2Ring AI Prompt Reference Guide

Use natural language to create powerful alerting expressions — no coding required.

This guide provides a library of ready-to-use prompts organised by category. Use them exactly as written, or adjust queue names, thresholds, and time ranges to match your environment. Click the Ai Assistant button in 2Ring Dashboards & Wallboards, then type or paste any prompt below.

You can combine prompts and add specifics. The Ai Assistant understands context — it will map queue names and metric terms to the correct data sources in your platform.

1. Threshold Alerts

Create alerts that trigger when key metrics exceed or fall below specific values. Ideal for monitoring handle times, wait times, abandonment rates, and other KPIs.

AI Prompt	What It Does
How to create an alert when average handle time exceeds 5 minutes	Creates an expression that monitors AHT and triggers when it goes above 300 seconds.
How to create an alert when the longest waiting call has been holding for more than 2 minutes	Monitors the maximum wait time in queue and alerts at 120 seconds.
How to create an alert when the abandonment rate goes above 10%	Tracks call abandonment percentage and triggers above 10%.
How to create an alert when average speed of answer is greater than 30 seconds	Monitors ASA and fires when the average exceeds 30 seconds.
How to create an alert when average talk time drops below 1 minute	Flags unusually short conversations that may indicate call quality issues.
How to create an alert when calls in queue exceed 15	Triggers when the number of queued calls surpasses your threshold.
How to create an alert when after-call work time averages more than 90 seconds	Monitors ACW duration to identify wrap-up efficiency issues.
How to create an alert when hold time averages over 45 seconds	Tracks average hold duration to ensure customers are not waiting too long on hold.

Adjust the numeric values to match your organisation’s SLA targets — for example, replace ‘5 minutes’ with your actual AHT goal.

2. Agent State & Activity Alerts

Monitor individual agent states and activity durations in real time. Use these to identify agents who may be stuck, disengaged, or need assistance.

AI Prompt	What It Does
How to create an alert when an agent has been on hold for more than 2 minutes	Detects extended hold times per agent, which may indicate the agent needs help or is struggling with a complex call.
How to create an alert when an agent has been in after-call work for more than 3 minutes	Flags agents spending excessive time in wrap-up, helping supervisors coach efficiency or identify system issues.
How to create an alert when an agent is in a Not Ready state for more than 10 minutes	Identifies agents unavailable for an extended period — possible unplanned absence or a forgotten state change.
How to create an alert when an agent has been on an active call for more than 15 minutes	Flags unusually long calls that may require supervisor intervention.
How to create an alert when an agent switches to Not Ready state more than 5 times in one hour	Detects frequent state toggling that could indicate avoidance behaviour or a system issue.
How to create an alert when an agent has been in a ready state with no calls for more than 20 minutes	Identifies agents available but not receiving calls — may signal routing or queue assignment issues.
How to create an alert when an agent’s average handle time exceeds the team average by more than 2 minutes	Highlights individual agents significantly above the team norm for targeted coaching.
How to create an alert when an agent has not taken a call in the last 30 minutes while logged in	Detects potential issues with call routing or agent availability configuration.
How to create an alert when an agent’s hold time per call exceeds 90 seconds on average	Monitors per-agent hold patterns to identify training needs or knowledge gaps.
How to create an alert when an agent has been in a custom aux state for more than 15 minutes	Tracks time in custom reason codes such as training or meetings to ensure schedule adherence.

Agent-level alerts are especially effective on real-time wallboard displays — supervisors can identify who needs attention at a glance.

3. Staffing & Availability

Monitor agent availability and staffing levels in real time. Stay ahead of understaffing before it affects service levels.

AI Prompt	What It Does
How to create an alert when fewer than 3 agents are available in the sales queue	Monitors available agent count for a specific queue and alerts when it drops below the threshold.
How to create an alert when no agents are in a ready state while callers are waiting in the queue	Critical alert for when zero agents are available to take calls.
How to create an alert when more than 50% of agents are in after-call work state	Identifies when too many agents are in wrap-up simultaneously.
How to create an alert when the number of logged-in agents drops below 5	Tracks total staffing levels to prevent understaffing.
How to create an alert when agents on break exceed 4 at the same time	Helps manage break scheduling to maintain coverage.
How to create an alert when the ratio of available agents to calls in queue falls below 1 to 1	Compares agent availability to demand for balanced workload distribution.

Replace queue names and agent counts with values specific to your contact centre.

