Advisory for distributors and manufacturers
The decision that hurts cash is rarely the write-off. It is the purchase approval that made it inevitable. Sales are fine, margin is fine, and the credit line keeps climbing, because nobody can see which buying decisions are holding the cash.
Find the cash that is already stuck, and test the next commitment before it is signed.
What you get
The cash sitting in inventory, in dollars, measured the same way on day 1 and on day 90, so the change is real.
Which buying decisions put it there, so next year's cash does not go the same way.
The commitment on the desk, tested first: the assumptions named, the week the cash leaves, and what it removes if it is wrong.
Why this works
Crestline Supply is a constructed $36M distributor, built so the method can be shown end to end with the data published. Its reports already said which items were stuck. None of them could say which kind of decision put them there. Once every purchase was traced back to the decision behind it, vendor programs turned out to be 15% of purchase dollars and 39% of trapped cash. 40 cents of every vendor-deal dollar was still on the shelf.
Read the worked case Test a purchase order before it is signed
How it works
Anyone can ask AI a question. This is a workflow designed end to end, from the raw exports to the decision: the same steps, the same rules and the same checks every time, so the result can be repeated, compared month to month, and defended.
Products, purchase orders, sales and stock on hand, from whatever system is in place. 12 months minimum. Messy data is fine. It is the normal starting point.
It cleans and joins the 4 exports. It works out why each purchase was made. It traces every dollar on the shelf back to the order that bought it. It measures what is trapped, ranks it by dollars, and drafts the report. Each step is tested against cases where the answer is already known, and it runs the same way next quarter.
The trapped cash in dollars, the decisions behind it, and the list that changes next month's buying.
Time from the team: About 4 hours across the engagement, mostly one kickoff and one readout.
This is how I run a diagnostic, with my own workflow, to get you the answer. Building a workflow inside your own business is the AI workflow implementation below.
Why now
Every finance team knows the questions worth asking about inventory. Which purchases are still sitting on the shelf, and who approved them? Which product lines give the cash back quickly, and which hold it for a year? What happens to cash if the next big order goes ahead? They go unasked because answering them meant reading every purchase order, every sale and every stock line, and nobody has those weeks. So the team looks at the top 20 items and hopes the other 4,000 behave. An AI workflow reads all of it, every item and every order, the same way each time. The answers arrive before the next buying meeting, and again next quarter without starting over. Your competitors are still looking at the top 20.
How it is measured
Trapped cash, identified and priced on day 1. The definition is written down, and the baseline is page 1 of the report. This is what I promise.
How much of it has come back, measured on the same items, the same way, at day 90. Your team's actions release it, and I measure it.
Purchases not placed because of the action list. Always labelled as an estimate.
Ways to work on it
Pick the one that fits. Its price, scope and timeline open underneath.
Written findings: Full Diagnostic
The true inventory position, the cash and exposure in dollars, and a plan ranked by cash released: a 10 to 15 page report with a one-page summary for a board or a bank.
Written findings: Focused Diagnostic
One purchase or one decision, named on the order. A 3 to 5 page report: whether the numbers behind it hold, what it costs in dollars if they do not, and the first fix. Plus a 30-minute readout. Not included: a second decision, or doing the fix.
Built and handed over: AI workflow implementation
One deliverable: a working AI workflow inside the business, run by its own team. One workflow, named at the start.
Scope and timeline
In scope: one workflow, named at the start, such as the inventory close, the buying review or the monthly report, plus the handoffs it depends on. Anything beyond that workflow gets its own quote. Most of the 3 months goes to mapping how the work really gets done, with your people, so the AI fits reality.
Coached: AI workflow build sessions
Live coaching while the team builds, on its own data, in its own Claude. One session unblocks one thing. 4 sessions build the whole workflow.
Sharon is an insightful, kind, and hyper-detailed analyst. She quickly recognizes core issues and provides a clear path to the answer ... she has developed a powerful intuition and sense of clarity in a very difficult subject matter: cash and inventory optimization.
Mark LuptonCFO and advisor to 8-figure companies. Founder, Greenhouse Business AdvisorsShe reads financial reports quickly and spots patterns, opportunities, and potential issues that others might miss, turning insights into actionable strategies. Her ability to notice hidden details and identify root causes instead of just surface issues makes her invaluable.
Charles TuCEO, technology manufacturer, TaiwanSharon is knowledgeable and approaches a situation from a real-world perspective.
She doesn't critique outputs, she diagnoses the system behind them: backtesting, train/test splits, auditability. That's the rigor senior finance professionals actually need. She's the rare coach who can speak to both the finance and the AI side at a level senior professionals respect.
She brings positive energy, strong subject matter expertise, and practical AI insights to every interaction. Her feedback was always constructive and actionable, helping participants elevate their thinking.
Start here
Free, if you want to look first
