惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

Apple Machine Learning Research
Apple Machine Learning Research
OSCHINA 社区最新新闻
OSCHINA 社区最新新闻
G
Google Developers Blog
博客园 - 司徒正美
J
Java Code Geeks
aimingoo的专栏
aimingoo的专栏
A
About on SuperTechFans
博客园 - 三生石上(FineUI控件)
WordPress大学
WordPress大学
T
The Blog of Author Tim Ferriss
D
Docker
大猫的无限游戏
大猫的无限游戏
D
DataBreaches.Net
腾讯CDC
V
Visual Studio Blog
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
C
Check Point Blog
M
MIT News - Artificial intelligence
Jina AI
Jina AI
I
InfoQ
雷峰网
雷峰网
The Cloudflare Blog
美团技术团队
Engineering at Meta
Engineering at Meta

Amplitude

I was the bottleneck Your agents are only as good as your data context The New Trust Economy in Financial Services See how Agent Analytics governs agent trace data with access controls and privacy redaction making it safe to send us your agent traces. Connecting Agent Performance to Product Outcomes The Hidden Cost of a Bad AI Answer Verification loops in software factories: What PMs need to know Your AI Chat Box Isn't a Strategy. Where You Put It Is. The Neobank Squeeze: Why Classic Banks Need to Learn to Move at Challenger Speed From Context to Intelligence: Building Self-Improving Products in Banking and Insurance Practical Guide to Evaluating Open Models: Achieving Sonnet-level performance with Kimi AI Can Build. Can It Know What Worked? Your Analytics Just Joined Discord How HYBRD Turned Agent Evals into a Retention Signal Use Amplitude with Your Existing Adobe Tags Beyond the Rate: Retail Banking's New Competitive Front How NS Prevented €1.8M in Revenue Loss Through Experimentation Go from Product Launch to Insight to Action in Minutes What Makes a Good vs Bad North Star Metric The Role of Feature Management in Successful Product Development Cohort Retention Analysis: Reduce Churn Using Customer Data 7 Steps to Measuring the Success of a Feature 14 Best Product Management Tools for 2026 (Plus Tips from Senior PMs) The Definitive Guide to Behavioral Cohorting Putting A Number On AI Quality Meet the Winners of the 2026 Amplitude AI Impact Awards Beyond Last-Touch Attribution: Find Out Which Interactions Really Matter Agent Connectors Are Better Together Agents That Act on What Actually Happened How Square Used Amplitude to Enhance the Seller Experience and Power Growth
Optimizing Core Web Vitals with Amplitude’s Global Agent
Alicia Chang · 2026-05-20 · via Amplitude

Your pages are slow, and users leave before your content loads. Google notices too. Core Web Vitals—Largest Contentful Paint (LCP), Interaction to Next Paint (INP), and Cumulative Layout Shift (CLS)—directly affect both user experience and search rankings. If you’re not monitoring them alongside real user behavior, you’re guessing at what to fix.

Amplitude Web Vitals connects your site’s performance scores to actual user sessions, so you see exactly where pages break down and who gets affected. Pair it with Global Agent, and you can diagnose issues, prioritize fixes, and pull up Session Replay evidence using plain-language prompts instead of building charts from scratch.

How it works: Three prompts, zero chart-building

Global Agent is Amplitude’s AI-powered assistant that answers questions about your data in natural language. Instead of configuring charts and filters manually, you type a question and get back analysis, visualizations, and direct links to sessions. Here's what that looks like for Core Web Vitals.

“How are Web Vitals looking this month?”

I asked Global Agent this question about my demo site, Amplifile. It pulled 617 Web Vitals observations from the month to date, identified my highest-traffic pages (/Dashboard.html, the home page, /sign-in.html), and generated a daily trend chart broken down by page. It also caught a traffic pattern: visits peaked March 3-4, dipped over the weekend (normal for a B2B product), and rebounded March 11.

That’s a site-wide performance snapshot from a single prompt. No chart configuration, no filter setup, no waiting on an analyst. While the trend was acceptable, Global Agent took it further by suggesting what to look into next.

“Show me which pages have poor Core Web Vitals scores”

Global Agent filtered to only poorly rated LCP, INP, and CLS events and grouped them by page. The home page was the biggest CLS offender with 17 poor events. /Dashboard.html had the most LCP issues at 10 poor events. INP was relatively contained across all pages.

More importantly, Global Agent handed back a prioritized list of pages to fix first, ranked by severity and traffic volume. That's the difference between “something feels off” and a ranked backlog your engineering team can act on today.

“Pull Session Replays from poor CLS sessions on home page”

Because I have Web Vitals and Session Replay consolidated in Amplitude, Global Agent found 10 sessions with poor CLS scores on the home page and returned a table with dates, session durations, and direct replay links. I opened the most recent one: a 30-minute session from March 12. The visitor was a new organic user in San Francisco on Chrome 144. In the replay, content visibly jumped as the page loaded, pushing buttons and text out of position.

Several of those sessions were under a minute, which Global Agent flagged as likely bounces caused by the layout shift. That’s the kind of connection that’s almost impossible to make with performance data alone. You need to see what the user saw so that you can validate the issue and describe the pain point to your engineering team. This gets you there without filing a ticket or switching tools.

What each prompt surfaces

Start diagnosing your site’s performance

Web Vitals, Session Replay, and Global Agent work together in Amplitude to connect performance scores to real user behavior. You don’t need a separate monitoring tool, a separate session replay tool, and a separate analytics platform.

Try It Now

Ask Global Agent about your Core Web Vitals and see what comes back. If you’re not on a plan that includes Web Vitals yet, explore Amplitude’s AI Agents to see the full capability set.