WW/SUPPLYCHAI
SEVOS Told a Client to Cut Three of 20 Roles Before Automating Anything
SEVOS told a client three of 20 outsourced roles were unnecessary, staffed them empty for three weeks as proof, and now runs 225 people across Atlanta, India and Serbia on an AI-last model.
- By
- Marcus Bennett
- Filed
- Length
- 891 words
- Read
- 4 min

Key points05
- SEVOS studied 20 roles for a client and concluded three were unnecessary, proving it with a three-week empty-seat trial.
- The Atlanta-based company soft-launched last year and has grown to 225 people under COO Jay Singh.
- SEVOS onboards each role in a fixed 30-day sequence before any automation decisions.
- Teams in Atlanta, Chandigarh (India) and Serbia all work Eastern time, covering six languages.
- Track-and-trace paperwork — PODs, bills of lading, scanning — is automated first; negotiation and relationship building stay human.
When a freight company asked SEVOS which of 20 roles it should outsource, the Atlanta-based firm returned an answer most vendors would not volunteer: three of those roles were unnecessary. The client pushed back. SEVOS responded by placing floaters in the three seats and assigning them no work, betting the workflow would hold without them. Three weeks later the client had its proof — and three fewer positions to pay for.
"That also builds trust with the customer where they trusted us with more," said Jay Singh, SEVOS' chief operating officer.
The anecdote is the company's clearest commercial pitch, and it comes with an obvious caveat: a three-week trial at one unnamed client is a demonstration, not an audited case study. But it frames what SEVOS says distinguishes it in a freight sector being pushed toward AI adoption with little clarity about what to automate, by when, or at what cost.
What is SEVOS, and who runs it?
Singh ran a trucking company before co-founding SEVOS with Michael McLary. The Atlanta-based company soft-launched last year and has grown to 225 people — a headcount Singh manages directly, though he describes the reporting line running the other way. "225 people report to me," he said. "However, I report to every single one of my customers."
Earlier this year, before the official launch, SEVOS hired Tom Carr as senior vice president of customer solutions to lead the commercial side. The company made its public debut at FreightWaves' F3: Future of Freight Festival in Chattanooga.
How does the 30-day onboarding work?
Singh's method inverts the usual automation pitch. Instead of selling software first, SEVOS fills the role, documents it, and measures it. Only then does the company decide which tasks software should absorb.
The sequence is fixed:
- Week one: identify the roles.
- Week two: collect and study the client's standard operating procedures.
- Day 30: a SEVOS team member is working the seat.
- Before any automation: each role gets KPIs and coaching tools.
In the sales development work SEVOS already performs, callers use AI-assisted scripts and receive real-time AI tutoring. Singh frames the human-first sequence as non-negotiable: "You have to bring the customer along for the journey."
Which freight tasks get automated first?
Track-and-trace paperwork is where automation lands first. Collecting proofs of delivery and bills of lading, scanning them, and moving them into accounting is repetitive and rule-bound — the textbook profile of work software handles well.
"That stuff is getting automated really, really fast, and it should," Singh said.
Three functions stay with people: carrier negotiation, relationship building, and sales development. The work surrounding those conversations does not. A proposal that follows a sales call, for example, can be generated and sent in real time rather than waiting until the end of the day.
Clients also choose the underlying technology — open-source or closed-source models, hosted on their own servers if they prefer. That choice matters for brokerages and carriers with data-security concerns, though SEVOS has not published pricing for either model. Over time, the company wants customers to pay a per-transaction fee rather than a headcount-based rate, a shift that would tie SEVOS' revenue directly to the volume it processes. No timeline for that transition was given.
Can clients watch the work happen?
SEVOS Insights, the company's reporting tool, shows the client every action a team member takes. It can also connect to the client's own systems, such as Tableau, so productivity is visible person by person. The transparency claim is the company's answer to the oldest objection in outsourcing — that quality drops once the work moves offshore — but the tool's adoption rate among clients was not disclosed.
What about shifts and time zones?
SEVOS runs teams in three locations, all working on Eastern time: Atlanta, Chandigarh in India, and Serbia, which fills the hours between India's shift and the U.S. morning. The model supports day-shift work, including sales development, as well as coverage outside normal business hours. Customers decide which seats and which hours to staff based on the work, not a fixed starting point.
The teams work in English, Spanish, Serbian, Punjabi, Hindi, and Haryanvi. Singh says the language coverage matters in carrier sales and in work with newly authorized trucking companies — a segment that has grown steadily as new entrants continue to register with the Federal Motor Carrier Safety Administration.
AI agents sit at the top of the stack as the last layer, not the first. Placed on top of a follow-the-sun team, they extend coverage across the full spectrum of hours and tasks.
What remains unproven?
The numbers SEVOS has released describe scale, not outcomes: 225 people, a 30-day onboarding clock, a three-week proof exercise. The company has not published client counts, retention rates, or cost savings measured against in-house staffing. The per-transaction pricing model remains an ambition rather than a live offering. What SEVOS has demonstrated, on its own account, is that a services firm willing to tell a client to cut three roles can win more business than one that simply fills all 20 — and that in freight AI adoption, the person in the seat still comes before the software that may eventually replace the task.
Original: sevos.ai
More from Marcus Bennett
Show full bio
Senior reporter covering regulation, policy and their impact on Ocean and air freight, container shipping at Waybill Wire.
940 articles
Related05
Hundreds of Vendors, Billions at Stake: Automated Forklifts Face a Shakeout
Vietnam's 13th Mekong-Japan Summit Sets Supply Chain Agenda
North American factories reset supply networks for resilience in 2026
AYK wins 165 MWh order for Molslinjen electric ferries
Amazon Cuts 20-Minute Manual Calls to Seconds With AI Agents