AI in Action: Time
51% of Small Businesses Save 6+ Hours Each Week With AI.
The stories in this chapter are fictitious and meant to illustrate concepts. Any resemblance to actual companies, projects, or individuals is an unintentional coincidence.
The CFO Stopped Doing $15/Hour Work
David was the CFO of Meridian Manufacturing, a precision metal parts company in Ohio with about $5 million in annual revenue. David's job was to track every dollar, make sense of the numbers, and present them in a way the CEO could act on.
The problem was how David spent his time. Every month, he collected expense receipts from a dozen employees. Each receipt needed to be categorized, matched to a project or cost center, and entered into a spreadsheet. It consumed hours every week.
Once the numbers were compiled, David spent even more time building monthly financial presentations for the CEO. Slide after slide of charts, tables, and trend lines. The information was critical for decision-making, but the process of turning raw numbers into polished slides was pure drudgery. David was spending the majority of his week on tasks that could be done by someone making $15/hour. The analysis and judgment that justified his $150/hour expertise kept getting squeezed into whatever time was left over.
David started using AI to handle the front end of his workflow. He already had scanned receipt PDFs organized in a folder. He pointed AI at those files, and it extracted the vendor name, amount, date, and category from each one, then pre-populated his Excel expense tracking template. David still reviewed every entry before it was finalized. He caught the occasional misread, such as a smudged receipt or an unfamiliar vendor name, but reviewing and correcting took a fraction of the time that manual data entry had required.
The presentations were an even bigger shift. Once David had validated the expense data in the spreadsheet, he fed AI that finalized Excel file along with the prior month's slide deck and asked it to generate an updated version. AI identified which trends had changed meaningfully and deserved emphasis, which metrics were tracking in line with expectations and could be summarized briefly, and where a new chart might tell the story more clearly. David reviewed every slide and adjusted framing when AI's emphasis didn't match his read of the situation. But the starting point was most of the way there. He was refining rather than creating from scratch.
The result wasn't that David worked fewer hours. It was that he worked on different things. The hours he reclaimed from data entry and slide building went into the work the CEO actually needed from a CFO such as analyzing whether a new production line would pay for itself, modeling the impact of a raw materials price increase, evaluating whether to bring a machining process in-house or continue outsourcing it. The kind of work where a wrong answer could cost the company hundreds of thousands of dollars, and where David's experience and judgment were irreplaceable.
The Property Manager Got Her Mornings Back
Rachel owned Bridgewater Property Group, which managed roughly 200 residential units across Milwaukee which included apartment complexes, duplexes, and single-family rentals. She was proud of her reputation for being responsive and fair, which kept both tenants and landlord clients loyal.
But the cost of that responsiveness was Rachel's time. Every morning she opened her inbox to find dozens of tenant emails. Some were urgent such as a burst pipe or a heating failure in January. Most were routine such as lease renewal questions or follow-ups on maintenance requests. Each one needed a response that was accurate, professional, and consistent with Bridgewater's policies and the specific terms of that tenant's lease. Answering them required deep familiarity with the company's operations, or time-consuming lookups across multiple files and systems. Either way, it consumed most of her morning.
So Rachel set up an AI system connected to her company's documents: every lease agreement, vendor contract, maintenance log, property policy manual, and tenant communication history. This setup used what's known as retrieval-augmented generation, or RAG, which is essentially a way to give AI access to a specific library of documents so its answers are grounded in actual company information rather than general knowledge. When a tenant emailed asking about their pet deposit, the AI didn't guess. It pulled the relevant lease, found the pet policy clause, and drafted a response citing the specific terms. And because the AI also had access to Rachel's years of prior email responses, it matched her tone and communication style.
Every morning, instead of forty emails that each required ten minutes of research and careful drafting, Rachel now opened a queue of AI-drafted responses. Each one referenced the relevant lease terms or policy documents. She read each draft carefully. Sometimes she sent it as written. Sometimes she softened the tone or added a personal note. Occasionally she rewrote a response entirely when the situation was sensitive enough to warrant her full attention.
The key was that Rachel never let a response go out without reviewing it. The AI drafted; Rachel decided. Every tenant still received a response that Rachel had read and approved. The difference was that instead of spending three hours each morning composing those responses from scratch, she spent forty-five minutes reviewing and refining them. The rest of her morning was now available for the work that actually grew her business such as meeting with prospective landlord clients, negotiating vendor contracts, and visiting properties.