4. Service Level & Performance

Track overall contact centre performance against your targets. These prompts focus on SLAs, customer satisfaction metrics, and operational efficiency.

AI Prompt	What It Does
How to create an alert when service level drops below 80%	Monitors the percentage of calls answered within your SLA target time.
How to create an alert when the percentage of calls answered within 20 seconds falls below 75%	Tracks the classic 80/20 service level with custom values.
How to create an alert when first call resolution rate drops below 70%	Monitors FCR to ensure quality of customer interactions.
How to create an alert when transfer rate exceeds 25%	Flags excessive call transfers that negatively impact customer experience.
How to create an alert when total calls handled today is less than expected by this time	Compares actual throughput against forecasted call volume.
How to create an alert when the callback queue has more than 10 pending requests	Monitors callback backlog to ensure timely customer follow-up.

Service level targets vary by industry. Customise these percentages to match your organisation’s SLA commitments.

5. Queue Management

Stay on top of queue health across your contact centre. Monitor volumes, overflow conditions, and routing efficiency.

AI Prompt	What It Does
How to create an alert when any queue has more than 20 calls waiting	Broad monitoring across all queues for volume spikes.
How to create an alert when the support queue wait time exceeds 3 minutes but the sales queue is under 1 minute	Compares queue conditions to identify resource rebalancing opportunities.
How to create an alert when a queue has had zero calls answered in the last 15 minutes	Detects potential routing or staffing issues in specific queues.
How to create an alert when overflow calls from the billing queue exceed 5 per hour	Monitors overflow frequency to identify capacity issues.
How to create an alert when the average wait time across all queues exceeds 2 minutes	Provides a global view of wait time health across the centre.
How to create an alert when the difference in wait times between queues exceeds 3 minutes	Identifies imbalanced workload distribution across queues.

Use specific queue names from your environment in quotes (e.g., “support queue”) for more precise results.

6. Custom & Advanced Combinations

Combine multiple conditions for sophisticated alerting scenarios. These prompts show the Ai Assistant’s ability to handle complex, multi-condition expressions.

AI Prompt	What It Does
How to create an alert when AHT exceeds 5 minutes and available agents are fewer than 3	Combines handle time and staffing thresholds for high-impact alerts.
How to create an alert when service level is below 80% or calls in queue exceed 20	Triggers on either condition for early warning.
How to create an alert when AHT is above 4 minutes, service level is below 85%, and fewer than 5 agents are available	Triple-condition alert for critical escalation scenarios.
How to create an alert when wait time exceeds 2 minutes but only during peak hours between 9 AM and 5 PM	Time-scoped alerting that factors in business hours.
How to create an alert when any queue’s abandonment rate is above 15% while staffing is at full capacity	Identifies efficiency problems even when fully staffed.
How to create an alert when the sales queue has more than 10 calls waiting and average handle time in that queue is over 6 minutes	Queue-specific multi-metric monitoring.

Start simple, then add complexity. Begin with the question before asking to create an expression — 2Ring has extensive out-of-the-box alerting built in. Validate each expression before saving.

7. SQL Expression-Based Alerts

Some scenarios require SQL expressions to query data sources directly. The Ai Assistant builds these from your plain-language descriptions automatically.

Real-Time Agent & Queue Queries

AI Prompt	What It Does
How to write a SQL expression to alert when agents in Not Ready state with reason code ‘Lunch’ exceeds 5	Queries agent state data filtered by specific reason codes for targeted monitoring.
How to write a SQL expression to count calls waiting longer than 3 minutes across all queues	Uses SQL COUNT with a WHERE clause on wait duration for granular queue analysis.
How to write a SQL expression to find agents who have handled fewer than 10 calls today	Queries daily call count per agent to identify underperformance or routing issues.
How to write a SQL expression to calculate the percentage of calls abandoned in the last 30 minutes	Uses SQL aggregation with time-window filtering for rolling abandonment monitoring.
How to write a SQL expression to alert when the average wait time for a specific routing group exceeds 2 minutes	Queries group-level metrics using SQL AVG for precision routing alerts.
How to write a SQL expression to alert when any agent has been in the same state for more than 20 minutes	Uses duration calculation to detect stuck or forgotten agent states.

Avoid using the word “Skill” in SQL prompt descriptions as it can cause a generation error. Use alternative terms such as ‘group’, ‘team’, or ‘routing profile’ instead.

