Zendesk

We’re spotlighting RET International as we count down to this year’s Tech for Good Impact Awards. RET International supports communities made vulnerable by violence, conflict, and disaster, with a focus on youth and women. With Zendesk, the organization is streamlining case management, enabling youth-led community help desks, and using data dashboards to better understand and respond to community needs. This work will help expand access to education, protection, employability, and participation for displaced and vulnerable young people. Stay tuned as we continue to highlight organizations creating meaningful impact. This year’s Tech for Good Impact Award winners will be announced soon: https://zdsk.co/4cwYDUr

AI can no longer be a black box. Brands must be able to stand behind every automated decision. In this Zendesk Insights partner blog, Route 101 Ltd.’s Luke Cuthbertson explores why transparency, clear reasoning, and strong governance are essential to building customer trust in AI-powered service: https://zdsk.co/4wY6aUw

Over the last three years, the conversation around AI in support has shifted massively: 2024: "Should we deploy AI?" 2025: "Where can we plug AI into our support tools?" 2026: "Is our AI actually resolving tickets on its own?" Having AI in CX is no longer a differentiator. The real divide is between answer-only bots and action-taking AI. In the Third Edition of our State of AI in CX Report, we map out how support organizations evolve across three stages of AI maturity—moving from basic FAQ assistance to systems that execute multi-step tasks and adapt through automated QA feedback loops. If you’re evaluating your AI setup or trying to move beyond basic deflection metrics, this report offers a clear framework for what to do next. Find out more about Forethought AI Agents by Zendesk and download the report here: https://lnkd.in/gZkwdHyh

It's 2026 and somehow 40% of customer support work is still eaten up by repetitive, low-complexity tasks. 🤯 Even after years of teams deploying AI, frontline agents are still stuck dealing with tools that are missing customer context and nonstop manual workarounds. This is because basic chatbots only answer questions. They can't actually finish the work that needs to be done. In our third edition of our State of AI in CX Report, we looked at data across hundreds of support teams to see what separates shallow AI chatbot rollouts from real operational ROI. The biggest takeaway is simply having AI in your tech stack is no longer an advantage — it's just the baseline. The teams seeing real gains are moving toward systems that take action, handle end-to-end workflows, and learn from service history. Get the full report here 👉 https://lnkd.in/gZkwdHyh

Two weeks ago, I wrote about a number of the specialist roles in customer service converging into a new role we call the AI service architect. This is one of the biggest changes we've seen to customer support roles and worth a deeper dive. The job of the AI service architect is to take what your best agents do to resolve complex issues and capture the process precisely enough that AI can replicate it reliably. Teaching AI is a lot like onboarding a new employee. You give a new hire context, knowledge, access to your systems, data, and some sense of when to act and what to ask. An AI agent needs all of that too. The difference is that a new hire picks up a huge amount of context through osmosis. They learn just by sitting near people who know the job. Organizations need to replicate this context graph for AI. Some of it you write down and some comes from systems integrations, but all of it is deliberate. Making this transition requires significant change management and the instinct is to hire for it. Usually that is wrong. The person who knows which cases go sideways and why is already on your team. What they lack is time, permission, authority, and guidance. Everyone in service can write a help center article or a standard operating procedure for another human, but writing context complete enough to hand to AI agents requires a different set of skills entirely. First, structured knowledge captures the policies and rules the business really runs on. These are more prescriptive and procedural than public-facing content. Next, data provides complete customer context, insight into the conversation at hand, and enough history to see how similar cases played out. Finally, systems integrations provide a deeper understanding of the customer and the ability to take action on their behalf. Underneath all three sits evaluation, the part almost everyone skips. A procedure should constantly be tested against interactions you have already resolved previously. That is what makes this work closer to engineering than administration. You pick one intent, get it right, test it, ship it, keep testing it, then start on the next. That is the hill climbing that raises automation rates, sprint by sprint. This entire role is new, fluid and still a work in progress. Agility and flexibility are the most important qualities when facing such a disruptive technology. There will need to be training, ramp time, and room to experiment and learn. Some agents may worry that they are training their replacement. They are not. They are handing over the part of the job that no longer needs a human and getting back to their core purpose. The work stops being about resolving one case at a time and becomes about making sure every case is resolved as well as it possibly can be.

Not every AI tool is right for every agent. With group permissions, you can control access to Copilot features like enhance writing, similar tickets, and merge suggestions at the team level — not just account-wide. Pilot features with specific groups, tailor AI usage by role, and maintain tighter governance across your organization. Watch the demo to see how it works: https://zdsk.co/4qfbPmO

Automation follows the rules. Agentic AI knows when the rules are not enough. Read this Zendesk Insights blog to see how agentic AI reasons, adapts, and resolves work across the full service journey — and why service leaders are moving beyond task automation: https://zdsk.co/4bLvnJz