Assertion Supported AI assessment confidence: 90% certainty 4/5 debate potential 2/5

Rory O'Driscoll: LLM-powered support tools double call resolution rates to 60%

Rory O’Driscoll · Anthropic Raises $30B from Microsoft & NVIDIA & NVIDIA’s Core Business Faces TPU Threat · 20VC with Harry Stebbings · Nov 27, 2025 · at 37:35

Rory O'Driscoll is a General Partner at Scale Venture Partners. He contrasts historical customer support software performance with current LLM capabilities.

0:00 / 0:20exact quote · 20.0s

▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark

“Before LLMs, and I had an investment in a company that was pre-LLM customer support. Our resolution rate was around 23%. In other words, we could solve 23% of calls, maybe up to 30. With LLMs, with these new modern things, you can solve 60%, which means on a number of call basis, you can significantly reduce your customer support.”

quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →

The claim, plainly

Rory O'Driscoll states that pre-LLM customer support software resolved 23% to 30% of customer calls, whereas modern LLM-based support software achieves a 60% resolution rate.

a quote alone can be ambiguous, so this restates it with the references and context of the conversation around it resolved. AI-written; the tape below the player is the source.

Assessment

Rory O'Driscoll's statements regarding pre-LLM and modern LLM-based customer support resolution rates are supported by independent industry benchmarks and case studies. Multiple industry reports confirm that legacy rule-based chatbots and pre-LLM support tools typically achieved automated resolution or FCR rates of 20% to 30% (e.g., MavenAGI and Zowie benchmarks). For modern LLM-based support systems, major platform benchmarks and case studies consistently confirm typical resolution rates of around 60% (e.g., Klarna achieving 60%-67%, Intercom Fin averaging 50%-67% with a 65% guarantee threshold, and industry benchmarks ranging from 40%-60% to 55%-70%). Because both legs of O'Driscoll's claim align closely with industry data, the claim is true.

MavenAGI and Zowie benchmark reports confirm legacy pre-LLM rule-based chatbots typically resolved between 20% and 30% of customer interactions.
Klarna's LLM-powered AI assistant reported automated resolution rates of 60% to 67%.
Intercom Fin AI Agent benchmarks report average production resolution rates of 50% to 67% (with a 65% performance guarantee), and Notch industry reports benchmark AI assistants at 40% to 60% and 55% to 70%.

Checked against

[1] kayako.com [2] lorikeetcx.ai [3] devrev.ai [4] getvocal.ai [5] irisagent.com [6] getzowie.com [7] notch.cx [8] twig.so

Klarna AI assistant resolution rate customer support 66 percent 60 percent "legacy chatbots" "20%" OR "25%" OR "30%" resolution rate customer support "resolution rate" "20% to 30%" OR "23% to 30%" customer support chatbot "deflection rate" "20%" OR "30%" pre-LLM rule-based chatbots "resolution rate" 60% LLM customer support provider report Intercom AI customer support resolution rate 60 percent chatbot resolution rate pre LLM LLM 60 percent resolution rate customer support pre LLM customer support resolution rate 20 percent 30 percent

assessed as of 2026-07-31

This assessment is an AI reading of a public statement, made against the cited references with the confidence shown. It can be wrong. The quote, the tape and the references are all on this page; if something is off, use report an error below. full disclaimer →

Mentioned

LLMs pre-LLM customer support

⚑ Spot an error? Report it

Rory O'Driscoll: LLM-powered support tools double call resolution rates to 60%

Turn any episode into a week of clips.