Every filter the Blacklane bot checks

Speed is only worth having if the bot accepts the right work. These are the rules an offer has to clear, what each one is actually for, and the mistakes that make a filter set behave in ways chauffeurs do not expect.

How the filters are evaluated

Filters are not a scoring system. There is no total that a generous price can push over a threshold. Every rule is a gate, the gates run in a fixed order, and a single failure rejects the offer. That is deliberate: it means you can reason about why something was rejected without modelling interactions between rules.

The order is exclusions first, then economics, then feasibility, then capacity. Exclusions come first specifically so that no price can override a place or a class you refuse.

Price rules

Minimum price

A hard floor per ride. Anything below it is rejected without further evaluation. Keep separate floors for transfers and hourly bookings — an hourly job priced like a transfer is usually a loss once you account for waiting time.

The common mistake is setting this from what you would like to earn rather than what you would actually decline. If the floor is above the market rate in your city, the bot correctly rejects everything and it looks broken.

Price per kilometre

Judges the offer by its rate rather than its headline figure. A three hundred kilometre transfer with a large total can still be poor work once you count the return leg you are not paid for. This filter is what stops long, cheap rides from passing the minimum price gate.

Distance and geography

Distance limits

Two separate caps: how far you are willing to travel to reach the pickup, and how long the ride itself may be. The first protects your unpaid time, the second stops a single job from swallowing a day you had planned to fill with several.

City whitelist

Name the cities, airports and zones you want. Addresses, city names, airport codes and postal codes can all be used. A whitelist is the safest way to start: the bot accepts nothing outside the areas you have explicitly approved.

City blacklist

The inverse, and evaluated before everything else. Areas you never want are rejected regardless of price, class or timing. If a zone appears in both lists, the blacklist wins — there is no ambiguity to resolve.

Vehicle and class rules

Accept only the classes you are licensed and equipped for — business, first, SUV, van, electric. Rules are per vehicle, so each car in a multi-car operation carries its own class list, its own price floors and its own working hours.

This is what makes a two-car operation practical to automate. One vehicle can run a strict high-value rule set while the other takes broader work, and neither rule set contaminates the other.

Timing rules

Minimum notice

Rejects offers that start sooner than you can realistically reach the pickup. The window is measured from the moment the offer arrives, not from a fixed cutoff time, which is the only measurement that means anything when offers appear at unpredictable moments.

Working hours and schedule

The hours and days on which the bot may accept. Outside that window it keeps watching and logging but answers nothing. Chauffeurs who work split shifts usually configure two windows rather than one long one.

Capacity and conflict detection

The final gate, and the one that separates a filter list from a dispatch system. Before accepting, the bot compares the offer against every ride already booked across all of your vehicles — including rides it accepted seconds earlier — and refuses anything that would double-book a car or a driver.

  • Overlapping pickup and drop-off windows are detected across the whole fleet, not per vehicle
  • Travel time between consecutive rides is accounted for, not just the booked hours
  • Two offers arriving simultaneously cannot both be accepted into the same slot
  • The losing offer is rejected with an explicit overlap reason in the log

Overbooking is the failure mode that costs a chauffeur their rating rather than just a fare, which is why this check runs last and cannot be disabled by a price rule.

Reading the decision log

Every decision — accepted or rejected — is recorded with the rule that produced it. That log is the tool for tuning your filters. A day of rejections tells you more about your rule set than any amount of planning: if you see repeated rejections for work you would have taken, the rule that fired is the one to loosen.

Set your own rules in the trial

Start conservative, watch the decision log for a day, then loosen the rules that rejected work you would have taken.

Start free trial