
Once an AI agent answers part of your support queue, a simple question gets harder: how many people do you need on shift for the rest? Intercom has now built a product for that question. Its blog announces Workforce Management, a set of forecasting, scheduling and coverage tools inside the Intercom helpdesk. The post is the newest item in Intercom's own blog feed, ahead of their 7 October 2026 Fin 4 post, and it carries no publication date, so we can't tell you exactly when it launched.
What makes the announcement worth reading isn't the feature list. Intercom's post spells out a problem that AI support created. Then it puts the fix in a paid add-on.
Staffing a support team used to be a fairly stable calculation: look at last month's volume, work out how many conversations one person can handle, and divide. Intercom's post explains why an AI agent breaks that:
"Fin resolves 76% of queries on average, and for many teams, this percentage increases every week as they improve their content and expand what Fin can do."
"This means the volume that reaches your human reps fluctuates in a way that's much harder to predict."
The 76% is Intercom's own average for its own AI agent, Fin. It isn't a benchmark for AI support in general, and your number will depend on your content and your customers. The second sentence applies to any team running an AI agent, though. Your total volume might be steady, but the share left over for people moves whenever the AI improves, whenever you add a knowledge-base article, and whenever a new kind of question shows up that the AI can't answer.
Intercom also describes what most teams do about it today:
"Many teams use spreadsheets as a way to estimate and keep track of what's landing with their human reps."
According to the announcement, Workforce Management has four parts: forecasting, scheduling, Teammate Coverage and team performance tracking. The forecasting part is the core of it:
"Workforce management builds forecasts automatically from your own conversation data and excludes everything Fin handles, so you can staff for human-bound volume only."
"The forecast builds automatically from your last eight weeks of conversation data, with no integration setup needed."
The post says you can export that forecast, adjust it for something it can't know about, such as a launch or a seasonal spike, and re-import it as a CSV. Scheduling holds shift patterns as reusable templates and syncs with Google Calendar. Coverage compares the two:
"Teammate Coverage maps scheduled headcount against required staffing, hour by hour, so you can see where you're over- or understaffed before the queue fills up."
It's a sensible design. The forecast is built from conversations that actually reached people, not total volume, which is the right input once an AI is answering part of the queue.
The last paragraph of the post says who gets it:
"Workforce Management is available now and included in our Pro add-on."
"If you're on Pro, it's already live in your workspace."
So it isn't on every Intercom plan. It comes with the Pro add-on. The announcement doesn't give a price for that add-on, so we won't guess one here. If you're on Intercom and want this, the post sends you to your account team.
Look at the arc of the post as a whole. The AI agent takes a share of the queue. That share keeps changing, so the human remainder gets harder to predict. Predicting it is now a separate product, sold in a paid tier. None of this is unusual for software, but it's worth noticing when you work out what an AI agent will really cost: the agent itself may not be the only new line item.
You don't need a forecasting product to start planning sensibly. You do need to plan around the right number. A few habits help, whatever helpdesk you use:
DeskLeap doesn't have workforce management. There's no forecast, no shift scheduler and no hour-by-hour coverage view. Our reports show daily tickets created and resolved, response and resolution times, and per-agent figures. That's enough to see a trend, but it isn't a staffing plan, and we won't suggest otherwise.
What is different is what adding a person costs on the software side. The DeskLeap platform is 100% free with unlimited human agents. The only paid part is the optional AI Agent add-on, at $29/month each, with 500 AI replies per month included per agent. So covering a busy day with one more person, or giving a part-timer a login for the holiday rush, adds nothing to your helpdesk bill.
That doesn't make the planning question go away. People's time is still the real cost of support, and you still need to know who's on shift when the AI hands a conversation over. It just means the answer to "how many humans do we need?" never also has to answer "how many seats can we afford?"
Source: Intercom, "Announcing Workforce Management" by Andrea Nagel, intercom.com/blog/announcing-workforce-management. Read 16 November 2026. The page carries no publication date.