Average Handle Time (secs) by Date — Dedicated Sub-View

The Average Handle Time (AHT) by Date sub-view provides a full-page, dedicated analysis of conversation duration trends over time. AHT measures how many seconds, on average, each conversation lasted on a given day. This metric is a critical indicator of AI assistant efficiency, conversation complexity, and potential bot performance issues.

Figure 1: Average Handle Time (secs) by Date — With Hover Tooltip (Mon, 3/23/26: AHT 956.0 secs)

1. Chart Details

Visualization Type: Single-line time-series chart. Color: Gold/amber. Y-axis: Average Handle Time in seconds (0.0 to 1,800.0). X-axis: Day of Conversation Date.

The hover tooltip visible in the screenshot reveals: Date: Mon, 3/23/26 | AHT (secs): 956.0. This confirms the tooltip format for this chart: date label and the exact AHT value in seconds.

Full data series observed (03/08/2026 – 04/09/2026, STG stage):

Date	Average Handle Time (secs)
Wed 3/11/26	~1,300 secs (approx. 21.7 min) — estimated from chart position
Mon 3/16/26	1,739.0 secs (peak — approx. 29.0 min)
Tue 3/17/26	~100 secs — sharp drop from peak
Wed 3/18/26	~110 secs
Thu 3/19/26	51.5 secs — labeled on chart (approx. 0.9 min)
Mon 3/23/26	956.0 secs (tooltip confirmed — approx. 15.9 min)
Wed 3/25/26	~100 secs
Thu 3/26/26	~100 secs
Fri 3/27/26	~1,320 secs (secondary spike — approx. 22 min)
Mon 3/30/26	~700 secs (declining)
Wed 4/1/26	~250 secs
Thu 4/2/26	~30 secs — lowest point

2. Interpreting AHT Patterns

High AHT values (above 900 seconds / 15 minutes) in an AI assistant context are unusual and typically indicate one of the following: (1) conversations involving complex, multi-step interactions where the user is navigating multiple flow branches; (2) bot loops where the assistant repeatedly fails to understand user input and re-prompts; (3) data anomalies where a single very long conversation inflates the average for a low-volume day (common in staging environments with 1–5 conversations per day); or (4) specific assistants with inherently long flows (e.g., detailed intake forms, guided troubleshooting processes).

The dramatic day-to-day variation in this dataset (from 51.5 seconds to 1,739.0 seconds) strongly suggests a staging environment data anomaly pattern rather than a production behavior, where large sample sizes smooth out individual outliers. In production, filtering by specific Assistant name will reveal which bots drive the highest average handle times.

3. Business Value

AHT is a dual-edged metric. For live agent operations, shorter AHT is always preferred. For AI assistants, the interpretation is more nuanced: an AHT that is too short may indicate conversations that end before the user’s intent is resolved (e.g., quick timeouts or immediate failures), while an AHT that is too long indicates the assistant is not efficiently guiding users to resolution. The optimal AHT varies by use case and assistant design. Trending AHT over time, especially after bot updates or configuration changes, reveals whether changes improved or degraded conversation efficiency.