Academica's assistant reads the conversation, decides what a request needs, and hands the work to one specialized capability. Each capability owns a complete job from start to finish. The internal steps a capability runs stay out of the conversation.
conversationcapabilityevidenceanswer
One clear decision at each step
| Part | Handles | Does not handle |
|---|---|---|
| Assistant | The conversation, your intent, and which capability fits | Running the job itself |
| Capability | One complete job and its result | Writing the conversational reply |
| Internal step | One bounded judgment inside a job | Anything the reader sees directly |
How a request becomes an answer
your request → assistant reads intent → chosen capability → complete job → answer in the conversation
What Academica can do for you
| Capability | What it does |
|---|---|
| Data and entities | Answers questions about companies, drugs, devices, and other tracked entities. |
| Finance | Retrieves quotes and history, runs calculations, and builds charts. |
| Search | Expands a question, searches the sources, and returns ranked evidence. |
| Reports | Gathers evidence, writes each section, and assembles a full report. |
| Fact-checking | Extracts a claim, searches for evidence, and reports whether it holds. |
| App control | Inspects an open app and proposes actions inside it. |
| Charts | Turns data into a chart. |
Why the split matters
- Each job has one owner, so results stay consistent between requests.
- The assistant chooses the fitting capability by the whole job you asked for.
- One capability can draw on others when a job needs more than one kind of work.
- The evidence behind a report, chart, or answer stays traceable to its source.