Our Phones Could Not Recognise Customers Whose Numbers Had Spaces in Them
When we moved our credit control team to a new telephone system, one of the features we were most pleased with was screen popping. When a customer rings, the phone system looks up the caller's number in our customer system and opens their account on the agent's screen before the call is answered. In the pilot it worked for the team's test accounts. After go live it recognised fewer than a third…
Our credit control team experienced issues when they transferred to a new telephone system. The system's screen popping feature, which displayed a customer's account when a caller rang, worked well with test accounts during the pilot but recognized fewer than a third of callers after implementation. Agents had to request account numbers from every customer, as they had done previously, leading to complaints.
The source of the problem was not the phone system itself, but rather the data in the customer system. The system received the caller's number in international format, including the plus sign and country code. However, the customer system stored phone numbers in free text fields that had been manually entered for years. This led to over forty different patterns for phone numbers, including spaces, brackets, international prefixes, dashes, extensions, and numbers separated by slashes.
The lookup system compared the incoming number with the field exactly, resulting in a failure when a number contained a single space. Each number could be dialled by a person, but the phone numbers were now stored in a standard international format with no spaces or punctuation, while the original text was kept for display. An automated cleanup process converted about ninety percent of the numbers, and the credit controllers corrected the remaining ones over two months as customers called.
New numbers are now validated and converted during entry, with extensions stored separately and each contact able to have multiple numbers. The new lookup system matched about eighty-five percent of calls, with the remainder being withheld numbers or numbers the system had never held. The source of the issue was the human-typed data, which was not suitable for machine comparison. The team was only made aware of the problem after the system started requesting account numbers.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
