In partnership with

ChatGPT gives you generic answers because you give it generic prompts.

You know the fix: longer prompts, more context, clearer constraints. But typing all that takes five minutes per prompt, so you shortcut it. Every time.

Wispr Flow lets you speak your prompts instead of typing them. Talk through your thinking naturally — include context, constraints, examples — and get clean text ready to paste. No filler words. No cleanup.

Works inside ChatGPT, Claude, Cursor, Windsurf, and every other AI tool. System-level, so there's nothing to install per app. Tap and talk.

Millions of users worldwide. Teams at OpenAI, Vercel, and Clay use Flow daily. Free on Mac, Windows, and iPhone.

Some weeks are slow news cycles. This was not one of them.

Between May 21 and 26, Google held I/O and announced the biggest change to its search product in 25 years. Microsoft quietly published a framework redefining what it means to be indexed on the internet. And Google dropped a major core algorithm update the same week. Here is what happened -- and why it matters.

Google replaced its 25-year-old search box with an AI-native interface powered by Gemini 3.5 Flash, and announced AI Search Agents that operate 24/7 without any user query.

 The search box is gone

The new interface dynamically expands to accept text, images, files, video, and open browser tabs simultaneously. Google describes it as anticipating intent rather than responding to keywords. This is not a redesign -- it is a different paradigm for how people interact with search.

Search Agents replace the act of searching

Starting summer 2026, Google AI Pro and Ultra subscribers can create custom AI agents that run continuously in the background, monitoring topics and surfacing relevant information without a human query. Your customer will not search for your business. An agent will decide whether to surface you.

The scale of AI Mode

      Google AI Mode has surpassed 1 billion monthly users

      Queries in AI Mode are 3x longer than traditional search

      Follow-up queries are up 40% month-over-month

      16% of searches are now multimodal -- images, files, and screenshots, not just text

      Planning queries (find a vendor, compare options) are growing at 80% the rate of overall usage

 Google also expanded agentic booking features for local services: the AI now surfaces live pricing and availability with a direct booking link. The human is increasingly just pressing confirm.

STORY #2
What did Microsoft change about how websites get indexed for AI?

Microsoft's Bing team published a framework declaring that the search index no longer exists to help humans find pages -- it now exists to help AI systems decide what to say. The new standard is called grounding.

 Grounding vs. ranking: what changed

The Bing blog post, co-authored by three senior Microsoft AI researchers, is the most technically explicit public statement any major platform has made about how AI indexing works. The key shift:

      Old goal: help humans decide what to read

      New goal: help AI systems decide what to say

      Old unit of value: a ranked web page

      New unit of value: a groundable fact -- a discrete claim with clear provenance that a model can responsibly cite

 What is abstention -- and why does it matter?

Unlike traditional search, which always returns results, the AI grounding index is designed to return nothing when content does not meet its evidentiary standard. Microsoft calls this abstention. If your content is vague, ambiguous, or unverifiable, the model declines to cite it entirely. This is a fundamentally different bar than ranking.

 

From the Bing team's official post (May 2026):

"The goal is no longer to rank for a keyword. It is to provide groundable information: discrete facts with clear provenance that a model can responsibly cite."

 STORY #3
What is the Google May 2026 Core Update and what does it signal?

 Google launched its May 2026 Core Update days after I/O -- the same week it formally declared AI Mode the future of search. The timing strongly signals the algorithm is being recalibrated for an AI-first index.

 What core updates do and do not tell us

Google does not publish detailed release notes for core updates. But the timing matters: this is the first major update since Google replaced its search interface and announced AI agents at I/O. Industry analysts read this as the algorithm beginning to reward content that serves AI's grounding requirements -- and deprioritize content optimized for the old keywords-and-links era.

The CTR data tells the real story

      Queries with an AI Overview: click-through rates have dropped 58% compared to pre-AI search

      Position one without a citation: losing approximately 18% of its historical clicks

      Position one WITH a citation inside an AI Overview: receiving 35% more clicks than a standard first-place ranking

 The divide between cited and not cited is becoming the new first page versus second page.

