AI Assistants Overview

Two families of contextual AI Assistant — one built for acting on table rows, one built for interpreting dashboard metrics — both powered by the same Agentic writer-engine settings.

What is AI Assistant?

TotalApp gives almost every screen its own AI Assistant — a contextual panel that understands the data on that specific screen and can act on it or explain it in plain English. Rather than one generic chatbot bolted onto the app, there are two purpose-built families, each shaped for the kind of screen it lives on.

Every AI Assistant, regardless of family, is powered by the same underlying Agentic configuration (see Agentic → Overview) — the Writer Engine you choose in Settings (Ollama, In-Browser/WebGPU, Cloud LLM, or Local CLI) is what actually answers the assistant's requests. This is why every AI Assistant in TotalApp respects your chosen engine instead of silently calling a cloud API: the client-side code checks the active Writer Engine before making any request, so switching to a fully local or in-browser engine keeps every assistant working without a single behavior change.

The core split

Table/List screens get an assistant built for row-level action. Dashboard screens get an assistant built for metric-level insight. Same underlying AI, two very different jobs.

Table Assistants vs. Dashboard Assistants

Aspect	Table & List Assistants	Dashboard Assistants (Insight Generator)
Where it appears	Any table/list screen — Asset Registry, My Projects, Purchase Orders, Employees, and more	Chart-heavy dashboard screens — starting with the MTBF/MTTR Reliability Dashboard
What it does	Acts on the rows in front of you — selects, edits, cleans up, filters, summarizes, converts	Reads the charts for you — summarizes metrics, explains trends, recommends next steps
Example action	"Set status to Approved for these 5 records and create a work order"	"MTTR spiked in Q3 — driven mostly by the Boiler System H8 asset group"
Output shape	Bulk edits, filtered views, exported files, inline suggestions	Executive summary, root-cause explanation, actionable recommendation

Learn More

Table & List Assistants

Multi-select bulk actions, data cleanup, natural-language filtering, quick summarization, and inline smart suggestions on every table screen.

Dashboard Assistants (Insight Generator)

Natural-language executive summaries, root-cause and anomaly detection, and actionable recommendations for every chart-heavy dashboard.

Dashboard Assistants Are Arriving Now

Table/List Assistants already exist across many screens today. Dashboard Assistants — the Insight Generator — are the newest addition to this family, rolling out across TotalApp's chart-heavy dashboards starting with the MTBF/MTTR Reliability Dashboard in Maintenance Operations, with the Manufacturing OEE Dashboard and other analytics screens close behind.

Same concept, screen by screen

Because every Dashboard Assistant is scoped to its own dashboard's own screen and data, rolling it out to a new dashboard is a matter of wiring up that screen's metrics — not building a new AI concept from scratch.

Frequently Asked Questions

How do I know which type of AI Assistant a screen has?

Table and list screens (Asset Registry, My Projects, Purchase Orders, etc.) show a Table & List Assistant. Chart-heavy dashboard screens (starting with the MTBF/MTTR Reliability Dashboard) show a Dashboard Assistant. The panel itself is always labeled and scoped to the screen you're on — you never have to pick a type manually.

Do I need an API key to use AI Assistants?

No. AI Assistants work with whichever Writer Engine you've selected in Settings → Agentic — Ollama, In-Browser (ONNX/WebGPU), Cloud LLM, or Local CLI. If you choose a fully local or in-browser engine, no API key or cloud call is required at all.

Why does my AI Assistant behave differently after I change the Writer Engine?

Every AI Assistant checks the active Writer Engine before each request, so switching engines in Settings → Agentic changes response speed and style immediately — no separate per-screen configuration is needed.

Will every screen eventually get an AI Assistant?

Table/List Assistants already cover most table screens today. Dashboard Assistants are rolling out screen by screen, since each one is wired to that dashboard's specific metrics — new dashboards are added to the rollout regularly.

Is my data sent anywhere when I use an AI Assistant?

That depends entirely on your chosen Writer Engine. Local CLI and In-Browser engines process everything on your own machine; Cloud LLM sends the relevant screen data to the configured provider. See Agentic → Overview for details on each engine.