Ways To Improve Response Quality In Omnichannel Support

Let’s say your support center is getting hammered with repeat calls about a new product feature. Historically, the team would escalate, create a task force, and maybe update a knowledge base weeks later. With the tech available today, you should be able to unify signals from tickets, chat logs, and social mentions instead. This helps you quickly interpret the root cause. Perhaps in this case it's a confusing update screen that’s triggering the same questions. Instead of just sharing the feedback with the task force that'll take weeks to deliver something, galvanize leaders and use your tech stack to orchestrate a fix in real time. Don't have orchestration in that stack? Start looking into this asap. An orchestration engine canauto-suggest a targeted in-app message for affected users, trigger a proactive email campaign with step-by-step guidance, and update your chatbot’s responses that same day. Reps get nudges on how to resolve the issue faster, and managers can watch repeat contacts drop by a measurable percentage in real time. But the impact isn’t limited to operations. You energize the business by sharing these results in a company-wide standup and spotlighting how different teams contributed to the OUTCOME. Marketing sees reduced churn, operations sees lower cost-to-serve, and leadership sees a team aligned around outcomes instead of activities. If you want your AI investments to move the needle, focus on unified signals, real-time orchestration, and getting the whole business excited about customer outcomes....not just actions. Remember: Outcomes > Actions #customerexperience #ai #cxleaders #outcomesoveraction

✴️ Improving response accuracy with LLMs and Agents may not result in ticket deflection. 🔷 We learned that the hard way—with a recent customer. As early adopters of AI, they introduced chatbots. They iterated on various retrieval patterns. We brought in Knowledge Graphs, Graph RAG, and Agents. We boosted answer accuracy from 60% to over 80%. 🔴 But ticket deflection? Still under 5%. 🔬 So we rolled up our sleeves and dug in. 💡 What we uncovered was eye-opening. They didn’t need better AI. They needed a better system. We ran a range of analyses. Here’s what we found: 📈 Ticket Analysis 48% of tickets were repetitive, “how-to” queries. 📈 Content Analysis But 50% of those queries couldn’t be resolved—due to missing content, missing context, or poor curation. 📈 That left 24% of tickets that could have been deflected—if only there had been contextual content. 📈 User Journey Analysis 80% of users never engaged with existing content. They didn’t: Read the documentation Complete onboarding Search the help center 🔍 They Googled. Found little. Raised a ticket. 📈 That’s why self-serve initiatives resulted in less than 5% deflection— despite search accuracy improving to 80%. The results still didn’t change. 🛑 What shocked us was that users who did search... still raised tickets. 📈 Search Result Analysis 40% of users said the results seemed “sensible but not specific.” This is after UAT with SMEs rated search at 80% accuracy. 🔴 The system returned generic answers. Missed context. Ignored intent. And get this: They were 50% right— 50% of them were still doing keyword searches using one or two words in 2025! The gaps? 🔷 User personalisation 🔷 Contextual precision 🔷 Intent disambiguation It’s like hiring a genius— then locking them in a room with no clues. So we changed the system. 🔷 Introduced intent detection agents, personalization agents, Agentic RAG. 🔷 If the detected intent was “informational only,” we routed to RAG. Otherwise, we raised a ticket immediately. 🔷 Built continuous analyzers—for tickets, content, queries, logs, search quality, and user behavior. 🔷 These identified whether the issue was with content, queries, user behavior, or search itself. 🔷 Automated real-time escalations to content authors, search engineers, and UX designers—backed by hard data. In short, we built a system—not just an answering machine. The result? 🔷 Deflection jumped from under 5% to over 28%. 🔷 Content coverage, search accuracy, and user experience improved—with intervention attribution. 🔷 Most importantly, we now have a system we can monitor and manage. We’re not there yet— But we’re on a clear and a confident path to 50% deflection. 🟢 The truth? Great AI in a broken system still fails. But even average AI in a smart system can win. 🟣 So before you ask, “Why isn’t our AI working?” Ask this instead: Have we built the right world for it to work in?” Happy to share details on DM. ServiceNow

Two weeks ago I said AI Agents are handling 95% of our sales and support and I replaced $300k of salaries with a $99/mo Delphi clone. 25+ founders DM’d me… “HOW?” Here’s the 6 things you MUST do if you want to run your entire customer-facing business with AI: 1. Create a truly excellent knowledge base. Your AI is only as good as the content you feed it. If you’re starting from zero, aim for one post per day. Answer a support question by writing a post, reply with the post. After 6mo you have 180 posts. 2. Have Robb’s CustomGPT edit the posts to be consumed by AI. Robb created a GPT (link below) that tweaks posts according to Intercom’s guidance for creating content for Fin. The content is still legible to humans, but optimized for AI. 3. Eliminate recursive loops - because pissed off customers won’t buy If your AI can’t answer a question but sends the customer to an email address which is answered by the same AI, you are in trouble. Fin’s guidance feature can set up rules to escalate appropriately, eliminate loops, and keep customers happy. 4. Look at every single question every single day (yes, EVERY DAY). Every morning Robb looks at every Fin response and I look at every Delphi response. If they aren’t as good as they could possibly be, we either revise the response, or Robb creates a support doc to properly handle the question. 5. Make sure you have FAQs, Troubleshooting, and Changelogs. FAQs are an AI’s dream. Bonus points if you create FAQ’s written exactly how your customers ask the question. We have a main FAQ, and FAQs for each sub section of our support docs. Detailed troubleshooting gives the AI the ability to handle technical questions. Fin can solve 95% of script install issues because of our Troubleshooting section. Changelogs allow the AI to stay on top of what’s changed in the app to give context to questins about features and UI as it changes. 6. Measure your AI’s performance and keep it improving. When we started using Fin over 1y ago, we were at 25% positive resolutions. Now we’re above 70%. You can actively monitor positive resolutions, sentiment, and CSAT to make sure your AI keeps improving and delivering your customers an increasingly positive experience. TAKEAWAY: Every Founder wants to replace entire teams with AI. But nobody wants to do the actual work to make it happen. Everybody expects to flip a switch and have perfect customer service. The reality? You need to treat your AI like your best employee. Train it daily. Give it the resources it needs. Hold it accountable for results. Here’s the truth that the LinkedIn clickbait won't tell you… The KEY to successfully running entire business units with AI? Your AI is only as good as the content you feed it. P.S. Want Robb's CustomGPT? We just launched 6-part video series on how RB2B trained its agents well enough to disappear for a week and let AI run the entire business. Access it + get all our AI tools: https://www.rb2b.com/ai