Lanesurf Is Letting AI Negotiate Your Freight Rates While You Eat Lunch

The Macro: Freight Brokerage Runs on Phone Calls and Spreadsheets

I spent a week talking to freight brokers about their daily workflow. Every single one described the same thing: wake up, check the load board, start calling carriers, negotiate rates, confirm compliance docs, book the load, repeat sixty times before 5 PM. It is relentless, repetitive, and almost entirely manual.

The US freight brokerage market moves roughly $100 billion a year. The core transaction is simple. A shipper needs something moved. A broker finds a carrier to move it. The broker takes a margin. But the execution of that simple transaction involves dozens of phone calls per load, real-time price negotiation, FMCSA compliance verification, insurance checks, and dispatcher coordination. Most brokerages still run this on spreadsheets and TMS software that was built in the early 2010s.

There have been attempts to automate parts of this. Convoy tried to build a “freight Uber” and burned through hundreds of millions before shutting down. Transfix raised $300 million to do something similar. Uber Freight, despite the brand and budget behind it, has not dominated the way people expected. The pattern is clear: trying to disintermediate the broker entirely does not work because shippers and carriers still want a human relationship at the center of the transaction.

The smarter play is not replacing the broker. It is giving the broker superpowers. That is where voice AI gets interesting.

The Micro: Two IIT Delhi Grads Teaching AI to Haggle

Pratham Bansal and Sarthak Singh Chauhan founded Lanesurf in San Francisco after coming through Y Combinator’s Summer 2025 batch. Both studied engineering at IIT Delhi. Bansal built inference systems for compute-heavy ML models, work that eventually sold to an NVIDIA-backed company. Chauhan co-authored transformer research and built production inference tools adopted by Fortune 100 biopharma companies. They are deeply technical founders attacking a market that is decidedly not technical. That mismatch is the whole point.

Lanesurf is voice AI that calls carriers, negotiates rates, verifies compliance, and books freight. Not one call at a time. The system talks to over 100 carriers simultaneously for a single load, running parallel negotiations to find the best rate within the broker’s acceptable margin. It checks MC registrations, FMCSA status, COI documentation, and dispatcher legitimacy in real time during the conversation.

The self-improving part is what separates this from a glorified autodialer. Every call generates data about carrier behavior, pricing patterns, and negotiation tactics. The system uses fuel prices, weather data, and historical lane performance to anchor its rate offers. Over time, it learns which carriers are reliable on which lanes, who tends to accept lower rates, and when the market shifts. A human broker builds this intuition over years. Lanesurf is building it across thousands of calls per month.

Setup takes about an hour. You upload a spreadsheet and the system starts dialing. They have already processed loads for shippers including Walmart, Sysco, and Lipman. Early case studies show an 80% reduction in booking time and an 8% margin increase for a produce brokerage.

I find the compliance verification angle particularly smart. Broker compliance failures can result in six-figure fines. Having an AI that automatically checks every carrier against federal databases before booking removes a category of risk that currently depends on a human remembering to run the check during a hectic day.

The integrations list covers the obvious bases: DAT, major TMS platforms, Gmail, Outlook, WhatsApp. SOC-2 certified with AES-256 encryption, which is table stakes for anyone handling shipping contracts at scale.

What I want to know is how the AI handles the messy middle of a negotiation. Carrier dispatchers are not reading from scripts. They have accents, bad connections, background noise, and a tendency to throw curveballs about equipment type or pickup windows. Phone negotiation is adversarial in a way that customer service calls are not. The dispatcher is actively trying to get a higher rate. Lanesurf needs to hold its own without being rigid.

The Verdict

Lanesurf picked the exact right slice of the logistics market to attack. Instead of trying to eliminate brokers (the Convoy mistake), they are making each broker dramatically more productive. A single broker using Lanesurf can theoretically cover the call volume of a five-person team. That math is compelling for any brokerage owner doing hiring math.

The competitive landscape is thin. There are general-purpose voice AI platforms like Bland AI and Vapi, but none are purpose-built for freight negotiation with built-in compliance checks and carrier vetting. Parade and Emerge do carrier matching algorithmically but without voice. The combination of voice plus compliance plus self-improvement is genuinely differentiated.

At 30 days, I want to see how many loads per month are being booked without human intervention, start to finish. At 60 days, the question is whether the margin improvements hold up as the system scales across different freight categories. At 90 days, I need to know if brokers are trusting the AI enough to let it run unsupervised or if they are still spot-checking every booking. The technology is impressive. Whether freight brokers will actually trust a robot to negotiate on their behalf is the human question that determines everything.