Those numbers aren't a CX issue — they're a design issue. Gladly's 2026 Customer Expectations Report breaks down exactly where AI-powered service loses customers, and what the architecture of loyalty-driven CX actually looks like.

Get the report

Hey!

You're paying for Claude Pro or Max.

But you keep hitting limits halfway through the day.

Here's what nobody tells you, you're probably wasting 70% of your budget on stuff that doesn't matter.

Let me show you exactly how to fix it. But before that, let's check what's happening this week👇

🎬 X adds simplified AI video generation that turns multiple images into a single short video using its Grok AI tools.

🤖 NVIDIA expected to unveil next-generation AI chip and major software updates at its annual GTC conference as the company pushes further into AI infrastructure and agents.

🔗 Instagram tests clickable links in post captions for paying users, potentially giving creators and brands new ways to drive traffic.

🧠 New research from Swansea University finds that AI tools can actually increase human creativity by helping users explore more ideas and design possibilities during problem-solving tasks.

🛡️ Meta introduces new measures to ensure original creators get proper credit when their content is reposted across its apps.

🎬 Google and Accel selected five startups from over 4,000 pitches for their AI accelerator program, specifically choosing companies building real AI technology rather than simple “AI wrappers.”

P.S. This image was created with ViralSky.ai. Just describe what you want, and it generates the image in seconds🙌

Claude doesn't count messages. It counts tokens.

One word equals roughly one token. One page equals about 300 tokens.

Here's what kills your budget: Claude re-reads your entire conversation every single time you send a message.

Your first message costs 200 tokens. Your 30th message costs 50,000 tokens because Claude is processing everything from the beginning.

Most people have no idea this is happening.

Haiku for quick tasks. Sonnet for real work. Opus only when Sonnet fails.

Opus costs 9x more than Haiku per message. Most people use Opus for everything and wonder why they run out.

Use Haiku for simple questions, quick edits, and basic research. Use Sonnet for coding, writing, and analysis. Use Opus only when you genuinely need the extra power.

This one change can double your effective usage.

When Claude gets it wrong, stop sending "can you fix that?"

Edit your original prompt instead. Delete the bad response. Try again.

Every follow-up message costs more because Claude re-reads everything. Editing the original keeps your conversation short and your token count low.

After 15-20 messages, start a new chat.

Long conversations get exponentially more expensive. Every new message processes thousands of tokens from earlier in the thread.

Starting fresh resets the counter. Your messages stay cheap.

Extended Thinking burns hidden tokens you never see.

It's great when you need deep reasoning. But most tasks don't need it.

Check the toggle before every conversation. Turn it off for simple work. Turn it on only when you're stuck on something complex.

If you’re a decision maker at your company, you need to be on the bleeding edge of, well, everything. But before you go signing up for seminars, conferences, lunch ‘n learns, and all that jazz, just know there’s a far better (and simpler) way: Subscribing to The Deep View.

This daily newsletter condenses everything you need to know about the latest and greatest AI developments into a 5-minute read. Squeeze it into your morning coffee break and before you know it, you’ll be an expert too.

Subscribe right here. It’s totally free, wildly informative, and trusted by 600,000+ readers at Google, Meta, Microsoft, and beyond.

Your budget refills every 5 hours from your first message. Not at midnight.

Burn through everything at 9 AM? New capacity opens around 2 PM.

There's also a weekly cap. Don't blow your whole week by Wednesday.

Quick math:

Plan your heavy sessions accordingly.

Before any heavy session, check Settings → Usage.

If you're at 80% of your weekly limit, switch to Haiku for lighter stuff. Save your remaining budget for what actually matters.

Don't find out you're out of capacity halfway through an important project.

Every PDF page costs 1,500-3,000 tokens.

Upload a 50-page document? That's 75,000-150,000 tokens gone before Claude even responds.

Only attach files you actually need Claude to read. Extract the relevant sections first. Don't dump entire books into the chat.

Web search and Research mode eat tokens fast.

They're powerful features. But every search adds overhead. Every link Claude fetches costs tokens.

Use them when you need current information. Don't use them for stuff Claude already knows.

If you're on Pro or Max, turn on Extra Usage in Settings.

This lets you keep working on pay-as-you-go when you hit limits.

But set a spending cap first. $20-50 is usually enough to cover overflow without surprises.

You'll never be completely stuck again.

Most people treat Claude like it's unlimited.

Then they run out by noon.

Pick the right model. Keep conversations short. Turn off what you don't need.

That's it.

All the best,

René