Most small businesses have more revenue forecasting capability than they realise. If you have a spreadsheet with 12 or more months of revenue history, you have enough data to produce a reasonable projection — and AI tools can help you both run the numbers and interpret what they mean.
This isn’t about building a complex financial model. It’s about taking data you already have, applying a few straightforward methods, and getting a defensible sense of where your revenue is likely to land next month. Here’s how to do it practically.
Start With Clean Monthly Data
Before any forecasting method will work reliably, you need clean monthly revenue data in a consistent format: one row per month, one column for revenue, no blanks, no merged cells, no footnotes mixed in with the data. If your current records are in an accounting tool, export them to a CSV. If they’re already in a spreadsheet, spend five minutes cleaning up any inconsistencies before you proceed.
The more history you have, the more reliable the forecast will be. Twelve months is a reasonable minimum. Two years or more lets you account for seasonal patterns. If you have less than six months of data, the output of any forecasting method will have wide uncertainty ranges — and you should treat the result as a rough directional indication rather than a reliable projection.
| Method | What you need | Output | Good for |
|---|---|---|---|
| Simple linear trend | Historical monthly revenue (12+ months recommended) | Straight-line projection based on average growth rate | Stable, steadily growing businesses |
| Moving average | Recent monthly revenue (6–12 months) | Smoothed projection that reduces short-term noise | Businesses with some variability but no strong seasonality |
| Seasonal adjustment | At least 2 years of monthly data | Forecast that accounts for predictable peaks and troughs | Retail, hospitality, seasonal services |
| AI-assisted forecast | Historical data + business context in your prompt | Narrative forecast with reasoning and scenario analysis | Any business wanting interpretation alongside numbers |
| Pipeline-based forecast | CRM deal data with close probability | Bottom-up forecast from specific deals | Sales-led businesses with active pipeline |
The Simplest Method: Linear Trend in Your Spreadsheet
The simplest forecasting approach requires nothing beyond your existing spreadsheet. In Excel or Google Sheets, the FORECAST.LINEAR function (or FORECAST in older versions) calculates a linear trend from your historical data and projects forward. To forecast next month: create a column with month numbers (1, 2, 3…), use your historical revenue as the y-values, and call the function with the next month number as the x-value.
This tells you where revenue would land if the historical growth trend continued in a straight line. It’s not sophisticated, but for businesses with steady growth it’s surprisingly accurate, and it takes less than five minutes to set up. The main limitation is that it ignores seasonality entirely — if your business has predictable peaks and troughs, the linear forecast will be systematically wrong in those periods.
Using ChatGPT or Claude to Interpret Your Data
Where AI adds the most value in revenue forecasting isn’t calculating the numbers — it’s interpreting what they mean and surfacing patterns you might miss. Paste your monthly revenue data into ChatGPT or Claude (or upload it as a file in ChatGPT’s Advanced Data Analysis) and ask: “Based on this revenue history, what trend do you see? Are there seasonal patterns? What would you estimate for next month, and what’s your confidence level?”
What you get back is more useful than a formula output: a plain-English interpretation of the trend, identification of any seasonal patterns in the data, a rough projection with explicit caveats about confidence, and sometimes a note about anomalies (months that were unusually high or low that might be distorting the trend). That interpretive layer turns a number into something you can actually use in a business conversation.
You can also ask for scenario analysis: “What if growth slows to half the current rate? What if we lose our largest customer, who accounts for roughly 20% of revenue?” The AI can walk through the implications of different scenarios in a way that a formula can’t. The caveats apply — this is pattern extrapolation, not prophecy — but scenario thinking is exactly where AI interpretation adds value over a spreadsheet alone.
⚠️ Honest Limitations of AI Revenue Forecasting
Pipeline-Based Forecasting for Sales Businesses
If your revenue comes from a sales pipeline rather than recurring subscriptions or direct transactions, the most accurate near-term forecast comes from your CRM rather than your revenue history. The approach: export your open deals with their estimated close dates and close probabilities, then calculate a weighted forecast — each deal’s value multiplied by its probability, summed for the target month.
You can do this entirely in a spreadsheet: a column for deal value, a column for probability, a column for expected close month, and a SUMIF formula that sums the probability-weighted values for next month. Ask ChatGPT to write the formula if the multi-condition logic is unfamiliar. The result is a bottom-up forecast grounded in your actual pipeline rather than a statistical extrapolation from history.
Combining Both Approaches
The most reliable near-term revenue forecast for most small businesses combines both methods: a historical trend forecast as a baseline, and a pipeline-based forecast as a reality check. If your trend model says $90k next month and your weighted pipeline shows $65k of expected deals, the gap is a useful signal — either you’re missing deals that will close, or the historical trend is overstating the near-term reality.
When the two approaches diverge significantly, that divergence is worth understanding before you plan around either number. Ask your AI tool to help you reconcile them: “My historical trend suggests $90k, but my pipeline only shows $65k weighted. What might explain this gap, and which number should I plan around?” The answer varies by business, but working through the question usually surfaces something useful.
The Role of Judgement in Forecasting
Every forecasting method described in this article extrapolates from historical patterns. None of them can account for events your data doesn’t contain: a new competitor entering your market, a key partnership you’re about to close, a product launch that will change your revenue mix, or a macroeconomic shift that changes buyer behaviour. The numbers give you a baseline; your judgment about what’s different now compared to the period your data covers is what makes the forecast actually useful for planning.
Build your forecasting habit to include a brief qualitative note alongside the numbers: “This forecast assumes current sales activity levels continue. It doesn’t account for the new enterprise deal we’re targeting, which could add 15–20% upside if it closes.” That context transforms a number into a planning tool and makes it far more useful for the decisions that actually depend on it.
Sharing Forecasts With Stakeholders
A revenue forecast is most useful when the people who receive it understand what it’s based on and what it doesn’t account for. When sharing forecasts with partners, investors, or team members, include a brief explanation of your method: “This is based on a linear trend from the last 18 months of monthly revenue. It assumes current growth rate continues and doesn’t account for the new product launch in Q3.” That transparency prevents overconfidence in the numbers and makes it easier to update the conversation when new information changes the picture.
Making It a Monthly Habit
A revenue forecast you update once and ignore isn’t useful. The value comes from updating it every month, comparing the forecast to actual, and understanding the variance. When actual revenue comes in higher or lower than the forecast, ask why — and update your forecasting inputs based on what you learn. Over six months of consistent practice, most small business owners develop a noticeably better intuition for where their revenue is heading, which is the real goal. The spreadsheet and the AI are tools for building that intuition faster, not substitutes for it.