in view.

Review and decide (module opened by hand)

This is the step that matters. Before producing anything, the module shows you what it understood from the conversation: the objective, the participants, the context. Correct it, complete it, remove what is wrong.

That stop belongs to the panel. When the assistant runs the module from the chat, it analyses and produces in one go: your review then happens on the deliverable, at the next step.

This checkpoint is not a formality: it is where you stop a context mistake from spreading through the whole deliverable.

The review form, pre-filled from the conversation, before generation.
The human checkpoint: the assistant shows what it understood, you correct it before generating.

Get the deliverable

The deliverable is produced from what you validated, or, if the assistant ran the module, from what it understood of the thread. For a meeting preparation: an agenda, elicitation questions grouped by theme and prioritised, the stakeholders to challenge, the blind spots and the points to watch.

The generated meeting preparation pack, with its sections.
The preparation pack. Every section stays editable.

Find your deliverables again

Every generated deliverable is kept in the library, reachable from the left menu. You can reopen it there, mark it as validated, duplicate it into a variant, or cross-check it against your other deliverables.

The deliverables library and its history.
The library: the history of everything you have produced.
The key idea: the AI proposes, you decide. No deliverable is frozen without your validation.

Organise: projects and context

Group your conversations, deliverables and context inside a project (the switcher at the top of the left menu). Enrich the project context (glossary, requirements, roles) for more accurate and consistent results across modules.

What’s next?

Go further with the chat, the modules and the deliverables library.