Monitoring & Forecasting
AI Anomaly Detection
Detects anomalies in transactions, traffic or operations data.
Monitoring & Forecasting
AI Forecasting Assistant
A forecast for demand, revenue or volume, with the assumptions behind each number explained in plain language.
Planning stock, staff or budgets on a guess is expensive in both directions. This assistant projects demand, revenue or volume from your own historical data and explains the assumptions behind each number in plain language. You can question the forecast instead of taking it on faith. Updates arrive on a schedule by email or Slack, and connections are read-only.
- Forecasts of demand, revenue or volume from your own history
- The assumptions behind every number explained in plain language
- Forecasts checked against your past data before launch
$6,900
20–30 working days
Optional maintenance $345/mo
What it is
AI Forecasting Assistant projects demand, revenue or volume from your historical data, and explains the assumptions behind each number so you can judge the forecast rather than take it on faith. Forecasts are checked against your history before launch and refresh on a schedule.
Why businesses buy it
Over-forecast and cash sits in stock or idle staff. Under-forecast and you miss sales or burn out the team. Spreadsheet forecasts are slow to update and hard to question. This assistant produces regular forecasts from your real history and shows the assumptions, so managers can challenge and adjust them with confidence. The written scope sets the sources, forecast lines and validation against your history before you order, so you know exactly what you will receive.
Problems this solves
Forecasts live in spreadsheets that only one person understands
The assistant projects demand, revenue or volume from your historical data. It explains the assumptions behind each number, so anyone can follow how the forecast was made.
Teams accept or reject forecasts without knowing why
Each forecast comes with its assumptions written out. Managers can judge the forecast and challenge an assumption rather than take the number on faith. Planning meetings start from shared reasoning.
Forecasts go stale because updating them takes too long
Forecasts refresh on a schedule and are delivered by email, Slack or a dashboard. Before launch they are checked against three months of your history.
A good fit if
You need a repeating forecast, such as monthly demand, revenue or order volume, and currently build it by hand in a spreadsheet.
You want the assumptions behind each forecasted number stated in plain language, not a single unexplained figure.
You already keep data in tools such as Google Sheets, BigQuery, PostgreSQL or Google Analytics 4, and can grant read-only access.
You want the forecast checked against at least 3 months of your own historical data before you rely on it.
Not the right fit if
You specifically need a weekly sales pipeline forecast for a sales team. See our AI Sales Forecasting product, which is built around pipeline stages and win rates.
You want unusual patterns in past or current data flagged, not future values projected. See AI Anomaly Detection.
You want alerts when a number crosses a threshold rather than a projection of where it is headed. See AI KPI Monitor.
You want messy source data cleaned before anything is forecast from it. See AI Data Cleaning, which cleans and standardizes datasets first.
How it works
- You tell us what to forecast, such as monthly demand for a product line, revenue by region, or order volume, and connect the relevant data sources.
- The assistant builds a forecast model from at least 3 months of your historical data, accounting for trend and seasonality where the data supports it.
- Each forecasted number is delivered with the assumptions behind it stated in plain language, such as which trend it extrapolated and which season it assumed.
- The forecast is checked against your recent actuals before launch, so you can see how closely it would have tracked reality.
- Forecasts refresh on the schedule you choose and are delivered to the channel you set, such as email, Slack or a dashboard.
- You review the first forecasts with us and adjust the assumptions before the 30-day support window closes.
What you provide within 72 hours
Your portal lists these the moment you order. Delivery starts the working day after everything is in and your access is checked.
- Read-only access to each data source the forecast should be built from.
- Your own definitions of what to forecast, such as demand, revenue or volume, and at what level of detail.
- At least 3 months of historical data, so the forecast model can be built and checked before launch.
- The list of recipients and the channel you want forecasts delivered to.
- Admin access and API keys for every connected system, shared by adding our team as a user or creating a key you control, never by email.
- Your business information (company basics, hours and brand voice), shared during intake.
- One decision-maker who can confirm the forecasting assumptions before launch.
- An API key for your own AI model account (for example OpenAI or Google Gemini), created for this build with a spending limit.
The 72-hour rule. Send everything on your checklist, including access, within 72 hours of ordering so your build starts on time.
Timeline
Start
Intake confirmed and what to forecast, and at what detail, agreed.
Weeks 1–2
Data sources connected and the forecast model built.
Week 3
Delivery schedule and alert thresholds configured.
Weeks 3–4
Forecasts checked against 3 months of your historical data and recent actuals.
Week 4
Your review and two rounds of changes to assumptions.
