I’ve got some exciting news to share! I’m thrilled to introduce the revamped Profound Index, your go-to leaderboard for AI Search. This update marks a new era in search, providing both clarity and authority in the rapidly evolving world of AI-driven solutions.
In this rebuild, we focused on enhancing the user experience and performance metrics. Whether you’re an AI enthusiast or a professional seeking the latest insights, the improved Profound Index is designed with you in mind. Its comprehensive data sets and intuitive interface make it an indispensable tool for anyone looking to stay ahead in the realm of AI Search.
When I think about how artificial intelligence is revolutionizing advertising, a common belief is that AI is killing advertising. But, in reality, AI is not the end of advertising; it’s merely transforming it into new dimensions. With AI seamlessly integrating into search, assistants, productivity tools, and beyond, it’s only natural for advertising to follow suit.
I’ve noticed that while the density of ads may shift in AI-led experiences, the opportunities for advertising are actually broadening. There are new surfaces emerging continuously, and they all offer exciting chances for advancers and advertisers alike.
To me, the divide between paid and organic isn’t as clear-cut anymore. The same AI systems powering search experiences are also driving ad campaigns and influencing brand visibility across Google’s expansive ecosystem.
This calls for a change in how we brands perceive visibility. Paid and organic aren’t just isolated competitors vying for clicks; instead, they’ve become alternative strategies influencing the same AI systems. As a result, the signals that shape organic visibility may also impact paid performance.
The traditional search engine results page (SERP) we once knew, consisting of 10 blue links, a handful of ad slots, and a side panel, no longer holds the same dominance. Back then, dedicated teams managed paid and organic strategies separately, each with its own set of tools and quarterly goals.
Things changed for me when Dynamic Search Ads (DSA) appeared, using my website’s content to cleverly create ad titles and determine bids, merging the lines between our organic strategies and paid efforts.
Stepping into the modern age, Performance Max (PMax) campaigns took the very logic of DSAs and applied it across every Google surface—importantly altering how ads are placed from Search and YouTube to Maps and more.
Of course, it isn’t without its nuances. If Google’s Gemini doesn’t have a thorough understanding of our brand, the system has to fill the gaps with assumptions, which may not align with our intended brand narrative. It’s crucial to train these AI systems deliberately, or we risk losing control.
Strategically, I’ve come to realize that paid campaigns help me discover which target audience-intent-profit combinations convert best. I can then build my organic content around these successful elements, creating a feedback loop where each strategy amplifies the other.
I recently discovered some exciting news from Microsoft Ads that could be a game-changer for advertisers like myself. They’ve expanded their LinkedIn targeting capabilities to include job seniority filters. This allows me to target audiences with more precision in both Search and Audience campaigns.
This new feature means that I can now target users based on their job seniority, a wonderful addition for those of us focusing on B2B marketing. Thanks to LinkedIn data, I can reach audiences at various levels of seniority.
What’s the scoop? According to Navah Hopkins, Microsoft Advertising has added job seniority targeting to its LinkedIn Profile targeting, allowing me to utilize it within Search and Audience campaigns.
This update provides me the ability to choose from 10 different seniority levels, ranging from CXO to Volunteer. This flexibility is available at both the campaign and ad group levels, making it easier to segment my audiences effectively.
Why is this vital for us? In the world of B2B marketing, it’s often challenging to separate decision-makers from operational staff in search campaigns. With this new job seniority targeting, I can better align my messaging and bidding strategies with the right audience segments, ultimately improving my campaign performance.
Understanding who is interacting with my ads is crucial, especially in organizations with long sales cycles or multiple stakeholders. It’s not just about conversions; it’s about knowing who is behind them.
A closer look: Unlike other platforms, Microsoft’s integration with LinkedIn provides a unique perspective of professional identity that allows me to better understand user interactions.
Not only can I apply these filters directly within my campaign settings, but I can also utilize them in observation mode to gather insights without limiting my reach.
How do I benefit?
Customize messaging by seniority: I can create targeted ad groups for different audience levels, like executives or individual contributors, tailoring my messaging to each group’s expectations.
