AI Analytics Assistant — Self-Service Dashboards

Overview

The AI Analytics Assistant is an AI-powered tool built into the Fr8Labs Analytics Portal that lets you create custom dashboards, modify existing reports, and drill down into your data — all by typing questions in plain English. No SQL knowledge required.

It opens as a conversational workspace with conversation history, the current chat, and a Generated query panel for SQL and result previews. You can hide the conversations or query panel to make more room.

What changed: Previously, creating a new dashboard or modifying an existing report required submitting a request to the Fr8Labs team and waiting days for delivery. The editor provides a self-service workflow; query generation and result availability depend on the requested data and your configuration.

Video Walkthrough

See the AI Analytics Assistant in action — from opening the Dashboard Editor to asking a plain-English question and reviewing a generated chart.

AI Analytics Assistant.mp4

The walkthrough covers the conversational workspace layout, the ⚡ Fast / 🔴 Live data-source toggle, and a live demo of asking "Show me total shipment count by month for this year as a line chart" — watch the AI write the query and return the results on screen. For step-by-step details, see How It Works below.

When to Use It

The AI Analytics Assistant is designed for moments when you need answers fast:

  • Board meetings — A question comes up about exposure to a specific trade lane. Instead of noting it down and waiting a week for your ops team to compile the data, you can create the report on the spot.

  • Market disruptions — Straits of Hormuz escalation, port congestion at Klang, Red Sea diversions. Filter your shipment data by affected ports or trade lanes and review the resulting exposure report.

  • Customer reviews — Preparing for a QBR? Ask the AI to build a customer-specific dashboard showing volume trends, on-time performance, and revenue over the last 4 quarters.

  • Finance deep-dives — Unexpected margin drop? Ask the AI to break down costs by carrier, lane, or shipment mode for the affected period.

⚡ Fast vs 🔴 Live — Choosing Your Data Source

When you ask a question, a toggle at the top of the chat lets you choose how your data is sourced:

  • ⚡ Fast (recommended for most questions) — answers come from a high-speed analytics copy of your data that is refreshed periodically. Using this copy reduces load on the live operational system; query time depends on the report. A freshness indicator next to the toggle shows exactly how current this copy is (for example, "updated 3 hours ago").

  • 🔴 Live — answers come straight from your live operational system, up to the minute. Reach for this when a booking made an hour ago, or a status that changed this morning, has to be reflected. Live queries can be a little slower on large date ranges.

Start with ⚡ Fast for historical analysis and check the freshness indicator. If the needed data is unavailable or too old, use 🔴 Live where supported. Read any source-change message from the assistant; a source switch does not guarantee that the requested data is available.

Real-World Example: Board Meeting Scenario

The situation: You're in a board meeting and a director asks: "What's our exposure to shipments transiting the Straits of Hormuz? How many active shipments, which customers, and what's the total revenue at risk?"

The old way (1-2 weeks):

  1. Email ops team requesting the data

  2. Ops manually queries the analytics tool or exports from FMS

  3. Finance cross-references with invoice data

  4. Someone compiles it into a spreadsheet

  5. Results presented at the next meeting

Using the assistant:

  1. Open the Analytics Portal on your laptop

  2. Go to the Dashboard Editor

  3. In the AI chat, type: "Show me all active shipments where the port of loading or discharge is in the Straits of Hormuz region — include customer name, shipment count, and total revenue"

  4. The assistant writes the query live on screen, runs it, and shows you the results in a preview

  5. Share your screen — the board has their answer before the agenda item closes

AI Chat Panel

AI Chat Panel


How It Works

Step 1: Open the Dashboard Editor

From FMS, choose Dashboard → Editor. From the client portal, open Analytics and choose Editor where available. The editor initially shows a grid of dashboards and reports.

Step 2: Open the Assistant

Select Ask AI above the grid. The chat offers Create a dashboard, Create a report, Ask about a dashboard, and Ask about a report. Choosing Create a report starts a report prompt for you to complete.

Specify the metric, date basis, inclusive date range, statuses, and desired table or chart. For example: "Count shipment jobs by month of job created date, from 1 January through 31 August 2026, including all statuses, as a table."

Choose Fast or Live and check the freshness message beside the source controls. A stale indicator means you should review whether the data is recent enough for your question.

Step 3: Generate and review the actual result

The assistant may ask clarifying questions and propose a query card. Select Generate this query to generate SQL and populate the Generated query panel. The proposal also offers Edit and Dismiss.

