So while AI for coding and software engineering has seen some of the most explosive start-ups, from Codeium to Loveable to Poolside to Claude and OpenAi’s own releases here and more, just as big if quieter has been the change to contact center and support.
AI has ripped through the contact center and every leader from Talkdesk to Dialpad to Zendesk to Gorgias have adopted AI as core to their platform. And in many cases, also seen the number of human customer success reps already reduced by -20% to -30%. Already.
How far can it go though? Can AI replace almost everything a human does today in customer support?
I don’t know but perhaps what’s just as important is what customers are expecting, thinking, and buying based on.
That’s why I found 2 recent announcements pretty telling.
First, Gartner has just predicted AI will autonomously reduce support support issues by 80% and the costs for support by 30% by 2029.
Second, Salesforce claims AgentForce can already resolve 84% of customer support issues via AI! Though … that does sound a bit high 🙂
More on that here:
Now honestly I think Salesforce is exaggerating a bit or maybe at least flattering the metrics a smidge. What Intercom, Gorgias, Zendesk, etc. are all seeing today is that AI can realistically close up to 40% of tickets today, but often a much smaller percent of the tough ones. Those are often escalated to humans.
Some customers are pushing this well above 40% today, but the quality of resolution does decline. 80% for most is impractical to resolve all customer support issues with a high CSAT today with AI. At least for most customers.
But 80% by 2029? Sounds pretty plausible given that many are at 30%-40% real, honest resolution by AI today.
