📅 December 18, 2025 | ⏱️ 4-min read
🎁 Need a last-minute gift idea? Share a subscription to this newsletter with your colleagues and teammates. It’s FREE!
Good Morning!
The Executive Hook: The days of just turning on AI and hoping for the best are over. New reports show that “Agentic AI” (bots that actually do tasks) might break your customer service before fixing it. And honestly? That might be exactly what we need to see where things are broken.
Last year, we gave the keys to the bots. Next year, we’re going to realize they don’t know how to drive yet. The new goal isn’t just “automated”—it’s “trustworthy.”
🤿 THE DEEP DIVE: Agentic AI Will Break Your CX
The big picture: A provocative new analysis from Forbes suggests that the rush to deploy “Agentic AI”—systems that can take action, not just chat—is exposing deep operational fractures.
What’s happening: Agentic AI doesn’t just answer questions; it tries to do work. In doing so, it is crashing into the “invisible workarounds” your human agents use every day to keep things moving.
The friction point: When an AI agent hits a process gap (like a policy that doesn’t match reality), it fails or hallucinates. Humans used to fudge this; AI exposes it.
The risk: Leaders are seeing a temporary dip in resolution as these hidden inefficiencies are brought to light.
Why it matters: You can’t automate a mess.
Strategic shift: The Forbes report argues you must “break” a journey in a controlled environment to fix it.
If you layer Agentic AI on top of broken processes, you are just scaling chaos at the speed of light.
The takeaway: Don’t blame the bot. Blame the process it just tripped over.
Source: Forbes
📊CX BY THE NUMBERS: The Headcount Paradox
Data Source: Gartner 2026 Service & Support Survey and interview with Brad Fager, Chief of Research for the Customer Service & Support Leaders practice group.
80%: The percentage of organizations planning to reduce agent headcount in the next 18 months.
91%: CX leaders under executive pressure to use AI specifically to improve CSAT (not just cut costs).
58%: Leaders planning to upskill remaining agents into “Knowledge Management Specialists” to curate the AI’s brain.
The Insight: We are trading volume for value. The remaining humans aren’t answering phones; they are training the models that answer the phones.
Source: Gartner
Gartner clients can learn more in the report: 2026 Survey Results: Service and Support Leaders’ Goals and Game Plans.
🛠️ THE AI TOOLBOX: Concierge AI
What it does: Think of this as a super-smart librarian for your company website. It learns everything from your documents and brand guides. Then, when a customer asks a hard question, it gives a specific, correct answer instantly.
CX Use Case:
Self-Help: It lets customers find answers to technical or pricing questions without waiting for a salesperson.
No More Guessing: Because it only studies your content, it doesn’t make things up (hallucinate) like other chatbots might.
Listening In: It keeps track of what people are asking, so you know exactly what information is missing from your site.
Source: Concierge AI
⚡ SPEED ROUND: Quick Hits
The Friction War: Fast Company reports that Agentic AI is finally ending the “era of customer friction” (e.g., long hold times designed to make you give up). Bots don’t get tired of holding. Source: Fast Company
Alteryx + GenAI: Alteryx has launched new “governed” analytics tools, bridging the gap between wild GenAI creativity and the strict data governance needed for enterprise trust. Source: PR Newswire
Stock Surge: CXAI shares jumped nearly 10% after their CEO revealed a massive rollout of Agentic AI pilots scheduled for Q1 2026. The market is betting big on autonomy. Source: Stock Titan
📡 THE SIGNAL: The “Process Tax” Bill is Due
We are entering a phase of “Operational Honesty.”
For the last two years, we treated AI like a shiny wrapper. We put it over our old, messy computer systems and hoped the AI would figure it out. It didn’t.
The shift for tomorrow: The best leaders in 2026 won’t be the ones with the fanciest AI prompts. They will be the ones who fix their workflows.
Your challenge: Pick one broken thing your customers have to do (like getting a refund or resetting a password). Don’t just add AI to it. Map it out. Find the step where a human usually has to “cheat” the system to get it done. That is where your AI will fail. Fix the road before you build the Ferrari.
See you tomorrow,
👥 Share This Issue
If this issue sharpened your thinking about AI in CX, share it with a colleague in customer service, digital operations, or transformation. Alignment builds advantage.
📬 Feedback & Ideas
What’s the biggest AI friction point inside your CX organization right now? Reply in one sentence — I’ll pull real-world examples into future issues.