Business owners wear too many hats. When time is scarce, they often delegate strategic decisions to the first “tech person” who appears competent—sometimes an individual whose expertise stops at replacing hard drives and rebooting routers. The result? Ill-informed advice, misunderstood reports, and avoidable hits to cashflow. Before you reach the end of this series you’ll decide for yourself whether the owner in our live case study acted wisely or walked into a trap. First, let’s level the playing field.
Two Kinds of Owners
1. Owners who work on their business
They treat the company as a product that must itself be engineered, tuned, and scaled. In addition to:
- Hunting for fresh growth levers—new marketing channels, smarter inventory controls, emerging data tools.
- Reading voraciously to stay ahead of market shifts.
…they also:
- Delegate day-to-day firefighting so their calendar reserves space for strategy.
- Track a short list of high-impact KPIs and revisit them weekly.
- Invest in talent and systems that will survive without their constant supervision.
2. Owners who work in their business
Immersed in daily operations, they rarely step back to view the broader landscape; they “can’t see the wood for the trees.” Common symptoms include:
- Personally chasing late supplier deliveries.
- Approving every petty-cash purchase “just to be safe.”
- Micromanaging the sales counter instead of modeling demand.
- Reconciling bank statements at 10 p.m. because “nobody else can do it right.”
- Acting as ad-hoc tech support when the label printer jams.
Why Today’s IT Labels Are Misleading
In the 1990s a “computer consultant” soldered memory chips in the morning and wrote FoxPro code that night. Fast-forward to 2025 and the field is unrecognisably specialised. A hardware reseller may think a modern AI-driven inventory model is just an expensive spreadsheet macro. Conversely, many brilliant software engineers could not identify a failing power supply if it smoked on their desk. The knowledge silos are real—and dangerous when crossed without caution.
The ERP Minefield
Enterprise-Resource-Planning vendors confront three structural hurdles:
- Legacy-data migration on shoe-string budgets. The old schema’s inefficiencies slide straight into the shiny new database.
- Ageing code bases. Systems architected in the 2000s never anticipated machine-learning pipelines or real-time API orchestration.
- Accounting-centric implementations. Consultants nail the debits and credits yet overlook schema design that turns raw records into decision-ready insight.
Python, TensorFlow, ChatGPT and the rest have vaulted AI from ivory-tower theory to shop-floor necessity—but most mid-market ERPs still behave as though predictive modelling were science fiction.
Setting Up the Case Study
In the next article we’ll meet an owner who, pressed for time, asked a respected (but purely hardware) technician to evaluate an AI model designed to optimise inventory. The technician assumed that yesterday’s Point-of-Sale software already exposed every lever the business needed and dismissed the model outright—without recognising the cashflow turbulence hiding beneath static reorder points and blunt ABC codes. ChatGPT-style contextual reasoning—which could have revealed seasonality, cannibalisation, and dead-stock risk—never entered the conversation.
Read on, weigh the evidence, and decide whether the owner safeguarded or sabotaged their own balance-sheet.