The AI Consumer Experience Index 3Q 2026 – In-Depth Report

Author: Layne Haaksma, Senior Research Analyst

The Metrigy AI Consumer Experience Index fell to 102.7 in the third quarter, down from 104.6 in the second quarter but still above the 100-point baseline established in the first quarter. All eight sub-indexes eased—somewhat unsurprising given the growth in activities the third-quarter months naturally bring. Across all three waves of the index, the second-quarter ratings represent the peak rather than the new normal, with consumer sentiment settling at a level that remains better than where the year started.

What moved in third quarter was the standard more than the service customers received. The evidence sits in which responses changed. Consumers did not shift into the negative answers this quarter; they shifted into the middle. The Relationship Index shows this shift most plainly. The share of consumers who say their treatment was about what they expected jumped 5.6 points to 42.3%, the largest single response movement anywhere in the study, while the share who rate it much better than expected fell to 18.7%. Consumers are telling us their treatment was unremarkable, not that it was bad. Competent service, which incorporates competent AI, has become the price of admission rather than something worth praising. Consumers have raised the bar for exceptional service.

AI is doing the work of holding the index up. AI Satisfaction at 109.4 and AI Trust at 106.9 are the only two sub-indexes more than six points above baseline, and between them they hold about 70% of the gain left in the overall index. AI Trust posted the smallest decline of the eight sub-indexes this quarter, and that matters more than it sounds. The survey field period ran through a stretch of heavy negative press about AI’s data center footprint, yet consumer trust did not break. Pair that with heavy summer travel, storm seasons, back-to-school shopping, and overall raised AI expectations, and the trust score speaks volumes for AI’s performance.

The majority of consumers still prefer human agents over AI agents, but the percentage has been declining since the first quarter. When a resolution is guaranteed, more consumers now will select AI than they did in the first quarter: 28.8% today vs. 23.0% in first quarter, showing a slow but steady shift toward preference for AI agents. What’s immeasurable is the impact AI has had on their human preference, given agent assist continues to improve human agent performance, though customers don’t realize what’s happening behind the scenes.

Table of Contents

• Executive Summary
• Key Takeaways
• Purpose, Methodology, & Respondent Profile
• The 3Q 2026 AI Consumer Experience Scorecard
  • AI Satisfaction Index: 109.4
    AI Trust Index: 106.9
    Proactivity Index: 104.9
    CX Ease Index: 101.9
    Empathy Index: 100.8
    Resolution Index: 100.8
    Momentum Index: 100.1
    Relationship Index: 98.6
• What Changed Between Second and Third Quarter
  • Exposure
  • Experience
  • Expectations
• The Consumer Buying Personas
  • The AI Superstars
  • The Premium Buyers
  • The Silent Multitaskers
  • The Human Loyalists
  • The Underserved Skeptics
• General Customer Service
  • Voice Still Leads, but Text is Closing
  • What Drives Channel Choice
  • The Stakes for Businesses
• The Human Safety Net: Why We Still Prefer People
  • What Consumers Say When We Ask Them Why
  • The Preference Is Earned, Not Held
  • The Business and Consumer Gap
  • Trust Over Resolution
  • The Escalation Pattern
• Industry Performance Analysis
  • Retail: Still the Leader in CX Delta
  • Financial Services: Best AI Implementation Data
  • Healthcare: High Volume, Polarized AI
  • The Bottom Tier: Government Agencies
• CX Tolerance and Breaking Points
• The Evolution of Proactive Outreach
• Strategic Recommendations
  • Design for the Rising Bar
  • Re-segment Your Digital Investment One Cohort Older
  • Sell Choice, Not Access
  • Make Comprehension the Roadmap
  • Fix the Handoff Before the Greeting
  • Protect the Two Segments You Are Most Likely to Lose
  • Cap Outreach at Three a Week, Then Segment Up
  • Watch the Vertical Tails, Not the Vertical Averages