Historical & Trend-Based Queries

AI Prompt	What It Does
How to write a SQL expression to compare today’s call volume to the same day last week and alert if it is 20% higher	Uses date functions and percentage calculations for trend-based volume alerts.
How to write a SQL expression to alert when total transfers today exceeds yesterday’s total by more than 15%	Compares daily aggregates across time periods to detect unusual transfer patterns.
How to write a SQL expression to alert when hourly call volume drops below the rolling 4-week average for that hour	Uses historical averaging with time-of-day grouping for forecast-based alerting.
How to write a SQL expression to track first call resolution per agent over the past 7 days and alert below 65%	Queries historical per-agent FCR data for ongoing performance monitoring.

Aggregation & Grouping Queries

AI Prompt	What It Does
How to write a SQL expression to group calls by disposition code and alert when ‘Callback Requested’ exceeds 20% of total calls	Uses SQL GROUP BY with percentage calculation on disposition codes.
How to write a SQL expression to rank queues by average wait time and trigger when the worst-performing queue exceeds 4 minutes	Uses SQL ranking to identify the worst queue dynamically.
How to write a SQL expression to alert when more than 3 agents on the same team are simultaneously in Not Ready state	Uses SQL COUNT with GROUP BY on team assignment.
How to write a SQL expression to calculate the standard deviation of handle times and alert when an agent’s AHT is more than 2 standard deviations from the mean	Uses SQL statistical functions for outlier detection in agent performance.
How to write a SQL expression to count repeat callers within a 24-hour window and alert when it exceeds 10	Queries call history with window functions to detect repeat contact patterns.

Start by describing what you need in plain English. The Ai Assistant decides whether a SQL expression is required and generates the syntax automatically.

8. Platform-Specific Expressions

Prompts, expressions, and considerations tailored to Cisco (UCCX / Webex Contact Center), Five9, and Genesys Cloud CX. Use these alongside the universal prompts in earlier sections.

Cisco UCCX / Webex Contact Center (WxCC)

Supervisor Prompts

AI Prompt	What It Does
How to create an alert when a RONA event occurs more than 3 times in one hour	RONA (Redirect on No Answer) is triggered when an agent fails to answer a routed call. Frequent occurrences indicate availability or focus issues requiring immediate attention.
How to create an alert when the IVR self-service containment rate drops below 40%	Monitors how many callers resolve their query in the IVR without reaching an agent. A drop signals a broken flow, a changed menu option, or a surge in complex queries on WxCC.
How to create an alert when a Webex Contact Center digital channel queue (chat or email) exceeds 10 unhandled interactions	WxCC routes voice, chat, and email together. Digital channel queues are often overlooked when supervisors focus only on voice metrics.
How to create an alert when a Precision Queue has no eligible agents due to routing criteria	Precision Queues use attribute-based routing. If no agents match the required attributes, calls stall silently. This alert catches the condition before customers are impacted.
How to create an alert when an agent has not returned to Ready after an automatic Not Ready transition within 5 minutes	After a RONA event, Cisco automatically sets the agent to Not Ready. This tracks agents who fail to reset their state, preventing prolonged unavailability.

Service Level Expression — Cisco WxCC

Use this in a Custom Column (Numeric type) to calculate service level on Webex Contact Center:

CASE
  WHEN {{QueueStatisticsTasksTotal}} = 0 THEN 0
  ELSE (
    ( {{QueueStatisticsTasksAcceptedInServiceLevel}}
    + {{QueueStatisticsTasksAbandonedInServiceLevel}} * 1.0 )
    / {{QueueStatisticsTasksTotal}}
  ) * 100.0
END

The division-by-zero guard (WHEN … = 0 THEN 0) remains best practice even though recent versions return N/A automatically.

Five9

Supervisor Prompts — Outbound & Blended Campaigns

AI Prompt	What It Does
How to create an alert when an outbound campaign’s contact rate drops below 15%	A sustained drop below 15% signals list exhaustion, DNC filtering issues, or incorrect dial settings — a fundamental Five9 outbound KPI.
How to create an alert when the live connect rate on a predictive dialling campaign falls below 8%	Measures answered calls as a percentage of dial attempts. Below 8% indicates poor list quality, high answering-machine rates, or compliance throttling.
How to create an alert when agents manually disposition calls as No Answer more than 20 times in one hour	Excessive manual No Answer dispositions can indicate agents skipping calls or a deteriorating campaign list.
How to create an alert when the talk-to-idle ratio for outbound agents drops below 50%	In blended or outbound campaigns, agents should be active roughly half the time. A sustained drop below 50% points to a dialling pace, list, or routing problem.
How to create an alert when a skill queue has agents assigned but none are actively accepting calls	A common configuration issue where agents are assigned to a skill but their availability or call acceptance settings prevent them from receiving interactions.
How to create an alert when an agent’s personal callback queue has more than 5 pending items	Personal callbacks represent committed return calls to customers. An overloaded queue directly affects customer satisfaction.

