# CraftCX Blog

[RSS Feed](/blog/rss.xml)

[![](/_vercel/image?url=%2Fblog%2Fog%2Fa-support-ops-guide-to-inventing-work.jpg&w=1200&q=75) A Support Ops Guide to Inventing Work How Support Ops can find recurring problems, spot missed customer needs, and turn those signals into changes that improve support. ![](/_vercel/image?url=%2Fblog%2Fjason-dugdale.webp&w=48&q=75) Jason DugdaleSeptember 29, 2026](/blog/a-support-ops-guide-to-inventing-work)

[![](/_vercel/image?url=%2Fblog%2Fog%2Fmoving-support-decisions-to-jev.jpg&w=1200&q=75) Small questions, measurable gains with Jev Two small evals found higher recall for ticket insights and lower cost and latency for signal grouping. ![](/_vercel/image?url=%2Fblog%2Fjason-dugdale.webp&w=48&q=75) Jason DugdaleSeptember 24, 2026](/blog/moving-support-decisions-to-jev)

[![](/_vercel/image?url=%2Fblog%2Fog%2Fhelp-center-grounding.jpg&w=1200&q=75) Keeping AI answers grounded in the help center you trust Check whether AI answers apply the right guidance, distinguish content gaps from answer errors, and review conversations after an article change. ![](/_vercel/image?url=%2Fblog%2Fjason-dugdale.webp&w=48&q=75) Jason DugdaleSeptember 21, 2026](/blog/keeping-ai-answers-inside-the-help-center-you-already-trust)

[![](/_vercel/image?url=%2Fblog%2Fog%2Ffin-analytics.jpg&w=1200&q=75) Getting more from Fin analytics with conversation evidence Turn a change in Fin's results into a specific follow-up by reviewing comparable conversations and checking what changed for customers. ![](/_vercel/image?url=%2Fblog%2Fjason-dugdale.webp&w=48&q=75) Jason DugdaleSeptember 18, 2026](/blog/getting-more-from-fin-analytics-with-conversation-evidence)

[![](/_vercel/image?url=%2Fblog%2Fog%2Fai-support-qa.jpg&w=1200&q=75) Support quality assurance when AI handles most of the volume Build an AI support QA process with clear review criteria, sensible sampling, and a way to check whether fixes worked. ![](/_vercel/image?url=%2Fblog%2Fjason-dugdale.webp&w=48&q=75) Jason DugdaleSeptember 18, 2026](/blog/support-quality-assurance-when-ai-handles-most-of-the-volume)

[![](/_vercel/image?url=%2Fblog%2Fog%2Fchatbot-analytics.jpg&w=1200&q=75) Chatbot analytics that connect the score to the conversation Choose useful chatbot metrics, compare like-for-like results, and use conversation evidence to decide what to improve. ![](/_vercel/image?url=%2Fblog%2Fjason-dugdale.webp&w=48&q=75) Jason DugdaleSeptember 17, 2026](/blog/chatbot-analytics-pairing-the-score-with-the-conversation)

[![](/_vercel/image?url=%2Fblog%2Fog%2Fdeflection-rate-evidence.jpg&w=1200&q=75) What else to look at when you have a strong deflection rate Deflection is useful. Conversation evidence explains whether a closed AI support conversation was a good outcome for the customer. ![](/_vercel/image?url=%2Fblog%2Fjason-dugdale.webp&w=48&q=75) Jason DugdaleSeptember 16, 2026](/blog/what-else-to-look-at-when-you-have-a-strong-deflection-rate)

[![](/_vercel/image?url=%2Fblog%2Fog%2Fintroducing-craftcx.jpg&w=1200&q=75) Introducing CraftCX CraftCX helps support and product teams review AI support quality and find customer problems in the conversations their agents handle. ![](/_vercel/image?url=%2Fblog%2Fjason-dugdale.webp&w=48&q=75) Jason DugdaleSeptember 15, 2026](/blog/introducing-craftcx)