Free Intensive Nov 21–22Free 2-day hands-on Vibecoding Intensive on Nov 21–22 at 16:00 MSK (Sat–Sun) · Live Cohort 3 starts Nov 23Join Free →Cohort 3 (Nov 23)
ReleaseInsight · On RadarElevenLabs Blog→ original

Voice AI Agent Lifted Logistics Broker Trinity's Margins by 25%

Logistics broker Trinity Logistics has put carrier calls in the hands of ElevenAgents, a voice AI agent from ElevenLabs. The agent negotiates rates and books loads without a human on the line. According to the company, margin (the broker's share of earnings per load) is 25% higher on these agent-booked loads than on standard ones, and the platform's per-minute cost fell 40% after Trinity moved off an external contractor. The takeaway: when your own team, rather than the vendor, controls prompts and integrations, the agent gets configured faster.

AI-processed from ElevenLabs Blog; edited by Hamidun News
Voice AI Agent Lifted Logistics Broker Trinity's Margins by 25%
Source: ElevenLabs Blog. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceElevenLabs Blog ↗
Editor-in-Chief
Key metric #1About 30,000 calls and 1,000 automatic bookings per month
Key metric #2On Saturdays, the on-call shift handles about 120 calls instead of 500–600
Key metric #3The agent resolves 60–70% of inbound calls before a human picks up

Picture a freight broker. It doesn't haul cargo itself. Instead, it connects cargo owners with carriers and earns the spread between the two prices.

Loads are posted to an industry exchange, carriers call in, negotiate rates and book the trip. The calls are numerous and repetitive, and operators spend their whole day on them. U.

S. broker Trinity Logistics handed exactly this work to ElevenAgents, a voice AI agent from ElevenLabs. According to the company itself, margin (the broker's share of earnings per load) is 25% higher on automatically booked loads than on standard ones, and the cost per minute of running the platform fell by 40%.

An important caveat: this is a case study published by the vendor, so the numbers should be read with marketing in mind.

What the agent does and why Trinity switched contractors

Trinity is a third-party logistics provider (3PL in industry parlance): the company doesn't own trucks, it organizes shipments. When a carrier calls about a posted load, ElevenAgents discusses the rate and books the trip within the parameters Trinity sets. For the driver, it's a single call with no transfer to a person.

Previously, an external provider handled these calls. As call volume grew, the service became too expensive. As part of an efficiency push, the company decided to bring the agent in-house and approached ElevenLabs directly. The platform was chosen for self-service: Trinity's team connected the agent to its own transportation management system (TMS) and builds the scenarios itself.

The migration was done carefully. The success criterion was that the operations department would not notice the platform change at all. Departments were moved one at a time, feedback was collected every week, and only then did the team move on to the next. According to the company, staff never realized they had been switched to a new system.

What the numbers show

The company reports the following results:

  • margin on automatic bookings is 25% higher than on prior bookings;
  • the platform's per-minute cost fell by 40%;
  • the share of successful bookings was 4%, on par with the previous system;
  • roughly 30,000 inbound calls and 1,000 automatic bookings per month.

The 4% figure looks modest, but it is consistent with the other data: 1,000 bookings from 30,000 calls works out to about 3–4%. Many carriers call simply to check rates rather than to take a load right away, so high conversion shouldn't be expected. The key point is that the agent did not lose ground to the previous setup; it became cheaper.

The second set of figures concerns people. Kimberly Punt, director of after-hours operations, leads a remote team of seven. On a typical Saturday they used to take 500–600 inbound calls; now it's around 120. The company estimates that ElevenAgents handles 60–70% of inbound carrier calls before a human could pick up. Staff focus on exceptions: a broken-down truck, a driver stuck at the shipper, a pickup number that doesn't work. They also book loads with carriers in advance. Staffing up for Saturday peaks is no longer necessary.

What developers and product managers can take from this

The main engineering lesson: Trinity treated the voice agent as its own system, not as a subscription. Previously, changing the agent's behavior meant a multi-week cycle with the contractor. Now the team adjusts prompts and logic itself, based on live feedback. The company links its margin gains specifically to a shorter loop between operations and developers: they spot faster where the agent gives in too easily during negotiation and fix the rules.

Practical takeaways:

  • if call volume is growing, compare not only the per-minute price but also the cost of each change with the vendor;
  • connecting the agent to an internal system (here, the TMS) sets the boundaries of negotiation and lets it book on its own;
  • move teams over one at a time, and treat "staff noticed nothing" as a separate success metric;
  • decide in advance which cases the agent hands off to a human.

Caveats

The source is the vendor's blog, with no independent verification. Margin is compared against "prior bookings," but it isn't stated how comparable those loads are in terms of routes and cargo types, so causation is not established. The absolute per-minute price, model costs and migration costs are not disclosed, so the full economics cannot be calculated; only the 40% drop in per-minute price is known. The case's strongest element is the setup itself: the agent closes routine calls, people handle the exceptions, and an in-house team owns the prompts. Trinity's next steps are to filter inbound calls and build a support agent for common questions.

Story

Key Questions Answered

What the agent does and why Trinity switched contractors?

Trinity is a third-party logistics provider (3PL in industry parlance): the company doesn't own trucks, it organizes shipments. When a carrier calls about a posted load, ElevenAgents discusses the rate and books the trip within the parameters Trinity sets. For the driver, it's a single call with no transfer to a person. Previously, an external provider handled these calls. As call volume grew, the service became too expensive.

What the numbers show?

The company reports the following results: margin on automatic bookings is 25% higher than on prior bookings; the platform's per-minute cost fell by 40%; the share of successful bookings was 4%, on par with the previous system; roughly 30,000 inbound calls and 1,000 automatic bookings per month. The 4% figure looks modest, but it is consistent with the other data: 1,000 bookings from 30,000 calls works out to about 3–4%.

What developers and product managers can take from this?

The main engineering lesson: Trinity treated the voice agent as its own system, not as a subscription. Previously, changing the agent's behavior meant a multi-week cycle with the contractor. Now the team adjusts prompts and logic itself, based on live feedback. The company links its margin gains specifically to a shorter loop between operations and developers: they spot faster where the agent gives in too easily during negotiation and fix the rules.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Private AI Builders Community · Vibe Club

Ship AI Faster with Founders, Engineers & Weekly Architecture Reviews

Inside Vibe Club (hamidun.com/club) — weekly live masterminds with Zhemal Khamidun, Tech Radar, full LMS access, and 370+ production Claude Code skills.

Join Vibe Club →Explore Hamidun Academy
What do you think?
Loading comments…