Go-live
Live, handover report sent, 30 days of support starts.
| Option | Delivery time | Price |
|---|---|---|
| Standard | 20–30 working days | $6,900 |
| Express | 15–25 working days | $8,100 |
| Rush | 10–20 working days | $9,300 |
What you receive at handover
- The AI forecasting assistant, live with scheduled refreshes in email, Slack or a web dashboard
- Read-only connections to up to three data sources
- Defined forecast metrics, one dashboard or recurring report and up to five alerts
- A test report comparing forecasts with three months of your historical data
- An admin guide and the written handover report
- 30 days of post-launch support
Expected benefits and how to measure them
- Can make planning less dependent on one person's spreadsheet
- Helps managers question forecasts with the assumptions in front of them
- Keeps forecasts current without manual rework
- Forecast accuracy
- Compare each forecast with actual results when the period closes. Track the gap over several periods.
- Time to produce a forecast
- Time the current manual forecast, then compare with the scheduled refresh.
- Planning decisions changed
- Note decisions on stock, staffing or spend that changed after a forecast review.
Before you start: Keep your last few manual forecasts and the actual results. Record how long each forecast takes to build.
These are things to track, not promised results. Outcomes depend on your data, your team and your market.
Running costs
The build itself is a one-time payment, shown before you buy. Running the assistant costs roughly $20 to $150 a month in your own accounts, covering AI model usage and hosting or scheduler fees, depending on how often forecasts refresh and how many sources feed them.
AI usage during the build and testing is billed to your own accounts too, so the provider relationship stays in your name.
Ask about running costsCustom requirements
If you need more than 3 data sources or forecasts at a finer level of detail than agreed, add a note at checkout for a written Change Quote. Your order still runs on its original delivery date. Larger or bespoke projects go through the Custom AI request form instead.
Request a custom versionKeep it running
Monthly maintenance for this build covers monitoring, fixes and change hours for assumption updates or a new forecast line. The hours scale with the build. Maintenance starts after the included support period and can be cancelled any month.
$345/mo (5% of the build price a month, includes 3 change hours)
Optional. Starts after the post-launch support period. Cancel any month.
How maintenance worksHow this compares
| Do it yourself with a SaaS tool | Hire a custom agency | Upmark Labs | |
|---|---|---|---|
| Price | A forecasting tool charges a subscription regardless of how much you forecast | A forecasting consultancy quotes per model built | Ours is one fixed price, shown before you buy |
| Delivery | A forecasting tool is instant to sign up but slow to fit your data | An agency model build usually takes several weeks | Ours has a fixed delivery time shown before you buy, with faster options |
| Assumptions | A forecasting tool often outputs a single number with no stated reasoning | An agency documents assumptions in a separate report | Ours states the assumptions behind every forecasted number |
| Validation | A forecasting tool rarely shows how it would have tracked past actuals | An agency validation depends on what was agreed | Ours is checked against 3 months of your own historical data |
Questions about this product
What does the AI Forecasting Assistant price include?
One fixed price, agreed before work starts, covers up to three read-only data sources, up to 25 defined metrics, one delivery channel and forecast checks against three months of your history. Faster delivery options are shown in the price panel before you order, and there are no other build fees.
What can the AI Forecasting Assistant forecast?
Demand, revenue, order volume or a similar recurring metric, at the level of detail you define during intake, such as by product line or by region. It is not limited to one department; the metric to forecast is agreed as part of your setup.
How long does the AI Forecasting Assistant take to launch?
The standard, Express and Rush delivery times are all shown in the price panel before you order, and the clock starts once your intake is complete. Delays only happen when data access or forecast definitions arrive late, which moves delivery by three working days per day of delay.
How does the AI Forecasting Assistant explain its assumptions?
Every forecasted number is delivered with a plain-language note on how it was reached, such as the trend it extrapolated and the seasonal pattern it assumed. The forecast is also checked against your recent actuals before launch, so you can judge its track record.
How much does the AI Forecasting Assistant cost to run each month?
Roughly $20 to $150 a month in your own accounts, covering AI model usage and hosting or scheduler fees. The figure depends on how often forecasts refresh and how many sources feed them. You pay these providers directly, with no markup from us.
Can the AI Forecasting Assistant add a new forecast line later?
Yes, as an add-on. A new forecast at a different level of detail, or a fourth data source, is quoted through a written Change Quote once you send your request. Adjusting existing assumptions fits inside the monthly change hours included with maintenance.
Who owns the forecast models the AI Forecasting Assistant builds?
You do. The forecast model, its assumptions and every delivered forecast run against your own historical data, in your own accounts. If maintenance lapses, forecasts keep running; you lose the monthly monitoring and included change hours, not the build. While active, those hours cover assumption updates or a new forecast line.
Plan from a forecast you can question, not a single unexplained number.
Order the AI Forecasting Assistant and a checked, explained forecast goes live on your data. The scope above is what you get, in writing, before you pay.
Works well with
