C.H. Robinson's third-quarter 2026 results put a number on what AI agents can do for a brokerage: the company said its agents let staff handle roughly 40% more shipments per person per day, and that productivity is up about 65% since the end of 2022 (C.H. Robinson 8-K; TT News). The quarter itself was strong, with revenue of about $5.4 billion, up 29.5%, and EPS of $1.80.

Days earlier, on October 26, the company launched AI agents for ocean, air and customs that cut quote times from about a day for air and two to three days for ocean to two to three minutes, with more than 80,000 quotes delivered to over 3,600 customers (C.H. Robinson press release, Oct 26, 2026). We covered its agent program in our earlier look at C.H. Robinson's AI agents and freight automation.

For the rest of brokerage, the takeaway is not that one company had a good quarter. It is that shipments per person per day is now a public benchmark. Shippers comparing providers and investors comparing brokers have a reference point, and mid-size brokers will be measured against it whether or not they publish their own number.

Why does shipments per person become the benchmark?

Shipments per person is the benchmark because it captures cost-to-serve in one figure. Brokerage margin is the gap between customer rate and carrier cost, minus the people cost of moving the load. When loads per person rises, cost per load falls, and the broker can either keep more margin or price more sharply.

This fall's results show the pattern from different angles. J.B. Hunt's brokerage grew revenue sharply while company margins tightened, which we analyzed in our post on J.B. Hunt's Q3 brokerage growth and cost to serve. C.H. Robinson's planned RXO acquisition rests partly on applying its "Lean AI" operating model to more volume, covered in our analysis of what the RXO deal means for mid-size brokers. Growth alone does not lower cost per load. Fewer manual touches per load does.

How do you measure your own baseline?

You cannot close a gap you have not measured. Most mid-size brokers know loads per month but not loads per person per day, or where each person's time actually goes.

Step 1: Calculate loads per person per day.

Loads per person per day = Loads covered in the period ÷ (Operations headcount × Working days)

Count the people who touch loads: customer reps, carrier reps, track-and-trace and billing staff, as applicable. Keep the definition consistent so the number is comparable over time.

Illustrative example: a brokerage covering 6,000 loads a month with 25 operations staff over 21 working days handles about 11.4 loads per person per day. A 40% improvement would take that to about 16.

Step 2: Count touches per load. Sample 50 recent loads and count every manual action: quote replies, carrier calls, rate cons, check calls, status emails, appointment requests, POD chases and invoice corrections. Most brokers find a long tail of small tasks.

Step 3: Time the touches. Estimate minutes per touch type and multiply by frequency. That shows where the hours go.

• Quote response — Typical trigger: Inbound email or portal request · Automatable?: Yes, for standard lanes
• Status update — Typical trigger: Customer email or milestone · Automatable?: Yes
• Check call — Typical trigger: Scheduled interval · Automatable?: Yes
• Appointment booking — Typical trigger: Load tender · Automatable?: Yes
• POD and document chase — Typical trigger: Delivery · Automatable?: Yes
• Rate con check — Typical trigger: Booking · Automatable?: Yes
• Carrier negotiation on hard loads — Typical trigger: Tight capacity · Automatable?: Partly; a person decides
• Customer escalation — Typical trigger: Service failure · Automatable?: No; a person owns it

Where do mid-size brokers close the gap?

Mid-size brokers close the gap by automating the high-frequency, low-judgment touches first and leaving negotiation and relationships to people. That does not require a large company's engineering team. It requires agents that work inside your existing email, TMS and carrier channels.

The usual order:

1. Status emails and check calls. Highest volume, lowest risk.
2. Quote intake and response for standard lanes. Quote speed wins business, and minutes-not-days quoting is now on the table.
3. Appointments and documents: booking, POD chasing and rate con verification.
4. Billing preparation: carrier invoice matching and accessorial capture.

Team size is not the barrier it once was. We covered how small brokers with fewer than 10 people deploy AI agents, and how to plan for measurable results in our guide to freight broker AI automation ROI in 90 days.

What does a realistic target look like?

Do not commit to 40% on day one. C.H. Robinson's productivity gain built up over several years. A practical plan sets a baseline this quarter, automates the two or three most frequent touch types and measures loads per person monthly. Report the number internally the way you report margin. If it moves, you will know which workflows moved it. If it does not, you will know where the time still goes.

Frequently asked questions

What did C.H. Robinson report about AI productivity in Q3 2026?

It said AI agents let staff handle roughly 40% more shipments per person per day, and that productivity is up about 65% since the end of 2022 (C.H. Robinson 8-K; TT News).

How do I calculate loads per person per day?

Divide loads covered in a period by operations headcount multiplied by working days. Keep the headcount definition consistent so you can track the number over time.

How fast are C.H. Robinson's new ocean and air quotes?

The agents it launched on October 26, 2026 cut quote times from about a day for air and two to three days for ocean to two to three minutes, with more than 80,000 quotes delivered to over 3,600 customers (C.H. Robinson press release).

Can a mid-size broker match large-broker productivity gains?

The gains come from automating repetitive touches per load, which does not require scale. Start with status updates, check calls and quotes, measure your baseline and track loads per person monthly.

The bottom line

C.H. Robinson has made shipments per person per day a public benchmark. Measure your own loads per person and touches per load now, automate the most frequent low-judgment work first and track the number monthly. The brokers who know their baseline will be the ones who can show they are closing the gap.

Debales deploys AI agents for quoting, status emails, check calls, appointment booking and document chasing, the per-load touches that set your shipments per person. Book a demo.