Mid-market 3PL adds Keelway as an overlay on McLeod LoadMaster.
Composite reference profile. Numbers drawn from real engagement patterns at mid-market 3PLs on McLeod; not attributable to a single named customer. Named references available on request under NDA.
The starting point
An 85-broker Midwest 3PL running ~7,200 loads / month across dry van, reefer, and flatbed. McLeod LoadMaster as system of record, deployed for nine years, deeply customized with internal scripting and a custom EDI pipeline to top shippers. Carrier vetting handled via a mix of FMCSA spot checks, Carrier411, and a manual three-tier broker-approval list. Inbound carrier email handled per broker in their own Outlook inbox.
The pain: load volume had grown 22% the previous year and coordinator productivity was flat. The COO had a budget request in front of the CEO for 8 additional coordinator hires — roughly $640K loaded — to keep up with the next year's growth projection. The CEO had rejected the request twice.
McLeod was non-negotiable. The 9-year investment, the EDI pipeline, and the deep customization made any replacement a 12-to-18-month program nobody had appetite for. The question wasn't "what TMS"; it was "what AI fits on top of the TMS we have".
Why Keelway fit the overlay shape
Most freight-broker AI tools assume they get to be the system. Voice-AI tools push brokers into a dialer UI. New TMSs push brokers into a new dashboard. Keelway is calibrated to the opposite: brokers stay in the inbox they already live in (Gmail or Outlook, in this case Outlook through a server-side connector), McLeod stays system of record, and the AI surfaces ranked decisions inside the surface the broker is already using.
The integration to McLeod LoadMaster ran in both directions — Keelway pulled active load context (origin, destination, equipment, target rate, posting time) from McLeod, and pushed accepted carriers + final rates back into McLeod via the LoadMaster API as standard booking events. See the McLeod integration page for the technical detail.
The rollout — 8 weeks, no floor disruption
- Week 1. Discovery, scope sign-off, security packet exchanged. SCIM provisioning from the customer's Okta tenant. Sandbox Keelway tenant provisioned in AWS us-east-1.
- Week 2. McLeod LoadMaster API integration configured. Historical carrier-data migration on the previous 18 months — about 9,400 unique carriers with FMCSA backfill. Label taxonomy and ranking weights reviewed per desk (reefer weights insurance heavier; flatbed weights equipment verification heavier).
- Weeks 3–4. First 10 brokers (the reefer desk) went live in shadow mode — Keelway reading and ranking, but the existing manual triage workflow continuing in parallel. Reefer desk lead reviewed the ranked top-5 vs. their own picks every morning.
- Week 5. Reefer desk cut over to Keelway-driven triage. Daily syncs for the first two weeks; 30-minute weekly check-ins after that.
- Weeks 6–8. Phased rollout across flatbed, dry van, and dedicated-lanes desks. Each desk took 4–5 business days from shadow mode to cutover.
Total elapsed time from contract signature to full-floor adoption: 53 days. No coordinators left, no broker workflow rebuilt, no McLeod customization touched.
What changed at month 6
- Load volume grew 38% over the next two quarters — from 7,200 to about 9,900 monthly loads — driven by reps having time to chase additional shipper relationships instead of triaging inbound carrier replies.
- Coordinator headcount stayed flat at 85. The budget request the CEO had rejected went off the table.
- Fraud incidents dropped sharply. Three double-brokering attempts and one chameleon-MC booking were caught in the first 90 days post-rollout. In the previous 12 months under the manual process, the brokerage had eaten roughly $42K in cargo claims tied to two confirmed double-brokering incidents and one stolen-MC fraud.
- Time-to-cover on standard dry-van loads dropped from a median of 73 minutes to 19 minutes — measured from posting time to first accepted booking.
The contracting shape
Enterprise contract, custom pricing modeled on load volume and seat count, dedicated tenant in us-east-1, SAML SSO via Okta, SCIM provisioning, audit logs streamed to the customer's Splunk deployment. Signed MSA, signed DPA, 99.9% uptime SLA. Named CSM, named SE, 24/7 Microsoft Teams shared channel for incident response.
See Keelway Enterprise for the full posture, or the security page for the compliance specifics.
The overlay is the path. Let's model it for you.
Related
Technical detail on the LoadMaster API integration — pull load context, push accepted bookings.
What it looks like when the brokerage runs three different TMSs across acquired brands.
Dedicated tenant, SAML SSO, SIEM streaming, 99.9% SLA — the posture this profile signs.
