AI agents for logistics and freight companies

Most freight and logistics websites do two jobs: turn shippers into quote requests and turn carriers into sign-ups. Both depend on speed. A shipper with a load to cover can ask several providers at once, and a long quote form at 7 p.m. is easy to give up on. This page covers where AI agents help a broker, 3PL, carrier or forwarder, and the limits they need.

Who visits a freight or logistics website

• Shippers with a load. They have a lane, an equipment type, dates, a weight and a commodity, and want a rate or at least a fast answer on whether you can cover it.
• Shippers vetting you. Before tendering freight they check your operating authority, insurance, modes, regions, tracking options and how you handle claims. Every US property broker must keep a $75,000 surety bond or trust fund in place (49 CFR 387.307), and shippers can look up a company's USDOT and MC numbers in FMCSA's public Company Snapshot. Expect questions about both.
• Carriers and owner-operators. They want your setup packet, payment terms, quick-pay options and how to see your loads.
• Existing customers. Tracking, proof of delivery, invoices, claims.
• Drivers and dispatchers reading your careers page.

Which ASDK agents fit

All four agents draw on one knowledge base built from your website: services, lanes, equipment, carrier requirements and policies.

Sales: turn quote requests into qualified opportunities (early access)

Sales collects what your pricing desk needs, in conversation instead of a 20-field form, and decides with your rules whether this is a spot load, a recurring lane or a fit for an account manager. Leads go to your team by email or Slack, and to your CRM or TMS through webhooks and Zapier or Make, and recurring business can be booked straight onto a rep's calendar.

Example flow

1. A visitor writes: "Need a rate for two dry vans, Chicago to Columbus, picking up Thursday."
2. Sales asks for the commodity, weight per truck, pickup and delivery hours, and whether this is a one-off or a weekly lane.
3. It does not invent a rate. It confirms that you run dry van in the Midwest (from your services page), says when a quote will arrive based on the response time you publish, and captures an email and phone number.
4. The visitor mentions the lane runs every week. Your rules treat that as an account opportunity, so Sales offers a call with the account manager for the region.

Chat: answer service, coverage and carrier questions (early access)

Chat handles the questions that clog your inbox, answered from your pages with a link to the source.

Example flow

1. A carrier asks: "How fast do you pay, and do you offer quick pay?"
2. Chat answers from your carrier page (standard terms, quick-pay fee and how to request it) and links the setup packet.
3. The carrier asks whether you take hazmat loads. Your services page lists the classes you accept, so Chat answers with that list and adds that every hazmat load is confirmed by dispatch.

Document requests work the same way. If your certificate of insurance, W-9 and carrier packet are published, Chat links them; if they are sent on request, it takes the request and the email address for your team. Chat does not see live shipment data. Asked "where is my load?", it points to your tracking portal and offers to have someone follow up.

Forms: conversational quote requests (coming soon)

Forms is planned to replace a static quote form with a short conversation that asks only the relevant questions (LTL needs pallet dimensions; full truckload does not) and can show a price range from rules you describe. Until it ships, Sales collects the same details. One tip for LTL: ask for the dimensions and weight of each handling unit rather than "what class is it?". NMFTA's 2025 docket, in effect since July 19, 2025, moved many items toward density-based classification (NMFTA), and many shippers are not sure of their class.

Agent-Ready: make your services readable to AI agents (early access)

Agent-Ready publishes an llms.txt file that points AI systems to your modes, lanes, equipment and carrier setup pages. A browser agent working for a shipper can use "ask a question" to check whether you cover a lane, or "contact sales" to pass on the load details, instead of guessing its way through a 20-field quote form. On Pro and above, you also see which AI agents read which pages. More on why that matters below.

Guardrails for freight websites

A wrong answer about a rate or liability limit costs real money. In 2024 a British Columbia tribunal held Air Canada responsible for what its website chatbot told a customer (Moffatt v. Air Canada). Treat your agent's words as your company's words:

• No binding rates from the agent. It can repeat published minimums or ranges, labeled as estimates. Real quotes come from your pricing desk or tool.
• Transit times are estimates. Say so every time, and avoid "guaranteed" unless you sell a guaranteed service.
• Escalate special freight. Hazmat, oversize and overweight, high-value cargo, temperature-critical pharma: capture the details and route them to a person.
• Liability and claims only as written. The agent quotes your published terms and insurance amounts and sends claims to the claims team. It never estimates what a claim will pay.
• Keep shippers and carriers apart. Ask early whether the visitor wants to ship or to haul, so a carrier is not run through shipper qualification.
• No banking details in chat. Carrier bank details and tax forms belong in your onboarding portal, not a chat window.
• International terms from your pages. If you quote under the Incoterms 2020 rules, list the ones you support. The agent should not explain what a term means for a specific contract.

How AI agents research freight partners

A logistics manager might ask an assistant: "find asset-light brokers with reefer coverage from California to Texas that pay carriers within 30 days and have active authority." Assistants search the web, read a few provider sites and compare them. Agentic browsers can go further and fill in your quote form for the user. They can only use what your pages state plainly; AI agents are already visiting your website explains how to spot them.

Put these on ordinary, crawlable pages, not only inside PDFs:

• Modes and equipment: FTL, LTL, reefer, flatbed, drayage, intermodal, warehousing.
• Lanes and regions, including cross-border service and the border crossings you use.
• USDOT and MC numbers as text, plus your insurance coverage amounts.
• Response times you can keep, such as when quotes go out after a request.
• Accessorial charges: detention, layover, lumper, liftgate, or at least how you bill them.
• Specialties and certifications, such as hazmat, food-grade or bonded warehousing.
• Technology: tracking links, EDI or API connections.
• How to request a quote and which details to have ready.

Terminal and warehouse addresses belong in structured data; your USDOT and MC numbers, modes and main lanes belong in llms.txt. Rate tools and tracking portals are often hidden behind scripts or logins, so run the free agent readiness checker to see which of your pages an AI agent can actually read. Agent-Ready keeps the summary and agent actions current as your lanes change.

Where to start

Start with Sales on your quote page and Chat on your carrier and services pages, with the guardrails above in their instructions. Read the first week of conversations, fix the answers your site was missing, then decide which quote flows to hand to Forms when it arrives. Pricing shows which plan includes each agent.