Review the SQL, row count, result table, and source label. A conversational description of the intended result is not proof that the generated query follows it. Check:

  • The requested date field: job created date, shipment date, ETD, and invoice date answer different questions.

  • The actual first and last periods returned.

  • Whether cancelled jobs or other statuses were included as requested.

  • Whether the result counts jobs, house bills, invoices, or another entity.

  • The selected source and its freshness.

If the preview differs from your request, describe the discrepancy in a follow-up. Ask the assistant to verify that the requested field exists in the selected source. Do not accept a substitute date field without deciding whether it answers your question.

The preview controls include Run preview, Save to reports, and Schedule as report. Scheduling is a separate action; generating a preview does not schedule delivery.

Step 4: Save and check the saved report

Save only after the preview matches the intended scope. Select Save to reports, review the save form and destination, and complete the save. Then return to the reports list and reopen the saved result to check its name and contents. Access depends on the destination and the permissions assigned to your users; saving does not grant new access.

The FMS entry route, report prompt, Generate this query, and preview controls were checked as Admin DEMO in demo company #18 on 21 September 2026. Save/reopen and restricted-user behaviour remain separate verification checks.

What the AI Can Do

Capability

Example

Create new dashboard cards

"Create a chart showing DO status breakdown by week"

Modify existing queries

"Add a date range filter to this report"

Filter and drill down

"Filter this to only show Ocean Export shipments via Port Klang"

Change visualizations

"Change this from a table to a bar chart grouped by month"

Cross-card dashboard edits

"Add a company filter to all cards on this dashboard"

Create full dashboards

"Build me a 5-card dashboard for monitoring air freight operations"

Financial & payment-timing questions (when your accounting data is connected)

"What's our average days to collect (DSO) this year, days to pay suppliers (DPO), and which customers are slowest to pay?"

Choose speed vs freshness

Select ⚡ Fast for prepared data or 🔴 Live for operational data; check the actual source and freshness

What the AI Cannot Do

  • Access other companies' data — All queries are locked to your company. The analytics service is designed to enforce company scope on both ⚡ Fast and 🔴 Live.

  • Modify the database — The AI can only read data (SELECT queries). It cannot insert, update, or delete records.

  • Read arbitrary external systems — The assistant works with the data sources connected to your analytics environment. Accounting questions require connected accounting data; this does not give the assistant general access to every screen or action in your accounting system or email.

Tips for Best Results

  1. Be specific about what you want to see. Instead of "show me shipments", try "show me Ocean Export shipments from Port Klang to any US port in Q1 2026, grouped by customer".

  2. Mention the time period. Do not leave the date range implicit. Add "last 3 months", "this quarter", or "since January 2026".

  3. Ask for the visualization you want. Say "as a bar chart", "as a table", or "as a line chart over time" — otherwise the AI picks a default.

  4. Iterate. Start broad, then refine. "Show me revenue by customer" → "Now filter to only the top 20" → "Add a month-over-month comparison".

  5. Leave it on ⚡ Fast unless you need today's data. Check the freshness indicator; use 🔴 Live where supported when you need more recent data.

  6. Use the Marketplace. Before building from scratch, check the Marketplace tab — Fr8Labs maintains a library of ready-made dashboard templates you can add with one click.

Access Requirements

  • You need Dashboard Editor permission (not just viewer) to use the AI chat

  • Ask your company admin to check your dashboard access and categories. If editor access is unavailable in the settings they can manage, contact Fr8Labs support.

  • See the Analytics Portal — Getting Started guide for setup instructions

FAQ

Q: What's the difference between ⚡ Fast and 🔴 Live? A: ⚡ Fast reads from a high-speed copy of your data that is refreshed periodically — suitable for trends and reporting when its freshness meets your needs. 🔴 Live reads your operational system up-to-the-minute. Check the freshness indicator and select Live where supported when the Fast copy does not contain the data you need.

Q: Are there usage limits? A: Follow any limit or retry message shown in the editor. Contact your administrator or Fr8Labs support if a limit prevents your work.

Q: Can other people on my team see dashboards I create? A: Yes — dashboards are saved to your company's collection and visible to anyone with the appropriate category access.

Q: What if the AI generates wrong data? A: The AI shows you the query and a preview of the results before you save, so you can verify before committing. If something looks wrong, ask the AI to adjust or contact Fr8Labs support.

Q: Do I need to know SQL? A: No. The AI translates plain English into SQL. However, if you do know SQL, you can view and manually edit the generated queries in the editor.

Documentation owner: Fr8Labs Customer Success. Text reviewed: 21 September 2026.