Summarize with AI:ChatGPTPerplexityGoogle AIClaude
Quick Answer
Give every routing rule a timed next step. When the specialist a rule targets is offline, offer the chat to the whole department after 45 seconds, reassign email after 4 business hours, and send unanswered calls to voicemail or a callback. Then name backup and supervisor roles as the next hops, so every path ends with a person.
How do I set routing fallbacks when the right agent is offline?
Give every rule a timed next step. LiveHelpNow routing already holds the parts: departments with schedules, skill levels 1 to 10, backup and supervisor roles, and re-queue timers.
What you need
- Admin access to departments, schedules, shifts and holidays
- A skill level from 1 to 10 set for each agent
- At least one backup role and one supervisor role per department
- A chosen wait for each channel: chat, email and calls
- A last stop for every path: a person, a voicemail or a callback
Steps
- List every rule that targets one specialist or one skill. These are the rules that strand customers when one person is away.
- Set department schedules, shifts and holidays. The router can only fall back when it knows who is off.
- Add a re-queue timer to each chat rule. In your live chat system for customer support, start by offering the chat to the whole department after 45 seconds.
- Name backup and supervisor roles as the next hops. Backups receive the contact first, and the supervisor receives whatever backups cannot take.
- Add a ticket automation for email. Reassign any unanswered email to the department after 4 business hours.
- Send unanswered calls to voicemail or a callback. No caller should ring out with nowhere to go.
- Keep a person one click away on the AI assistant path. The assistant identifies itself as an AI assistant. When no article in your knowledge base software covers a question, it gives the configured fallback, "I'm not sure about that, let me connect you with a human agent", and hands off to a person with a summary.
- Turn on queue alert emails and test each path. Set the specialist offline, send a test chat, email and call, and confirm that each one lands where you expect.
Keep a person at the end of every path. A 2026 Gartner survey found that 87% of customers consider an option to reach a human agent essential when a company uses GenAI in customer service.
The evidence does not settle the ideal timer length for every team. Treat the 45-second and 4-hour values as starting points, and let your own wait and abandonment reports move them after the first month of live traffic.
Common mistakes
- A rare skill with no time limit. A CrestApps routing guide drops a skill requirement once the caller has waited a set number of seconds, so give your specialist rules the same limit.
- A chain that loops. The same CrestApps guide never sends a caller back to a queue already visited, so check that no fallback points to an earlier queue.
- A bot as the last stop. Point the final hop at a person, and use customer service automation tools only for the steps before it.
Frequently Asked Questions
What do teams ask after setting routing fallbacks?
Should a call fall back faster than a chat?
Yes. In a November 2026 Medium post on Amazon Connect, Ambuj writes that on voice, "every millisecond spent waiting for the primary to fail is dead air." Worst-case delay equals the first timeout plus the fallback's own response time, so keep the first call timeout short.
Do fallbacks work for a team with fewer than 10 agents?
They matter more on a small team. Make the whole department the first fallback and the supervisor role the last person in the chain. It works like a spare house key: one holder is enough, as long as everyone knows who has it.
What should I check once fallbacks are live?
Open a sample of moved contacts. Each one should arrive as a handoff to a person with a summary, meaning the agent sees what was already said. Then compare wait times by department for the month before and after, and adjust any timer that still leaves contacts waiting.
Written by
Michael Kansky
Founder
Michael Kansky is a serial entrepreneur, software founder, and AI-driven business operator with more than two decades of experience building companies at the intersection of customer engagement, automation, software, digital services, and data-driven growth.
Connect on LinkedIn