Skip to content
Freight market analyst

Freight market analyst

Agent: the freight market analyst (freight_market), run from the Freight market card on Why (/simulate?tab=knowledge&card=freight). The old /market_diligence address redirects to that card (owner, 2026-10-31).

What it does

It benchmarks each of your lanes against that lane’s own history, and suggests re-bid, consolidation and fuel-surcharge reviews for your procurement team to decide. Its jobs:

  • Rate trend. The median linehaul rate per mile or kilometre over the last 90 days of a lane’s records, against the 275 days before. A lane above its own baseline by more than your target, with enough loads in both windows, is flagged. With no target set, the change is shown and nothing is flagged.
  • Seasonal comparison. The same 90 days a year earlier, so a seasonal peak is not mistaken for a trend.
  • Tender rejections. How many of your tenders on the lane were rejected.
  • Fuel surcharge against diesel. How your fuel surcharge per distance moved against the US EIA weekly diesel price over the same weeks.
  • Transit and accessorials. Mean transit days, on-time share against planned transit, and accessorial charges per load.
  • Candidates. Re-bid candidates (above own trend, or many rejections), consolidation candidates (thin lanes from one origin to one destination region), and fuel-surcharge reviews.

What it reads

Your own lane history: the loads and tenders your systems record, with origin, destination, dates, tender outcome, linehaul, fuel surcharge, accessorials, distance and its unit, and transit days. A connection is marked as lane history by the operator.

Two public sources, as context:

  • US EIA weekly retail on-highway diesel price (U.S. Energy Information Administration, public domain).
  • BTS Freight Transportation Services Index (U.S. Bureau of Transportation Statistics, public domain), shown as national context only.

What it produces

  • Lane benchmarks with every figure, the windows they cover and the number of loads behind them.
  • Flags with the figures that raised them.
  • Recommended actions, each naming the Transportation Procurement Manager as the person who decides.
  • The list of sources consulted, each marked live, cached, stale or not checked, with its retrieval time, and the rows of your records that could not be used and why.
  • On request, a short written summary. It phrases the figures above and adds none.

Human oversight

  • Every recommended action is a draft. Sending one for approval puts it in the approval queue, where a named person approves or rejects it.
  • Approving records the decision. FACE contacts no carrier, sends no tender and awards no freight.

Model

The figures and the recommended actions are fixed arithmetic over your records and the public sources. No model computes them.

The optional summary uses the general reasoning tier.

Where it runs and data handling

On the deployment’s own model plane: Runink’s shared machines on a Trial, your own cloud on Cloud and your own servers on On-premises. No outside AI service is called on any plan. The two public sources are fetched for every organisation, unless the deployment has turned live public data off. They are cached with their retrieval time; nothing from your records is sent to them. See Facts shared by every card.

Guardrails

  • The comparison is your own history plus public fuel data. It uses no proprietary rate index and refuses to make up a market or spot rate.
  • Refuses to steer a re-bid or consolidation towards a favoured carrier.
  • Recommends only. It does not book, tender, award or pay.
  • A written summary that states a figure the analysis did not compute, or proposes an action outside its allowed set, is not shown.

How to read its results

  • A lane is measured against its own history. A lane above its own trend is not proof it is above the market.
  • The BTS index covers all US freight. It is context, not a lane benchmark.
  • A fuel surcharge moving faster than diesel is a reason to check your contract’s fuel table, not proof of an error.
  • A row without a linehaul rate, a distance or a distance unit is not used, and is counted with the reason. Nothing is filled in.
  • With no network or no permission to reach the public sources, the lane figures still work; the fuel comparison is marked not checked.

Thresholds

There are no built-in thresholds. Each is a target set for a deployment, or for your organisation by an admin under Admin → Public registries & market data. Until a target is set, the figure it would judge is shown as measured, nothing is flagged on it, and the result names it as having no target. Every result shows the thresholds it applied.

ThresholdDefault
Lane above its own baseline by more thanNo target set
Tender rejection share that flags a laneNo target set
Fewest tenders before the rejection share countsNo minimum
Fuel surcharge moving beyond the EIA diesel average by more than (points)No target set
Loads per 30 days below which lanes sharing a region are proposed for consolidationNo target set