Average Speed of Answer Expression — Five9

Use this in a Custom Column (Numeric type) with Five9 custom reporting Skill Data:

CASE
  WHEN {{SkillStatisticsCalls}} = 0 THEN 0
  ELSE {{SkillStatisticsTotalSpeedOfAnswer}} / {{SkillStatisticsCalls}}
END

Cloud connectors such as Five9 supply pre-aggregated values from the API. Avoid creating an average of averages — instead, sum raw counts and divide within the expression as shown above.

Genesys Cloud CX

Supervisor Prompts — Omnichannel & Routing

AI Prompt	What It Does
How to create an alert when an omnichannel queue has voice and chat interactions both exceeding their thresholds at the same time	Genesys Cloud CX routes voice, chat, email, and social through a unified queue model. Monitoring both channels simultaneously prevents supervisors from addressing one at the expense of another.
How to create an alert when predictive routing is assigning more than 20% of interactions to agents with a below-average predicted outcome score	A high proportion of interactions going to lower-scoring agents can indicate the model needs retraining or that the agent pool is understaffed for the demand type.
How to create an alert when an email queue has interactions older than 4 hours that have not been handled	Email SLA violations are common in Genesys deployments where voice takes priority. This alert surfaces aging emails before they breach SLA.
How to create an alert when a queue group’s aggregate service level falls below target while individual queues appear healthy	Queue groups can mask systemic problems. An aggregate alert on the group catches issues that are invisible at the individual queue level.
How to create an alert when IVR flow containment in a specific Architect flow drops below 30%	A containment drop signals a broken intent, a changed menu option, or a spike in queries the bot cannot handle.
How to create an alert when an agent’s concurrent interaction limit is reached across all channels	When the concurrent limit is reached the agent is effectively unavailable for new interactions — this makes that condition visible in real time.
How to create an alert when the combined abandonment rate across voice, chat, and email in a queue group exceeds 12%	Aggregates abandonment across all interaction types for a single omnichannel health metric.

Genesys Cloud connector data is sourced from the Analytics API. Pre-aggregated metrics (e.g., average handle time) cannot be re-aggregated across queues. Build cross-queue averages by dividing the SUM of raw counts by the SUM of interaction counts within your expression.

9. Best Practices for Writing Prompts

Be Specific

Include exact metric names, threshold values, and queue names where possible. The more specific your prompt, the more accurate the generated expression will be.

Use Plain Language

Write prompts as if explaining to a colleague. The Ai Assistant is designed for natural, conversational descriptions — technical jargon is not required.

Start Simple, Then Refine

Begin with a single-condition alert and verify it works. Then ask the Ai Assistant to add conditions or adjust the expression. A useful instruction to refine output:

Modify your expression to return 1 and 0 instead of text labels.

Review the Output

Always validate the generated expression before saving. The Ai Assistant shows exactly what it created — ask it to make adjustments if anything looks off.

Common Modifiers

AI Prompt	What It Does
How to limit this alert to only trigger during business hours	Restricts the alert to specific time windows.
How to scope this alert to the [queue name] queue specifically	Scopes the expression to a single queue.
How to change the threshold value of this alert to [value]	Overrides a default or previously generated threshold.
How to exclude weekends from this alert	Adds day-of-week filtering to the expression.

10. Customer Questions Log

Use this log to record questions received from customers using the 2Ring AI Assistant. Tracking patterns helps identify prompts worth adding to this guide and surfaces known limitations across platforms.

AI Result options: ✅ Works correctly | ⚠ Partial / needs edit | ❌ Fails / incorrect | 🔄 Needs prompt update

Date	Customer / Company	Platform	Question Asked	AI Result	Notes / Follow-up
01-04-2026	Cisco	WxCC	SEE chapter 8
17-04-2026	Five9	Five9	See Chapter 8		Need to discuss with customer – standard 2Ring Functionality instead of AI