Cognitive Load for Concurrent Chats

Note: This feature is available with the Premium WFM License.

When agents handle more than one chat at a time, switching between conversations takes a mental toll. Responses slow down the more chats an agent handles. Cognitive Load for Concurrent Chats teaches your CXone Workforce Management scheduling engine to recognize that toll and account for it. As a result, your staffing plans reflect what agents can actually handle, not just a theoretical chat count.

With cognitive load enabled, your Service Level Agreement (SLA) and Average Speed of Answer (ASA) forecasts become more realistic, and your staffing requirements match real-world agent capacity in multi-chat environments.

Pre-requisites

Before you enable cognitive load, verify the following:

• You have a Premium WFM License.

• Chat concurrency is enabled in your tenant through the ACD, configured by team or by agent.

• Your tenant has enough historical chat concurrency data for the model to learn from. The model trains itself on data your tenant is already generating. There is nothing to upload, tag, or configure.

Understanding the Cognitive Load Model

The cognitive load model uses machine learning trained on your tenant's own historical chat data. It analyzes how long agents take to respond as their number of active chats changes and learns the pattern: as concurrent chats increase, response times to each individual chat tend to increase too.

The model applies that pattern going forward by adjusting simulated contact end times and agent capacity based on how many chats an agent is handling at once. It does not assume every agent handles concurrency identically and instantly.

Because the model is trained on your own tenant's data, it reflects your agents' real behavior, not a generic industry average. Longer chat Average Handle Time (AHT), longer response times, and higher chat message intensity could indicate that more staffing is required based on your tenant's chat trends.

Forecast, Schedule Manager, and Intraday Integration

Once enabled, the scheduling engine factors cognitive load into every simulation that involves chat concurrency:

• Forecast: Uses cognitive load when calculating staffing requirements for chat skills with concurrency.

• Schedule Manager: Uses cognitive load when simulating how well a schedule will perform.

• Intraday: Uses cognitive load when re-evaluating staffing needs against real-time conditions.

There is no new page or setting to configure. The model runs in the background, adjusting how the simulation treats an agent's capacity as their concurrent chat load increases.

What Changes in Your Staffing Results

• Staffing numbers may shift: Since simulations now account for the real cost of multitasking, required staffing levels in concurrent-chat environments may look different than before.

• Existing published schedules are not rewritten: Cognitive load affects how future simulations, forecasts, and staffing requirements are calculated. It does not retroactively change schedules you have already published.

• SLA and ASA targets do not change: What changes is how confidently your plans can meet those targets, because the simulation behind them is more realistic.

Considerations

• If you change the concurrency value in WFM Step 4 Staffing parameters for Chat under Forecasting, or change agent concurrency in the ACD, the model still considers cognitive load. However, with the updated values, it takes time for the data to be retrained: a minimum of three weeks and up to three months.

• The model does not use data from other companies or tenants. It is trained only on your own tenant's historical chat data.

• You do not need to retrain or update the model yourself. It uses your tenant's ongoing chat history automatically.

Frequently Asked Questions (FAQs)