Nova AIS Assistant
The copilot that speaks your ontology.
Nova AIS Assistant is the natural-language layer inside Nova AIS. Ask anything about your data, your workflows, or your agents, and it answers from your ontology, not a generic model. Then it can invoke the agent that does the work, and hand you a receipt you can audit.
Where it lives
Not a chatbot bolted onto a page. The language layer of the platform.
The Assistant sits on top of every Nova AIS workflow, built on the same ontology architecture that runs the rest of the platform. Builders and end users both reach the same data through it, without learning the platform first.
See the platform it lives inside, Nova AISOne copilot, two kinds of user
For the builder, and for the person at the desk.
A copilot that knows your platform
Ask in your data, and it can invoke the agent that does the work. No SQL, no training week.
Ontology queries
Ask which loans flagged for review this quarter, in plain language. It pulls from your objects, relationships, and events and returns rows you can act on.
Multi-agent invocation
It calls the phone, docs, collections, or lab agent on your behalf and returns the result inline. No SQL and no platform training required.
Workflow drafting
Describe a pipeline that scans new docs and flags failed checks, and it returns a runnable spec you can review before it ships.
An audit trail
Every question, every answer, and every agent call is logged with the model and prompt context, so you can show how any result was produced.
Skip the platform tour
An end user, a lab tech, a call-center agent, a planner, asks in plain Tagalog or English and gets an answer already in their data.
Plain-language input: ask for yesterday's critical results and it is done.
Code-switches the way Filipino users actually speak, mid-sentence.
Contextual to the surface: it knows whether it is looking at the dashboard or the lab.
A receipt for every answer: click to see the query and the model behind it.
What ships today
Already running, not a roadmap.
Each of these is live inside the platform now. The schools edition opens as a separate pilot in Q3 2026.
Knows your objects
It loads your ontology on the first request. Ask which patients had a critical result this week and it reads the actual Patient object, not a generic model's guess.
Invokes the right agent
Phone, docs, collections, lab. It routes the request to the agent that does the work and returns the result inline, in the same thread.
Speak data, get answers
Plain language to query to answer, across call transcripts, lab samples, collected accounts, and scanned documents. One query language for the whole platform.
Same copilot, every surface
It lives inside the dashboard, the lab view, and the platform itself. The context shifts with the surface; the engine underneath is one.
English, Tagalog, Cebuano
Native, not translated. It code-switches mid-sentence the way people actually speak, tested on call-center, lab, and planning workflows.
A receipt for every answer
Click any answer to see the underlying ontology query and the model that produced it. Logged for retention and review.
Ask the Assistant your own question.
Tell us the workflow you would want to ask about. We set up a sandboxed account, walk you through three real queries on your kind of data, and you decide if it earns a place in your stack.