Everyone has been sharing ChatGPT “slash commands” lately — /flowchart, /infographic, /matrix, /roadmap. Type one word before your question and ChatGPT changes how it answers: it draws instead of writes, plans instead of lists, compares instead of describes.
Most people point these commands at holiday plans and study notes. We pointed them at a live SAP Business One system.
This guide gives you every prompt from our 8-episode series, word for word, so you can run them on your own system. Each one opens with a slash command, runs on the same ChatGPT subscription you may already pay for, and reads live data through TEKAI — no report writer, no developer, no export to Excel.
📄 Download the full reference: The Slash Command Reference for SAP Business One — 13 pages, 200+ commands across 20 function areas. Jump to the download.
Watch the series: ChatGPT Prompts for SAP Business One — full playlist
Three things, once:
After that, every prompt below is one line typed into ChatGPT. The word before the colon is the slash command. The words after it are your question.
A note on wording. Throughout, we say “system”, not “database”. The AI is not querying tables — it reads the business documents and master data the way a consultant would, and it names them the way you do: invoices, deliveries, BOMs, purchase orders.
/cheatsheet — Your CEO brief, every morningcheatsheet: CEO morning business brief
What comes back: the number that matters most first (in our demo, over $10 million in receivables, almost all overdue past 90 days), the customers behind it, working capital (receivables against payables), negative and stuck inventory, a data-freshness alert so you never manage on stale numbers, and five questions to ask your team this morning.
Then automate it:
Can you send this to my email?
Make this a daily task at 8 AM.
Two sentences. The brief now lands in your inbox every morning, connected to the live system. The system reports to the owner — not the other way around.
/matrix — Which customers are risky?matrix: customer sales vs collection risk, last 5 years
What comes back: every customer plotted in four quadrants — sales against collection risk — scored on collection percentage, overdue exposure and payment delay, with the scoring method shown. In the demo, one large account carried most of the risk at 78% collected. The AI then suggests what to add for production use (credit limits, DSO, concentration risk) and offers to refresh the matrix monthly.
Then automate it:
Refresh this matrix on the 1st of every month and email it to me.
/timeline — How a customer really paystimeline: customer 360 for [customer name], invoice to payment history
What comes back: each invoice linked to the real incoming payment — invoice date, due date, payment date, days to collect — with the current open, overdue and promised position underneath. Settlement history, not an average.
Then act on it:
Send this customer their statement of account.
Change this customer's credit terms to 30 days.
/flowchart — Why was this delivery late?flowchart: why is order [number] delayed?
What comes back: the document trail — order date, promised date, stock at the time, delivery posting, invoice posting — drawn as a flow, with the delay classified (in the demo: an operational fulfilment delay, delivery posted 12 days after the promise and invoice 41 days after that; stock was never the problem). It also tells you what the system cannot prove — the human reason — so you know exactly which question to ask.
Then automate it:
Alert me on any open sales order about to miss its promised date.
/explodedview — What’s blocking production?explodedview: BOM + stock, shortages and incoming supply
What comes back: every active bill of materials opened out — parent item, components, warehouse stock, committed quantity, shortage, open purchase orders — and the one constraint that matters. In the demo, a single component was blocking four production lines at once.
Then act on it:
Create purchase requests for the shortages.
/comparison + /infographic — Month-end review in one shotcomparison: customer vs vendor vs product concentration
One table: how concentrated sales are in your top customers, how dependent purchasing is on a few vendors, which products actually carry the margin.
infographic: monthly business review
One management visual for the latest month — sales, purchases, gross profit, receivables, payables, inventory exposure, top movers — against the prior month. When a month has no activity, the AI says so and anchors on the last real month instead of inventing numbers. That honesty is the difference between a tool you check and a tool you trust.
/roadmap — A 90-day improvement planroadmap: sales and collections improvement plan
What comes back: a phased plan built from the current position. Days 1–30: make receivables trustworthy and chase the biggest collectible balances. Days 31–60: tighten credit control at the order. Days 61–90: restart systematic sales growth. Each phase names the metric it moves.
Then push harder:
Collections or inventory — which plan should I start first?
Give me this week's actions as a task list.
/cycle — Live procure-to-paycycle: procure to pay
What comes back: purchase request → quotation → purchase order → goods receipt → quality check → invoice → three-way match → payment → vendor performance, drawn from the live system with today’s counts on each step.
Then automate it:
Watch this cycle and only show me the exceptions: overdue POs, partial receipts, goods received but not invoiced, price mismatches, unpaid vendors.
Everything else runs without you.
These didn’t make the series but work the same way.
/beforeafter — Did the price change help?
beforeafter: sales volume and margin for [item group], 3 months before and after the price revision on [date]
/pyramid — Where does the profit come from?
pyramid: customers by gross profit contribution, last 12 months
/stickynotes — What does my team need to do today?
stickynotes: open items by owner — overdue receivables, late deliveries, pending approvals, stock shortages
Does this need a special version of ChatGPT?
No. A standard paid ChatGPT plan that supports connectors. TEKAI also works with Claude, Perplexity and other AI models — the prompts are the same.
Does it change anything in SAP Business One?
Not by default. TEKAI is read-only until you switch on specific write actions for a company (for example, creating purchase requests or sending statements). No add-on is installed in the SAP client.
Will it work on my version of SAP Business One?
Yes, on standard SAP Business One systems, SQL or HANA, on-premise or cloud.
What does it cost?
TEKAI is a monthly subscription per company — see pricing. You bring your own AI subscription.
The eight commands above are the ones we filmed. The full reference goes much further — over 200 commands across executive, sales, customer 360, collections, inventory, procurement, production and finance, plus the end-to-end cycles, the /dashboard → /xray → /roadmap diagnosis pattern, exception management, and the command syntax that combines them into a single line.
Enter your work email below to download. Business domains only — we don’t accept Gmail, Outlook or other personal addresses. If you want the background on how the AI reads your system, start with AI in SAP Business One.
Every prompt on this page ran on a standard SAP Business One system, with the AI we already pay for. The fastest way to know what it does on yours is to bring your own questions to a 30-minute live demo.
Book a live demo — bring your own questions
TEKROI is an official SAP Business One partner (SAP PartnerEdge — Sell & Service) with 350+ implementations across India, Nepal, the Middle East and Africa. TEKAI is TEKROI’s AI suite for SAP Business One.
13 pages. 200+ commands across 20 function areas — executive, sales, customer 360, collections, inventory, procurement, production and finance — plus the end-to-end cycles, the /dashboard → /xray → /roadmap diagnosis pattern, and the command syntax that combines them into a single line.
Your download is ready — we have emailed you the same link.
Download the PDF →This link is valid for 15 minutes.