Start free

Sources

Every number and rule we quote on this site, with its source and the day we checked it.

  • 1Arizona State University, 2020: solving one problem took customers about three contacts, and about one in ten needed six or more.Checked 3 October 2026Open the source →
  • 2SQM Group: in call centres, every 1% gained in first-call resolution brings 1% more customer satisfaction.Checked 3 October 2026Open the source →
  • 6Gartner, 2024: 64% of customers would prefer that companies didn't use AI for customer service.Checked 3 October 2026Open the source →
  • 7COPC: an 86% overall quality score hid 60% accuracy on what mattered to customers.Checked 3 October 2026Open the source →
  • 10Eurogroup Consulting: repeat contacts within 7 days under 10% are good, under 5% excellent, over 15% a warning.Checked 3 October 2026Open the source →
  • 11Consortium for Service Innovation, KCS v6: a new fix is written in the customer's own words so it can be found again.Checked 3 October 2026Open the source →
  • 12York University, 2024: warmth at the start and end of a sales chat, efficiency in the middle.Checked 3 October 2026Open the source →
  • 13Journal of Consumer Research: concrete words (an action, an item, a date) lift satisfaction and spending.Checked 3 October 2026Open the source →
  • 14American Marketing Association: after most service failures, “thank you” works better than “sorry”.Checked 3 October 2026Open the source →
  • 15UK Trading Standards allergen guidance: no answer from memory; “I can't confirm that” is acceptable.Checked 3 October 2026Open the source →
  • 16Auto Care Association: data standards for parts and vehicle fitment.Checked 3 October 2026Open the source →
  • 18PCI Security Standards Council: taking card numbers over messaging brings that channel into the card security standard.Checked 3 October 2026Open the source →
  • 20MaestroQA: more than 265,000 tickets, and no link between customer satisfaction and quality scores.Checked 3 October 2026Open the source →
  • 22Crolic and others, Journal of Marketing, 2022: a human-like bot lowers the satisfaction of angry customers.Checked 3 October 2026Open the source →
  • 23Meta for Developers: the “typing…” indicator on WhatsApp, Messenger and Instagram.Checked 3 October 2026Open the source →
  • 27EU AI Act, Article 50: people must be told when they're dealing with AI.Checked 3 October 2026Open the source →
  • 28California Business and Professions Code: a bot may not mislead a buyer into thinking it's human.Checked 3 October 2026Open the source →
  • 29Messenger Platform policy: disclose automated chats where the law requires it.Checked 3 October 2026Open the source →
  • 34US health privacy rule, 45 CFR 164.514: verify identity before disclosing results.Checked 3 October 2026Open the source →
  • 35Qatar, Law No. 13 of 2016: health data is personal data of a special nature.Checked 3 October 2026Open the source →
  • 36Saudi Arabia, Personal Data Protection Law: health data is sensitive data.Checked 3 October 2026Open the source →
  • 37Qatar Central Bank AI guidelines: customers must be told when they deal with an AI system (Pinsent Masons, 26 September 2024).Checked 3 October 2026Open the source →