Know how busyyou're going to be
Forecasting shows your expected support traffic for the next 90 days, day by day, and where it comes from. It is rebuilt every night from your own history, and it tells you how far it has typically landed from the real day.
500 conversations, no credit card.
What's in Forecasting
- StaffingHow many people each day needs, per channel and per team.
- Why this dayThe weekday baseline plus the named causes that add up to the total.
- ForecastExpected traffic for the next 30, 60 or 90 days, day by day.
- BandsLow, Normal, Busy, High and Peak, against your typical weekday.
- Typical missHow far the forecast has landed from the real day, measured on your own history.
- Track recordThe last 30 days, each one as forecast against what arrived.
- Event ImpactWhat each event did to your volume, and the window it acts over.
- UnexplainedLift with no known cause. Name it once and the forecast learns it.
- Demand CalendarPublic holidays, school holidays, sales events and your own campaigns.
- Add eventName a day, set its markets and its window, and let the lift be learned from past dates.
Grounded in your history
Forecasting reads up to two years of your own conversation traffic — your weekday baselines, your seasonal shape, and how each event actually ramped up and decayed for you. Nothing is borrowed from an industry template. When an event has only happened once, the product labels it a guess, not a forecast.
It also tells you how wrong it has been. Typical miss is the average distance between the forecast and the day that arrived, scored on your own history at the lead time you are reading, and Track record lays the last thirty days out one by one. Until enough days have been scored it says Not measured yet.
- Horizon: the next 30, 60 or 90 days, rebuilt every night at 04:30 UTC
- Each day is banded Low, Normal, Busy, High or Peak against your typical weekday
- Cut the forecast by channel, by team or by topic, one at a time
- About 60 days of traffic is enough to start; a channel with too little history is left out and the day panel says so
A forecast that shows its work
Every forecast day breaks down into your weekday baseline plus named causes — public holidays, sales events, your own campaigns, spillover from a busy day next door — and the parts always sum to exactly the total. Extra volume with no known cause is labeled Unexplained. Name it once and the forecast learns its lift and its shape for next time. Event Impact is where that learning is on show: what each event did to your volume, the window it acts over, which channels and topics it landed in, and how many past occurrences it was learned from.
Staff the peaks before they hit
Staffing forecasts the work a person will actually touch — what is left after the Agent Stack — and divides it by the conversations one agent closes per day. The table then reads in people: suggested headcount for each day, by channel or by team.
Your typical headcount is what the team has had on that weekday recently. When a day needs more than that, and the gap is wider than the forecast's own measured miss, the page says which channel is short of people and when. The line under the table is the product's own: Forecasting suggests capacity. It does not schedule anyone.
- Suggested staffing by day, cut by channel or by team
- An alert only where the shortfall is larger than the forecast's own error
- Where no day is short, it says so instead of filling the page
The calendar your traffic already follows
Sources is the list the forecast draws its causes from: public holidays and school holidays, which you import per country; sales events like Black Friday, which come curated; and your own campaigns, which you add. Each source has its own switch. Add event names a day, sets the markets it applies to and the days it acts over, and can learn its lift from the dates it already happened on.
- Public and school holidays import per country, for this year and next
- An event can be excluded from the forecast, or marked as needing confirmation each year
- Every change applies at the next nightly run, and the page says so
Forecasting sits on top of the traffic the rest of the platform generates. Conversations are the history it learns from, Analytics tells you what already happened while Forecasting tells you what's coming, and the numbers flow into the Agent Workspace and Monitor where the team feels the peak.
Conversations
Every conversation across every channel is the raw history the forecast is built from.
Analytics
Analytics reports what already happened; Forecasting projects what's coming next.
Agent Workspace
The expected load lands where agents work, so a busy week is never a surprise.
Monitor
Watch the forecast meet reality in real time as the day the model called Peak arrives.
Expected traffic, and where it comes from.
See Atender in action
Try the full platform for free. Or book a demo and we will walk you through it.
500 conversations, no credit card.