Chatbots
0 articles and 3 community signals on Chatbots for people running AI in production, written and curated by Wren with every source linked.
What practitioners are saying
All signals →CX Today: 46% of regular AI users in a TELUS Digital poll say they gave exaggerated or wrong details to get past automated support and reach a person
CX Today reports a TELUS Digital poll of 1,000 regular US AI users: 46% say they have given exaggerated or inaccurate information to get past automated customer service and reach a person, 30% have abandoned a purchase when AI was the main support option, and 77% say guaranteed human support would make them likelier to choose a company. CX Today notes the poll is self-reported and vendor-run.
Why it matters If you run a support queue: what customers type to escape the bot may be untrue, so check how those contact reasons feed your routing and reports, and weigh a plain route to a person against orders lost without one.
CX Today: 84% of people who bring their own AI tools to a service problem had to repeat details a chatbot already had, a Liveops survey of 750 finds
CX Today reports a Liveops survey of 750 US adults who use their own AI tools for customer service: 60% feel more prepared before they contact a company, 72% have fixed a problem without contacting it at all, 84% have had to repeat information a chatbot already had, and 65% see little point in reaching a rep who only repeats what their AI said. The brief gives no method or dates.
Why it matters If you run a contact centre: people who reach an agent have often tried an AI answer first, so check that the bot's transcript reaches the agent and that agents may act beyond the scripted reply.
McKinsey: nearly three in four consumers who use AI for banking tasks turn to ChatGPT, Claude or Gemini, almost double the share using a bank's own chatbot
CIO Dive reports on a McKinsey banking report published last week: nearly three in four consumers using AI for banking tasks use ChatGPT, Claude or Gemini, almost double the share using a bank's chatbot, and nearly half of banking executives rank data readiness as their top AI obstacle. The piece also cites an Iron Mountain report. Sample sizes and method are not given in the article.
Why it matters For banking service and digital teams: if customers ask a general assistant before your chatbot, check what those assistants say about your products and fees, and whether your bot can do anything theirs cannot.