 STORY #4
Why does Perplexity launching a cross-platform browser matter?

Perplexity completed the rollout of Comet -- its AI-native browser -- to iOS, Android, Mac, and Windows this week, making it the first AI system to replace the browser itself across all major platforms simultaneously.

 What makes Comet different from a browser with AI features

Comet is not a browser with AI added on. Every page a user visits is being interpreted, summarized, compared, and contextualized in real time. Comet Assistant performs multi-step tasks autonomously -- booking flights, managing email, filling out forms -- without the user manually navigating anywhere.

The practical implication: when someone using Comet lands on your website, their browser is cross-referencing your content against other sources and deciding in real time whether you are the best answer to their underlying need. The user does not trigger this. It happens automatically.

The scale Perplexity has reached

      100+ million monthly active users across all Perplexity products (Sacra, April 2026)

      50% revenue growth in a single month -- March 2026

      Comet is now available for enterprise deployment via device management software

      Companies are issuing Comet to employees as a standard browser

 THE NUMBERS
What does the data show about the state of AI search in May 2026?

 AI search has crossed from early adoption into mainstream use. These figures reflect the current state, not projections:

 

48%

of all Google queries now show an AI Overview -- up from 34.5% in December 2025

60%

of Google searches now end without a click, answered directly by AI

1B+

monthly users on Google AI Mode; queries are 3x longer than traditional search

73%

of brands have zero AI mentions despite ranking on Google page one

25.6%

domain overlap between ChatGPT low-reasoning and high-reasoning citation sets

35%

more clicks for brands cited inside a Google AI Overview vs. a standard #1 ranking

38%

longer average sessions from AI-referred visitors vs. traditional search visitors

+527%

year-over-year growth in AI-sourced website sessions (2024 vs. 2025)

50%

Perplexity revenue growth in a single month (March 2026)

16%

of Google searches are now multimodal

 GOING DEEPER
Why do different AI queries cite completely different websites?

 New research shows that when ChatGPT uses high-reasoning mode versus low-reasoning mode, only 25.6% of cited domains overlap. The AI is effectively drawing from two different versions of the internet depending on how complex the question is.

 What the Semrush reasoning study found

Semrush's analyzed 20 buyer journeys across four industry verticals. When ChatGPT processes a simple query, it cites one set of sources. When it processes a complex query -- the kind involved in a real purchasing or research decision -- it draws from a near-entirely different pool of content.

      Low-reasoning mode: faster, broader, more surface-level citations

      High-reasoning mode: stricter evidentiary standard, narrower citation pool

      Overlap between the two: only 25.6% of cited domains appear in both

What this means for brand visibility

There are effectively two AI search landscapes running simultaneously. Most brands optimizing for AI visibility are optimizing for the casual query. But the queries that drive actual purchasing decisions, vendor comparisons, service evaluations, high-consideration research, trigger high-reasoning mode and pull from a fundamentally different set of sources.

The 25.6% overlap finding in plain terms:

Being cited in low-stakes AI queries does not mean you will be cited when a buyer is close to a decision. The most important queries have the highest citation standards, and they are pulling from a different internet than everything else.

 Why Microsoft's grounding framework explains this

The Bing team's May 2026 post provides the structural explanation: high-reasoning queries trigger a more aggressive abstention filter. The model applies a stricter evidentiary standard when the stakes of the answer are higher. This is not a ChatGPT quirk, it appears to be a structural feature of how advanced AI reasoning works across platforms.

 The search box is gone. The index has a new job. The browser is making decisions before you ask.

The week of May 21 through 26, 2026 will not be remembered as the week AI search arrived. It will be remembered as the week it became impossible to pretend otherwise.

check out how AI interviews the founder about the questions AI couldn't answer about them: https://airecommendable.monicaisystems.com/

Learn all about GEO and AEO on our learn hub: https://www.monicaisystems.com/learn

That's the Firehose for this week. If this landed, forward it to one person who needs to hear it. Till next time, stay visible!

— Monica

Keep Reading