AI Employees for logistics

One AI Employee for every leader.

SN3 builds AI Employees for 3PLs and fleet companies, one per leader: a chief of staff for the CEO, one for finance, one for the leader who runs customer onboarding. They work in Slack or Telegram, inside the systems you already run on, and draft for your approval.

We’ve launched them at a 3PL serving 40 to 45 clients, and at a fleet-tracking company that onboards 50 to 70 new customers a year.

Book my 15-minute Launch Plan See the seats

15 minutes · No purchase required

You: Anything I missed yesterday?, 6:52 AM, read
Sage: Needs a reply

• Brand client: why Monday’s retail order didn’t ship.

• Carrier: Q1 rate change, wants an answer by Friday.

New business

• A brand shipping 800 orders a month wants a call this week.

From the floor

• 2 inbound receipts came in short on the count., 6:53 AM
Sage: Replies to the brand and the carrier are drafted for your OK., 6:53 AM

Buttons: Review & approve, Open in Gmail

A sample morning check-in. Names and details are illustrative.

6AI Employees launched across two logistics companies

1,200+Real emails a day to one CEO, by her estimate

Slack + TelegramWhere each leader works with theirs

Your OKBefore anything reaches a client, carrier, or customer

Where the hours go

The freight moves. The leaders are stuck.

What leaders at our two logistics clients told us about their own work.

Inbox

An inbox no one can keep up with

One CEO estimates she gets over 1,200 real emails a day. An important client email that came in at 3 p.m. wasn’t seen until 9:30 that night.

Execution

Leaders doing the execution

“All of us are fully aware that we’re very much in doing mode instead of like thinking strategy and how we execute strategy,” one leader said of the leadership team.

Spreadsheets

Processes that won’t scale

Much of one 3PL ran in Google Sheets, which a leader said wasn’t scalable across 40 or more clients.

Data

Data in fifteen places

One CEO said company data lived in “probably like 15 places, if I’m being honest.”

Finance

Reconciliation in too many steps

Reconciling a vendor’s statements against QuickBooks takes several hand-offs, even with an AI tool the finance lead already trained. In July, the steps behind one ten-input report were “all in like, someone’s head right now.”

Onboarding

Slow time to first value

Three people on onboarding, with 10 or more accounts possibly in progress at once. Time to first value was an estimated 45 days; the CEO’s aspiration is billing every customer within 30.

The seats

Six AI Employees, five leaders.

Both clients started with one seat, then added more. Here’s the work each one is set up for, with anything still being connected or tested labelled.

For the leadership team

A chief of staff for each leader

For the CEO and the VP of Technology at the 3PL, and two for the CEO at the fleet-tracking company.

Briefings: Long briefings written up as a ClickUp doc, a setup one leader asked for so they aren’t reading everything as text messages.
Work by chat: Real work from wherever the leader is. One CEO marked up a serious legal document over Telegram from his backyard.
Proposals: The CEO’s own proposal method, run through his chief of staff: scope it in chat, say “build it,” and it’s live as a web page.
Finance pulse: A second, finance-focused chief of staff, launched in October. The CEO wants it to keep him “dialed in daily” on what’s selling, what isn’t, and who’s selling it.

For finance and onboarding

Seats for the people behind them

For the head of finance and the operations lead who runs customer onboarding at the fleet-tracking company.

Ticket tracking: The onboarding lead called the help-desk connection “crucial”: knowing when a ticket is created and whether it’s been escalated to partners. That connection is being validated.
Email drafts: On day one, the onboarding lead asked it for a correction “in an email draft… And then I can approve and send.”
Finance trackers: It ran its first shift on launch day. Next: its findings go on a Notion dashboard, and daily tracker updates are proposed.
Reconciliation: Proposed next, not running yet: vendor statements matched against QuickBooks, with differences flagged for a person.

Out of the box they draft and report. Anything going to a client, a carrier, or a customer waits for your OK.

From our logistics launches

They started with one. Both added more.

What our two logistics clients told us on their launch and check-in calls, including what’s still being tuned.

3PL · second seat

A second AI Employee, for the CEO

A couple of weeks into the first launch, the VP of Technology asked to set one up for the CEO next. It launched in August, connected to the CEO’s email and calendar.

3PL · build or buy

Building leverage

The VP of Technology on the pricing: the architecture behind these agents “would take me months, years to figure out on my own.”

Fleet company · four seats

One seat in June, four by October

It started with the CEO’s chief of staff. In October it added seats for the head of finance and the onboarding lead, and a second, finance-focused one for the CEO.

Fleet company · proposals

Proposals from a chat

The CEO built his own proposal method, and SN3 packaged it to run in his chief of staff. He scopes a proposal there, says “build it,” and it’s live as a web page. “For the most part, it knocks it out of the park.”

Fleet company · ROI

Close to paying for itself

About two weeks after launch, the CEO said it was “probably getting close to providing ROI on what it costs us to get it there already.”

Honest note

Still being tuned

In September, one CEO said his chief of staff couldn’t handle scheduling yet, and that even in daily briefings, “it just misses stuff,” like a document he’d already signed and sent back.

