AI Dispatch Inside X-TMS
Stop guessing which driver to assign. X-TMS scores every option, shows the reasoning, and lets you assign in one click.
Driver assignment is the highest-leverage decision you make — let X-TMS make it sharper
Every load you dispatch is a multi-variable optimization problem: which driver is closest, which has the right equipment, who's running on-time today, who's most efficient on this lane, who's running over hours, who's been waiting longest.
Most dispatch decisions get made on gut feel — and gut feel is wrong 20–30% of the time. A misassigned load can cost hours of detention time, an overspent fuel budget, or an unhappy driver.
X-TMS AI scores every available driver against every open load using six weighted factors. The recommendation is explainable — you see exactly why driver A scored 87 vs driver B's 71, and you can override with one click.
How it works
-
1
Define your assignment strategy
Three preset strategies — Efficiency-first (maximize on-time delivery), Balanced (multi-factor average), or Cost-first (minimize fuel + pay). Customize weights per division.
-
2
AI scores every driver-load pair
Six factors: driver historical efficiency, on-time rate, equipment capability match (hazmat, refrigeration, oversized), cost optimization (fuel + driver pay), load priority weighting, and hours-of-service compliance.
-
3
Top recommendations surfaced
Top 3 driver candidates per load, each with a numeric score and full reasoning — exactly which factors drove the score.
-
4
Dispatcher confirms or overrides
Most dispatchers accept the AI recommendation 80%+ of the time. The 20% override cases get logged so the model improves over time.
-
5
Assignment commits with full audit
Every assignment records the AI score, alternative candidates considered, and the dispatcher's choice. Searchable history for performance review.
By the numbers
What's included
- Six-factor driver scoring (efficiency, on-time, equipment, cost, priority, HOS)
- Three preset strategies plus custom weight configuration
- Equipment capability filtering (hazmat, refrigeration, liquids, oversized)
- Hours-of-service compliance check before suggestion
- Top-3 candidate display per load
- Explainable scoring (full reasoning per candidate)
- One-click assignment from recommendation panel
- Override logging for model improvement
- Per-division weight customization
- Real-time re-scoring as drivers complete loads
Best fit for
AI handles the routine 80%, dispatcher focuses on exceptions.
AI recommendations train new staff on assignment logic — see the reasoning live.
AI prevents non-certified driver suggestions before they happen.
Different scoring weights per division — efficiency for one, cost for another.
What to expect during rollout
Most teams underestimate the people-and-process work that surrounds any new technology. AI Driver Matching is technically straightforward to switch on, but a smooth rollout still benefits from a short coordinated effort across dispatch, IT, and ownership. Below is what we typically see in successful deployments.
Week 0 — Stakeholder alignment
Identify a single internal owner for the rollout. Confirm the metric you intend to improve (calls placed per day, hours saved per dispatcher, load-to-driver lead time, settlement cycle time — whichever applies). Align ownership, dispatch leads, and any affected drivers on what's changing and why. This step takes one or two short meetings, not weeks.
Week 1 — Pilot setup
Connect AI Driver Matching to a narrow scope first — one dispatcher, one lane, or a subset of customers. Validate that the integration behaves as expected on your real data. Capture any edge cases your operations have that the standard configuration didn't anticipate. X-TMS support is available throughout this phase.
Weeks 2–4 — Scale up gradually
Expand to more dispatchers, more lanes, or higher volume. Most teams scale to full production within 2–4 weeks of the initial pilot. Track the metric you committed to in Week 0; it's the honest signal of whether the deployment is doing what you bought it for.
Ongoing — Iterate
Review AI Driver Matching performance monthly with your team for the first quarter. Tune configuration (criteria, thresholds, weights) based on what's working and what isn't. This is normal — every AI-driven workflow benefits from a few iterations as it learns your specific operation.
Common pitfalls to avoid
Skipping the pilot. Teams that try to flip the switch globally on day one tend to discover edge cases at the least convenient moment — under live operational load. A one-week pilot prevents this.
No defined success metric. If you can't articulate what "good" looks like, you won't know whether the deployment succeeded. Pick one number and track it.
Treating AI as a black box. AI Driver Matching provides reasoning behind every recommendation. Take advantage of it during the first few weeks — your team learns the AI's logic, and the AI learns your team's preferences.
Frequently asked questions
Is the AI a black box?
No. Every score is decomposed into its factor contributions — you see exactly why driver A scored higher than driver B. Dispatchers can sanity-check every recommendation.
What if the AI recommendation is wrong?
Override with one click. The override is logged with your reasoning and feeds back into the model. Over time the AI learns your team's judgment patterns.
Does it work with my custom driver pay structure?
Yes. Driver pay (percentage of load, fixed per load, or per-mile) is configurable per driver and feeds directly into the cost-optimization factor.
How does HOS (hours of service) integration work?
X-TMS reads ELD data from Samsara (or via direct ELD API) and excludes drivers who can't legally complete the load within their remaining hours.
Can I disable AI recommendations and assign manually?
Yes. AI recommendations are surfaced as suggestions — dispatchers always have full manual control.