— Process improvement · Lean Six Sigma
Process improvement consulting: fix it, then automate it.
Automating a broken process makes it fail faster. We map the work, measure where it breaks and redesign it from first principles, then build the lakehouse pipeline or AI workflow on the version that works.
30+ years inside IBM · EY · J&J · McKesson
Lean Six Sigma Master Black Belt
AIMSolve from $2,500
60-Day Ship Guarantee
— The pattern · Three failure modes
Three ways automation makes things worse.
01
The workarounds break
The old workflow had workarounds nobody wrote down. Automate it as it stands and the workarounds break, the exceptions pile up, and the team goes back to spreadsheets.
02
No baseline, no verdict
Without a baseline, nobody can say whether the new pipeline or agent helped. The project gets judged on opinion, and opinion rarely renews a budget.
03
Defects come back
A data quality sprint cleans the tables, then the same defects return within a quarter, because the process that produced them never changed.
— The work · What we do
Process improvement consulting that holds after we leave.
Lean Six Sigma and first-principles redesign come first. Only then do we decide what to automate, and the automation lands on the lakehouse with a control plan so the gain holds.
Value stream mapping
We walk the process with the people who run it and map every hand-off, wait and rework loop, so the real workflow is on paper before anything changes.
Measured baseline
Cycle time, defect rate and rework measured on real work, so every later change is judged against a number, not an opinion.
First-principles redesign
We ask what the workflow would look like if you designed it today, and remove steps before automating any of them.
DMAIC for data pipelines
Define, Measure, Analyze, Improve and Control applied to the pipeline itself: defect rates per source, late loads and reconciliation breaks.
Quality checks on the lakehouse
Improvements land as Lakeflow jobs and data quality checks on the lakehouse, so the fix is enforced by the platform, not by memory.
Control plan
Owners, measures and review dates written down, so the improvement is still holding a quarter after the project closes.
— Why the process comes first
The platform reproduces whatever the process does.
A pipeline over a broken hand-off moves bad data faster, and an agent built on it repeats the mistake at scale. That is why we measure the process before we touch the lakehouse, and why the redesign and the automation ship together.
The long version is our field guide, don’t automate chaos, and the cost of skipping it is set out in deploying AI in broken processes.
— The method · AIM-IT
DMAIC improves the process. AIM-IT ships it.
Assess, Innovate, Model, Implement, Track. AIM-IT carries a Lean Six Sigma improvement into a data or AI build, and each phase ends in a named deliverable.
01 · ASSESS
Assess
Current-state map and baseline
Walk the process, measure the baseline, and find where the defects actually start.
02 · INNOVATE
Innovate
Future-state workflow
Redesign from first principles, removing steps before automating any.
03 · MODEL
Model
Pilot results against baseline
Pilot the new flow on real work and compare it to the baseline in numbers.
04 · IMPLEMENT
Implement
Rolled-out process and automation
Roll out the redesigned process and the automation together, with the people who run it.
05 · TRACK
Track
Control plan
Owners, measures and review dates, so the gain holds.
— The deliverable · What you get
What you own at the end.
- A current-state process map with the measured baseline.
- A redesigned future-state workflow, with steps removed before anything is automated.
- An automation or lakehouse pipeline scope built on the redesigned process.
- A written control plan: what is measured, how often, and who owns it.
- Your team able to run the improvement cycle without us.
1 to 2 wks
Typical first build
5 phases
AIM-IT, end to end
$2,500+
Starting price
60-Day
Ship Guarantee
On every Sprint engagement: if we don’t deliver the agreed working artifact in 60 days, you don’t pay the final invoice.
— Who does the work
Led by a Lean Six Sigma Master Black Belt.
Frank “Rio” Shines spent 30+ years implementing enterprise systems inside IBM, Ernst & Young, Johnson & Johnson and McKesson. He runs the assessment himself, on your real process, with your team, so they learn the method by doing it.
When the redesign needs data behind it, the pipelines are built under data engineering, and the AI that reads it under AI on the lakehouse.
— Related reading · From the blog
Go deeper on process before platform.
Field guide
Don’t automate chaos: a field guide to AI process improvement
How to fix the workflow so the automation actually sticks.
Method
Lakeflow through Lean Six Sigma: a repeatable method
How DMAIC and Lakeflow fit together on the lakehouse.
Data quality
A data quality framework for the lakehouse
Expectations, DQX and DMAIC applied to pipeline defects.
Cost control
Databricks cost governance with system tables
Tags, budgets and chargeback measured from system tables.
Framework
AIM-IT: the 90-day framework from assessment to production
The five phases, and what each one delivers.
— Straight answers · The objections answered
Frequently asked questions (FAQs)
Who leads the work?
Frank “Rio” Shines, a Lean Six Sigma Master Black Belt with thirty years of delivering enterprise systems inside IBM, Ernst and Young, Johnson and Johnson and McKesson. He runs the assessment himself.
Is this a Six Sigma training programme?
No. We run the improvement on your real process with your team, and they learn the method by doing it. If what you need is belt certification, that is a different engagement and we will say so.
Why fix the process before a lakehouse or AI project?
Because the platform will faithfully reproduce whatever the process does. A pipeline over a broken hand-off moves bad data faster, and an agent built on it repeats the mistake at scale. The pattern is set out in why AI pilots fail.
What does AIM-IT add to DMAIC?
DMAIC is how we improve a process. AIM-IT (Assess, Innovate, Model, Implement, Track) is how we carry that improvement into a data or AI build, so the redesign and the technology ship together.
What does process improvement consulting cost?
AIMSolve starts at $2,500 for a fixed-scope workflow charter and first build. Larger programmes are quoted after the assessment, because an honest number needs to see the process first. Every engagement carries the 60-Day Ship Guarantee.
Start with the process, not the tool.
One call, your real workflow on the table, and an honest read on what to fix before anything gets automated on the lakehouse.