Analytics AI Assistant

The Analytics AI Assistant lets you explore your data by describing what you want in plain English. Instead of building a query field by field, you ask a question and the assistant drafts a query and a result you can then refine.

You can reach it via the assistant button in the bottom-right corner of a dashboard, which opens it alongside what you are viewing, scoped to that dashboard's data.

Asking a question

1. Open the assistant — from Analytics Assistant in the Analytics sidebar, or the assistant button in the bottom-right corner of a dashboard.
2. Type what you want to know in plain language, for example “total samples processed per project over the last 90 days”.
3. The assistant builds a query and shows the result.
4. Refine it — adjust the wording, or switch to the field picker to fine-tune the query by hand.

The assistant is a good starting point when you are exploring a dataset you do not know well. Once it has drafted a query, everything you can do with a point-and-click query applies, including charting, saving, and adding the result to a dashboard.

Skills: reusable instructions

A skill is a set of instructions you save once and reuse with the assistant, so you get the same steps, data choices and output format every time instead of rewriting the same prompt.

• Open Skills in the sidebar and click New skill, then choose Create with AI, Write your own or Upload a skill. You can also ask the assistant in chat to save a way of working as a skill.
• To run one, type / in the assistant chat and pick it, or type its identifier (for example /weekly-sample-summary) anywhere in your message. You can still add filters or follow-up questions.
• Skills you create are visible only to you.

For the full details, see Omni's documentation on Omni Agent skills.

Routines: scheduled and conditional updates

A routine runs a prompt for you automatically and sends you the assistant's written answer — for example a summary, an answer, or an explanation of what changed. Unlike a scheduled dashboard delivery, which resends the same file, a routine writes new text each time it runs.

• Scheduled — runs every time the schedule occurs, for example a Monday-morning summary.
• Conditional — checks a condition you describe in plain language and only runs when it is met, for example when inventory for a reagent falls below a threshold.

1. Open Routines in the sidebar and click New routine.
2. Enter a name and the instructions the assistant should follow each time. You can include a skill by typing its /identifier.
3. Choose the trigger: a schedule, or a condition to watch for. For a condition, test it before you can save.
4. Choose where the result goes (email or Slack, where enabled), then save.

A routine runs with your own access, so it can only report on data you can access. Each routine has one delivery destination, and results do not include links to dashboards.

For the full details, see Omni's documentation on AI Routines.

Skills and Routines depend on your permissions and on what is enabled for your organization. If you do not see them or cannot create them, contact Cenevo Support or your account representative.

Good to know

• The assistant answers from the curated topics (datasets) that are available to you in Analytics. It does not currently apply row-level Labguru data visibility rules (such as project or group membership) — see the note on data visibility.
• A drafted query is a normal query — continue working on it as described in Querying Data, and present it using the guidance in Visualize & Present.
• For the other way to build a query (point and click), see Querying Data.
• The assistant consumes AI credits from your package

Treat the assistant's output as a starting point. Always sanity-check a generated query against what you expect before relying on the result.