Automation for logistics, distribution, and manufacturing
Your best people are
holding the process together.
The re-keying. The chasing. The reconciling. I map the manual workflow, price the hours it consumes, and automate the highest-cost step — starting with a 90-minute session, before you buy another system.
No free discovery calls. The 90-minute strategy session is the first step — $1,500, credited toward the two-week diagnostic.
The gap
The work is manual. The cost is invisible.
Most operators do not have an AI problem. They have a workflow problem — and buying a tool before mapping it is how the tool becomes another queue.
You already know where it hurts. What you do not have is the hours and the cost per transaction attached to it, in one document, with a ranked fix.
The work arrives as email, PDF, and phone calls
Orders, carrier packets, invoices, and spec sheets get re-keyed because no system owns the front door.
The handoff is the cost
Quote to order. Order to dispatch. Dispatch to invoice. Each handoff has an owner, a queue, and a rework loop.
The 5% that breaks eats the margin
A lapsed certificate, a PO that arrives by email, an accessorial with no proof. The exception is where the hours go.
The diagnostic
Two weeks. One document. A decision.
Most operators start with the 90-minute session. The diagnostic is the next step, and the session fee is credited.
Workflow map
Intake → process → handoff → exception → audit, with volumes, owners, and the manual steps named.
Manual-cost model
Hours per transaction, error and rework, cycle time, cost per transaction. Every assumption labelled.
Automation stack review
What you already own — ERP, TMS, WMS, OCR, RPA — what it actually does, and where it stops.
Vendor and tooling options
OCR/IDP, RPA, workflow, ERP add-ons. Build vs buy vs partner, with real prices and trade-offs.
90-day roadmap
Ranked by payback. The first pilot is scoped, with the baseline and the success measure agreed.
Go / no-go
Including “do not automate this yet” if that is the honest answer. The document is yours either way.
WHAT IT IS NOT
- A tool demo
- A strategy deck
- A staff-augmentation contract
- A free audit
THE GUARANTEE
If the diagnostic does not identify at least two automation opportunities with a payback under 12 months, I refund it. The $1,500 session fee is credited toward it, and 50% of the diagnostic fee is credited toward a sprint signed within 30 days.
Already know you need the diagnostic? Request it directly.
The ladder
Diagnose. Fix. Hand off.
The diagnostic produces the scope for everything after it. I do not quote a sprint before I have seen the workflow.
Paid Strategy Session
$1,500
A working session on your workflow: recorded, one-page memo, three prioritized recommendations. The fee is credited toward the diagnostic if you sign within 14 days.
Automation Diagnostic
$12,500
Workflow map, manual-cost model, stack review, vendor options, 90-day roadmap, go/no-go. The core offer.
Automation Sprint
$25K–$50K
One workflow automated end to end: OCR/IDP, RPA, API, human review, QA, and handoff. Scoped by the diagnostic. I architect and QA; a delivery partner builds.
Fractional Automation Ops
$5K–$8K/mo
Vendor management, QA, cost and quality, roadmap. After a sprint — not instead of one.
Fixed fee. No hourly billing. The $1,500 session fee is credited toward the diagnostic if you sign within 14 days. The tool is secondary — OCR, RPA, API, or a model, chosen after the workflow map, not before it. I do not sell software or take referral fees.
Fit
A strong operation with a manual seam.
Ops owns the problem. IT owns the systems. The diagnostic gives both a written map and a prioritised backlog — which is why it is usually the COO who forwards it to the CIO.
A FIT
- +Logistics and freight, distribution and wholesale, manufacturing, construction and building materials, facilities and field services
- +$10M–$250M revenue, 100–1,000 employees
- +Multi-site or high transaction volume
- +A CIO, VP IT, or COO with an IT lead
- +At least one process that still runs on email, spreadsheets, or a person who “knows how it works”
NOT A FIT
- −Software companies — that is the other practice
- −Pure strategy work with no access to systems or data
- −Staff augmentation or hourly development
- −A business with no one who owns the systems
- −Buying a tool before mapping the work
Method
Five steps. No black box.
Workflow map
Intake → process → handoff → exception → audit, with volumes and owners for each step.
Manual-cost model
Hours per transaction, error and rework, cycle time, cost per transaction. Every assumption labelled.
Vendor and tooling selection
OCR/IDP, RPA, workflow, ERP add-ons. Build vs buy vs partner, with prices and trade-offs.
Pilot
One workflow, measured against a baseline agreed before the work starts. No pilot without a measure.
Handoff
Runbook, cost model, failure modes, named owner. Your team keeps the system.
Straight answers
The questions a COO actually asks.
“We already have RPA.”
RPA automates the happy path. The cost is in the exceptions — the bot stops, and a person finishes. The diagnostic maps where it stops and what that costs.
“Our data is messy.”
Then the first fix is intake and normalisation, not a model. That is scoped as step one, not hand-waved.
“We tried a pilot.”
Most pilots automate a task and leave the handoff manual. The diagnostic maps the handoff and names the owner before anyone buys a tool.
“We have no IT team.”
Then the output has to be tools your ops lead can run, with a runbook. If you need a build, I bring a delivery partner and stay accountable for the architecture.
Why me
Operator first. Architect second.
I spent nine years in Amazon operations before I moved to the systems side — then built the classification and exception layer for a platform doing 2M requests an hour.
That order matters. It means I design for the person doing the work and for the cost of running the system, not just for the benchmark.
Nine years in operations
Operations, programme management, and product at Amazon. I have been the person who has to make the process work at 6am, and the person who has to explain why it did not.
Throughput, quality, and cost
100,000+ classes classified, accuracy up 35%, processing time halved. More than $20M in savings delivered. The 2M requests/hour system held at 99.99% because the exception layer was designed, not hoped for.
No software to sell
No referral fees, no partner programme, no reseller margin. Where the right answer is a product you buy without me, I say so in writing, in a document you keep.
No borrowed logos. The Amazon work is labelled; client automation work is published only with permission. Chennai, India · 8–11am ET overlap · 16 years working with US teams.
Start
Start with 90 minutes, not a software purchase.
Two weeks, $12,500, a written document: where the manual cost is, what it costs per transaction, what to automate first, and a 90-day plan.
I reply within 24 hours, Monday–Friday, with a one-page scope and a start date — or a clear no with the reason.
WHAT TO SEND
- 01Company and your role
- 02The manual workflow
- 03Volume per week
- 04Systems involved
- 05Where it breaks
- 06Timeline
For software teams and AI systems, this is the other practice: AI that survives production →