Request a demo

Ask your operation anything.

Ask answers plain-English questions from your own systems (TMS, ELDs, spreadsheets) and shows its work: the tables it read, the filters it applied, the SQL one click away. Ask once, or pin the answer to a live dashboard.

Request a demo
Ask · thread Read-only
Which lanes pay us the least per mile?
Ask
Inspected columns of dbo.orderheader
Revenue per mile by lane, July deliveries 41 rows · 388ms

WITH orders AS (

SELECT DISTINCT ord_hdrnumber, ord_originpoint, ord_destpoint, ord_totalcharge, ord_totalmiles

FROM dbo.orderheader WHERE ord_completiondate >= @start_date AND ord_status = 'CMP')

SELECT TOP (50) …, SUM(ord_totalcharge) / NULLIF(SUM(ord_totalmiles), 0) AS rev_per_mile

BIS → FSD is your weakest lane — $0.38 under the $2.24 fleet average.

Revenue per mile vs fleet average — July Saved · Ops dashboard
BIS → FSD −$0.38/mi AMA → OKC −$0.26/mi GRR → CLE −$0.12/mi DAL → MEM +$0.15/mi ATL → CLT +$0.33/mi
read orderheader · legheader filters completed deliveries · July 41 rows fresh 6m ago View SQL refreshes every 5m
AI-generated query — verify before acting on it.
How are those lanes doing on time?
Ask Thinking…
On-time delivery by lane, July 412 rows · 241ms

BIS → FSD runs 71% on time — 23 points under your fleet.

On-time delivery — July Save to dashboard Saved · Ops dashboard
71% BIS → FSD 94% fleet average 1h 48m median late
read stops · orderheader derived arrival vs appointment, ±15 min 412 rows fresh 6m ago View SQL
How are those lanes doing on time? Ask a follow-up…
A looping sequence of the Ask thread over a read-only bridge to the fleet's own systems. The thread already holds a delivered answer: asked which lanes pay the least per mile, Ask read orderheader and legheader across 41 completed July deliveries and found BIS to FSD the weakest lane at $0.38 under the $2.24 fleet average, charted as bars diverging from that average, with AMA to OKC and GRR to CLE also below it and DAL to MEM and ATL to CLT above; that answer is saved to the Ops dashboard as a live tile. A follow-up then types into the composer — “How are those lanes doing on time?” — and Ask writes and runs a single T-SQL statement over dbo.stops joined to dbo.orderheader, comparing each delivery arrival against its appointment window, returning 412 rows in 241 milliseconds: BIS to FSD runs 71% on time against a 94% fleet average, with median lateness of one hour 48 minutes. Its receipts name the tables read, note that on-time is derived from arrival against appointment within a fifteen minute tolerance, and put the SQL one click away. That second answer is saved to the dashboard in turn.

How Ask earns trust.

  1. Ask in plain English

    Ask writes the query and runs it against your systems through a read-only bridge. Nobody on your team needs to write SQL, or wait for someone who can.

  2. Read the receipts

    Every answer shows its work in plain words: the tables it touched, the filters it applied, how many rows, how fresh the data. The raw SQL is one click away for anyone who wants it.

  3. Keep watch

    Monitors are questions that never sleep: described in plain English, re-run on a schedule, posting findings to Slack, Teams, or email with the same receipts attached.

  4. Send in the researchers

    Deep Research puts a fleet of agents on your data. They crawl the whole operation and come back with the money: idle-fuel burn, detention leakage, lanes that quietly stopped paying.

  5. Trust the no

    If your data can't support an answer (no cost data means no margin claim), Ask says so instead of guessing. An honest refusal beats a confident hallucination.

The freight ontology, pre-built.

Generic analytics tools spend their first months learning what your data means. Ask arrives already knowing, tuned to heavily customized TMS installs, because that is where it lives.

TMS TMW Suite · McLeod
  • orders & rates
  • customers & stops
  • dispatch legs
  • drivers & trucks
ELD Samsara · Motive
  • GPS pings & miles
  • hours of service
  • duty status
  • arrivals & departures
Accounting QuickBooks
  • invoices & payments
  • fuel spend
  • driver settlements
  • receivables aging
Spreadsheets the back office
  • rate sheets
  • fuel surcharge tables
  • driver pay rules
  • everything else that runs the fleet
The ontology
Read
  • Loads
  • Lanes
  • Stops
  • Drivers
  • Trucks
  • Revenue
Derived
  • Deadhead
  • Detention
  • On-time
  • Margin

No source stores a deadhead, detention, or margin column. Those numbers exist only where the systems join.

FAQ

How do we know the answers are right?

Every answer carries its receipts: the tables it read, the filters it applied, how many rows, how fresh the data — and the SQL one click away. If the data can't support an answer, Ask says so instead of guessing.

What if nobody on our team writes SQL?

That's the point. You ask in plain English; Ask writes and runs the query. Pin any answer to a live dashboard, or turn a question into a monitor that re-runs on a schedule and posts to Slack, Teams, or email.

Does it work with our systems?

Ask reads from the systems you already run — the TMS, ELDs, and spreadsheets — through a read-only bridge, and it's tuned for heavily customized TMS installs. Messy is the normal case.