Ergora's AI Forecasts tab uses Google's TimesFM to turn a metric's monthly history into a six-month outlook, with an 80% confidence band around it. Every workspace sees sample forecasts across eight business areas straight away, and forecasts built on your own numbers appear once you connect the right tools. This guide covers what you see, what the forecasts are built from, how often they update and where their limits are.
What TimesFM is
TimesFM is a time-series foundation model released by Google Research. It is pre-trained on a large collection of time series, so it can forecast a series it has never seen without being trained on your data first. Ergora runs the 200 million parameter model on its own servers and feeds it monthly history for a metric. It returns a point forecast for each of the next six months, plus a lower and an upper bound.
What you see in AI Forecasts
The AI Forecasts tab shows one card per business area. Each card has:
- The history and the outlook. Monthly history for the metric, then a six-month forecast.
- A point estimate for the next month and the change from the latest month.
- An 80% confidence band, shown as a lower and upper value. Read it as a range the next month is likely to land in, not a promise.
- A sample label on any card that is built from example data rather than yours.
Sample forecasts across eight business areas
Until you connect your own data, the tab shows sample forecasts so you can see what the tool does. There are eight, each built from twelve months of example history:
| Business area | Metric in the sample |
|---|---|
| Ecommerce | Revenue |
| Finance | Cash balance |
| Local | Foot traffic |
| Sales | Pipeline value |
| Content | Organic traffic |
| Ads | ROAS |
| HR | Open positions |
| PR | Media mentions |
The sample forecasts are refreshed weekly. They are examples, not predictions about your business, and they are marked as samples for that reason.
Forecasts on your own data
When you connect tools to Ergora, your workspace syncs their history and builds forecasts for the matching metrics. A forecast built on your data replaces the sample for that business area once it exists.
| Connected tool | Metric Ergora can forecast |
|---|---|
| Shopify | Monthly revenue and orders |
| Stripe | Monthly revenue |
| Xero | Financial metrics, including cash balance |
| Meta Ads and Google Ads | Blended ROAS across your connected ad accounts |
| Google Analytics | Organic traffic |
| HubSpot | Pipeline value |
| Klaviyo | Email campaigns sent per month |
A metric needs at least six months of history before Ergora will forecast it. Forecasts built on your data are refreshed when your workspace syncs your connected integrations. The full list of what Ergora connects to, and what is live, is on the integrations page.
Three ways teams use a forecast
Ads and traffic. If the organic traffic forecast shows a dip in a season you already know is quiet, you can plan paid spend or a content push around it, and check the blended ROAS outlook before you commit budget.
Revenue planning. A revenue outlook for the next six months gives a hiring or stock conversation a number to start from. Treat the confidence band as the range to plan for, and compare it with what you expect from launches or promotions the model cannot see.
Pipeline. The pipeline forecast is about the monthly value of the pipeline in your CRM, not about individual deals. It helps you see the direction of travel. It does not tell you which deal will close.
What it does not do
- It is not real time. The sample forecasts refresh weekly, and your own forecasts update when your workspace syncs.
- It does not forecast without data. The samples are examples. Your own forecasts need a connected source and at least six months of history.
- It does not predict individual deals, customers or products. The forecasts are at metric level, monthly.
- It does not know what you plan to do. A launch, a price change or a stock-out is not in the history, so it is not in the forecast.
- The horizon and the interval are fixed. Forecasts run six months ahead with an 80% band.
- It is not the same as Live Forecasting. The Finance pack has a separate tool, Live Forecasting, which shows burn, runway and scenarios from your connected Xero data. See the finance page.
A statistical forecast is a starting point for a decision, not the decision. Use it alongside what you know about your business.
What it costs
Seat plans start at $99 a month, with specialist packs from $59 a month, 50% off the first month and a 30-day money-back guarantee. There is no free plan. See the pricing page for what each plan includes, or the features overview for the wider workspace.
Frequently asked questions
Does Ergora use Google's TimesFM?
Yes. The AI Forecasts tab uses Google's TimesFM, a pre-trained time-series model, running on Ergora's own servers.
What can I forecast with Ergora?
Metrics with at least six months of monthly history from tools you have connected: revenue, orders, cash balance, ad ROAS, organic traffic, pipeline value and email campaigns sent. Every workspace also sees eight sample forecasts.
Are the forecasts based on my own data?
Once you connect a supported tool and it has enough history, yes. Until then you see sample forecasts, which are labelled as examples and are not about your business.
How often do forecasts update?
The sample forecasts refresh weekly. Forecasts built on your data are refreshed when your workspace syncs your connected integrations.
How far ahead do the forecasts look, and how reliable are they?
Six months ahead, by month, with an 80% confidence band. They are statistical estimates from past patterns and cannot account for changes you have not made yet, so treat the band as a range, not a guarantee.
Do I need a data science background?
No. You connect your tools and read the cards. There is no model to set up or tune.