Jason Hope, CPA, CFA, began thinking differently about accounting technology while working in private equity. He was assigned full-time to one portfolio company and spent the early months cleaning up the work he inherited. Then the job changed. Once the processes were in place, much of the accounting became routine and needed less of his attention.
That experience shaped his view of how smaller companies should pay for finance work. A business may not need a full in-house team when an experienced outside professional can establish the process, monitor it, and step in where judgment is required. Hope bills by the hour currently, though he expects his firm to move toward a fixed-fee arrangement. He said that could make monthly costs more predictable for clients while letting the firm share the benefit of greater efficiency.
AI has not produced a dramatic cut in hours at Hope Financial Consulting, Hope said. The more immediate change is what the firm can deliver within those hours. Its monthly work can include finer detail, clearer financial visuals and more useful analysis without increasing the client’s cost.
Finding The Transaction That Does Not Fit
Reconciliations and variance analysis are tedious because the clue may be one transaction buried among thousands. Hope uses AI to review general ledger activity and compare transactions across months. The software can surface an inconsistency for a person to investigate instead of asking that person to find it manually.
Forecasting benefits from the same ability to sort a large amount of data. Hope’s firm also uses AI in budgeting and financial planning and analysis. His team prepares a cash flow forecast every week that looks 12 to 16 weeks ahead. The team can then change assumptions and show what different choices may mean for cash and return on investment. An owner considering a hire, a new market, a change in prices, or another financing option can see the possible consequences before committing.
Hope puts the questions plainly: “Where is the money going? Where in my business am I winning and where am I losing? What will my cash look like in various scenarios?”
A real estate client faced that last question. The company had committed capital from investors, but a preferred return would begin accruing as soon as it called the money. The owner wanted to wait without risking a cash shortage or missing payroll. Concern about the bank balance nearly prompted an early capital call.
Hope’s team prepared a cash flow analysis showing that the company had enough cash through the end of the year. It identified the lowest projected balance and the signs that would indicate it was time to call the next tranche. The owner could decide how much cushion to keep instead of reacting to the current bank balance alone.
Some warnings can appear even earlier. Hope cited margin pressure in current sales quotes. Quote details and accepted order terms can reveal an adverse trend while sales are still being made. The business does not have to wait for the next monthly or quarterly close to see it.
Showing A Prospect The Work
AI also appears early in Hope’s sales process. With a prospect’s permission, the firm reviews detailed general ledger data and financial statements. Its internal process looks for bookkeeping problems the prospect may not know about. The same historical data can populate a sample interactive dashboard that resembles the firm’s monthly deliverable.
The prospect gets more than a description of the service. It can compare the sample with the reports it already receives and judge whether the added detail would help its leaders make decisions.
The Books Still Need A Human Gatekeeper
Hope draws a firm line at unsupervised transaction entry. Day-to-day accounting still requires manual work, and he would not trust AI to record entries without a person checking them. A bad entry may sit unnoticed in the books and weaken every forecast or analysis built on top of it.
He sees a better division of work. AI can digest historical records, identify variances and trends, and offer predictions with confidence levels. A senior finance professional must decide whether the assumptions make sense, whether the analysis needs more information, and whether the conclusion matches what the records support.
“We read and review everything AI does,” Hope said.
The controls also extend to the information placed in an AI system. Hope said the firm uses enterprise-level tools that do not train the model on customer data. It does not upload employee information, Social Security numbers, bank account numbers, or other identifying numbers. When a client asks for additional separation, the firm removes customer and client names from the material used in the AI work.
Questions To Ask Before Sharing Financial Data
Business owners should look past the phrase “powered by AI” when choosing an accountant or fractional CFO. Hope recommends asking whether the professional is licensed and whether financial data will be shared outside the organization or the country. He also suggests asking to see a deliverable made with historical company data.
Those questions reveal who is accountable for the work, where sensitive information may go, and whether the technology produces anything useful. A polished claim about AI does not answer any of them.
More Capacity, With No Reliable Long-Range Forecast
Hope expects AI to change how bookkeepers, accountants and fractional CFOs use their time rather than remove them from the work. He compares it with the arrival of spreadsheets and the internet. Each changed accounting, but the result was more work for professionals who learned to use the new tools.
He is much less certain about the next three to five years. His own work had changed sharply in the previous six months, he said, and he would not pretend to know what comes next.
The useful choice for a small business is available now. Routine financial data can support cash forecasts, margin checks, scenario models, and dashboards built for decisions. AI can do much of the sorting. An experienced person still has to protect the data, test the assumptions, and stand behind the answer.