I am feeling optimistic and clear-headed and energetic for the first time in a while… [She] has helped me with so much. Like in the last week, I was like, all right, I’m gonna get out of, you know, inbox management and, like, really give you some work… she’s just been tremendously helpful.

VP of Technology, 3PL fulfillment company · /SN3 Group client

Just like the amount of work I can get done when I’m not sitting at my laptop is outrageous.

CEO, fleet-tracking company · /SN3 Group client

Your stack

Connected to the systems your operation runs on.

It works inside the accounts you name, with the access you approve.

Chat

Slack and Telegram

Each leader talks to theirs in Slack or Telegram. One CEO connected his to Slack himself and set up separate channels for briefs, notifications, and alerts: “that’s so much more effective.”

Email & calendar

Gmail and Google Calendar

Connected to each leader’s inboxes and calendar. One CEO asked hers to cover three inboxes: her own, sales, and accounting.

Work management

ClickUp and Notion

Briefings written up as ClickUp docs, and findings for a Notion dashboard.

Business systems

Zoho

Business records through Zoho, connected as each account’s API access allows. Help-desk ticket tracking is still being validated.

Publishing

Netlify

One CEO’s proposals go live as web pages on Netlify, straight from a chat.

Books

QuickBooks

Matching vendor statements against QuickBooks is proposed as a next job, not running yet.

1,000+

Your AI Employee can connect to more than 1,000 business apps through supported integrations. If yours isn’t a standard connection, we’ll review whether a custom integration makes sense. Additional installation fees may apply.

Guardrails

What it can touch, and what it can’t.

We map exactly what each AI Employee can access and act on before anything gets built.

Drafts

Drafts first

Replies to clients, carriers, and customers land as drafts. Nothing goes out until the owner of the account approves it.

Money

It flags the money, it doesn’t move it

It can flag differences for a person to check. It doesn’t pay, bill, or move money.

Seats

Every AI Employee answers to one leader

Each AI Employee connects only to the accounts its leader approves, and talks with that leader in a private chat.

Customers

Customer systems are your call

Connecting anything customer-facing is a decision made with your customers in mind, before it’s built.

Testing

Tested before it runs alone

Each seat runs a first shift with you watching, and new jobs are tested before they run on their own.

Reading

What it reads is data, not orders

Emails and documents are content to summarize. A message telling an AI to do something gets reported to you, never acted on.

How to start

Start with one seat. Add the next when it’s working.

Both of our logistics clients started with one AI Employee for one leader, then added seats.

01
Book the Launch Plan
Fifteen minutes to pick the first seat: the CEO, operations, finance, or onboarding.
15 minutes
02
Map the work
For bigger teams, a few working sessions to inventory your systems and pick the first jobs.
A few sessions
03
SN3 builds it
We set it up on that leader’s accounts and test it on real work.
About two weeks
04
First shift, then the next seat
A live launch session connects its accounts, and its first shift runs with you watching. When it’s working, the next leader gets theirs.
Launch day

Straight answers

Questions from logistics leaders.

Can every leader have their own?

Yes, and that’s how both of our logistics clients set them up. Each leader gets an AI Employee built around their own inbox, tools, and approvals, with its own private chat.

Does it work in Slack?

Yes. Our logistics clients use their AI Employees in Slack, Telegram, or both. One CEO connected his to Slack himself and has it send briefs, notifications, and alerts to separate channels.

Does it connect to ShipHero?

Not yet. We’ve discussed a ShipHero sync with a 3PL client whose own team planned to take the first pass, and nothing is built with ShipHero so far. What’s possible depends on your account’s access, so we confirm it on your Launch Plan call before we promise anything.

We run on Zoho. Does that work?

It can. A fleet-tracking client’s AI Employees are being connected to Zoho, and a deeper connection through the Zoho integration its CTO built is in progress. What’s possible depends on your account’s API access, and we confirm it on your Launch Plan call.

Can it talk to our customers?

Out of the box it drafts for the person who owns the account, and nothing goes to a client, carrier, or customer without their OK. One client’s leadership named a client-services AI Employee a top priority; it’s planned, not built.

What if it gets something wrong?

It happens, and we tune it with you. On its first day, one AI Employee told its user to prepare for a meeting that didn’t need to happen, because it read only the latest email instead of the earlier context. Drafts-first is why a mistake reaches you before it reaches a customer.

How do we start?

Usually with one seat, for one leader. Both of our logistics clients started with one AI Employee and added seats later.

Do I need another AI subscription?

Usually not. Your AI Employee can run on an AI plan you already pay for, such as ChatGPT, and switch to pay-as-you-go API usage if that plan runs out. Some clients run theirs on API usage alone. We’ll tell you which fits on your Launch Plan call, along with scope, setup, support, and investment.

The 15-minute Launch Plan

Start with one leader.

Fifteen minutes, no purchase required. We pick the first job, map where the work waits for your approval, and give you a straight answer on scope, timing, and investment.

Not sure where to start? Book anyway. Finding the first job is what the call is for.

Book my 15-minute Launch Plan

Opens our booking calendar in a new tab

You leave with a Launch Card

Workflow: The first job, mapped step by step
Tools: Their inboxes, workspace, and business systems
Approvals: Where the work waits for you
Success: How we’ll measure it, like hours back for that leader
Go-live: A date on the calendar