Notes from building AI for customer operations
How to run a churn analysis with ChatGPT: prompts, checks and the mistakes to avoid
October 15, 2026 · Sergio Palmero
ChatGPT runs a churn analysis in minutes. The prompts for each step, example outputs, and the checks that catch the numbers it gets wrong.
Why (and how) to build your own specialized AI on top of Zendesk (or any ticketing system)
October 13, 2026 · Sergio Palmero
Your helpdesk is the platform and the AI is the hire. When the built-in AI is enough, when it falls short, and how to build a specialized agent on top.
How to convert more free users to paid in PLG, and when to say no
October 9, 2026 · Sergio Palmero
Free users convert when they aren't left alone and when you know who not to sell to: onboarding around their problem, follow-ups and support that spots buyers.
8 keys to take your support AI from 20% to 95% resolution
October 8, 2026 · Sergio Palmero
Connecting docs to Claude or ChatGPT resolves the easy 20 to 30%. Guards, account context, clean docs and permission to act take a support AI to 90%+.
How to reduce and prevent churn with Claude: the analysis and experiment loop
October 6, 2026 · Sergio Palmero
Reducing churn with AI starts with analysis: find what churn is made of, which signals predict it, and test interventions against a real baseline.
What is the best AI agent for customer service? A comparison for SaaS teams
October 2, 2026 · Sergio Palmero
There's no single best AI agent for customer service. How Fin, Zendesk AI, Agentforce, Sierra, Decagon, Ada and Zentesimal fit different SaaS teams.
How to use AI chatbots to sell more: turn support into a sales channel
September 30, 2026 · Sergio Palmero
Most support chatbots are built to answer. How to make them spot buyers, start the conversation and know when not to sell.
Which sources to connect to an AI Assistant (and which to filter out)
September 29, 2026 · Sergio Palmero
Answer rate climbs in steps: 35-50% with docs alone, 50-65% with past conversations and team chat, 65-80% with customer data.
How to restructure a support team for AI without firing anyone
September 17, 2026 · Sergio Palmero
Restructuring support for AI moves people up a ladder, not out the door. The five stages, the KPIs that change at each one, and where the freed hours go.
The CRM Was Built for Human Memory
September 15, 2026 · Sergio Palmero
AI doesn't need a clean CRM. It reads tickets, calls, and product data straight from the source, so nothing depends on someone typing it in.
Building your own AI for customer ops: when does it pay off?
September 10, 2026 · Sergio Palmero
Building your own AI for customer ops pays off with the right team, timeline, and maintenance plan. See when buying instead works out better.
How to reorganize your customer experience team when AI arrives
September 8, 2026 · Sergio Palmero
How CX and post sales teams restructure when they bring in AI: which roles shift first, how fast each profile adapts, and how to lead the change.
How to Add AI to Your CS Stack Without Paying the Tool Switching Cost
September 4, 2026 · Sergio Palmero
Most AI rollouts price in the software, but forget to price what it costs the team to relearn a new tool.
Beyond "Not Escalated": How Not to Measure AI Support Success
September 2, 2026 · Sergio Palmero
Containment rate looks like success but hides silent churn. Here's the four state model that catches it before customers walk away.
How to Measure AI ROI in Customer Support
August 28, 2026 · Jose
Cost per ticket is only one lever. Here’s how CX teams measure AI worker ROI across OPEX, revenue, and the hours the team gets back.