An executive-focused strategy might highlight business growth, while campaigns targeting practitioners could focus on efficiency gains.
Analyze conversions by seniority: Observation mode helps me assess conversion performance across different seniority levels, answering questions crucial to my strategy:
Where are my conversions coming from? Are they decision-makers or influencers? Is my budget effectively spent? Which seniority levels bring in high-quality leads?
Enhance audience testing: This feature offers an extra layer of reporting, helping me make informed optimization and expansion decisions. If I’m importing from other platforms, this insight is invaluable for discovering performance patterns unique to Microsoft Ads.
Availability: This powerful tool is now accessible in select markets across the Americas, EMEA, and APAC regions, including countries like the United States, Canada, Brazil, and more.
Americas: Argentina, Brazil, Canada, Chile, Colombia, and others.
EMEA: Egypt, Nigeria, Saudi Arabia, and South Africa.
APAC: Australia, India, Japan, among others.
The takeaway: Microsoft Ads continues to leverage its LinkedIn integration as a standout feature in B2B advertising. By aligning search intent with professional profiles, I gain deeper insights into not just what my audiences search for, but who the searchers are.
I recently came across some notable updates from Google Ads that could impact a number of advertisers like me. From July 15, Google is making a big shift in how it charges for Demand Gen campaigns on Discover, specifically those aimed at view-through conversions (VTC). Instead of the traditional cost-per-click (CPC) model, we’ll be billed on a cost-per-thousand impressions (CPM) basis.
What happened. Google Ads informed me, along with other advertisers, that this shift will directly affect campaigns using VTC optimization. If you’re like me and use this optimization, be prepared for the billing change. This only impacts campaigns with VTC enabled, so if you’re not using it, you’re in the clear.
Luckily, no action is required on my part for this transition to take place; it’s automatic.
Why we care. For those of us focused on efficiency in Demand Gen campaigns, this switch could mean we’ll need to closely monitor changes in spend, impressions, and reporting metrics since the basis for billing is changing from clicks to impressions.
This shift in billing might prompt some of us, who primarily look for click-driven performance, to reassess if VTC optimization aligns with our goals.
Why Google is making the change. According to Google, aligning billing with campaign objectives is key. View-through conversions rely heavily on ad impressions. Thus, billing on a CPM basis could more accurately reflect the actual value generated from these campaigns.
Moreover, Google believes this shift will enhance the system’s ability to optimize for VTC goals more effectively.
Opt-out option. If the new billing structure doesn’t suit you, there’s an opt-out. Disabling VTC optimization in campaign settings will prevent this change from affecting your campaigns.
The bottom line. With Google tying payments more closely to the behaviors Demand Gen campaigns are crafted to optimize, those of us leveraging VTC will now focus on impressions rather than clicks for billing and optimization on Discover.
First spotted. This update first came to my attention through Adsquire founder, Anthony Higman, who shared details on X.
Recovering from a manual action is no quick fix; it can take months of rigorous cleanup and multiple reviews. I’ve learned that regular compliance audits are key to avoiding a crisis altogether.
Google penalties—or manual spam actions—are those unpredictable disruptions that can shake up a thriving online business overnight.
For businesses like mine that rely heavily on organic traffic, the impact is quite severe. It goes beyond just losing rankings; revenue takes a hit, customer acquisition costs spike, expansion plans are halted, and the effects linger long after the policy issues have been addressed.
With Google’s consistent 90% market share, it remains my main source of traffic, much like it is for many publishers, e-commerce platforms, and lead generation companies.
Unfortunately, direct traffic seldom makes up for significant visibility losses, and Bing isn’t enough to fill the gap. This means a manual spam action is not just an SEO risk but a grave operational concern.
Manual Actions Aren’t Algorithm Updates
It’s essential for me to clarify that manual spam actions and algorithmic updates are two different beasts. Manual penalties result from specific violations identified against Google Search Essentials and demand entirely different responses.
Manual actions involve considerable internal review at Google. When violations are suspected and verified, these actions are taken, because proven policy breaches aren’t taken lightly by Google.
The real issue lies in recognizing accumulated policy violations over time, something I’ve seen many businesses fail to address adequately.
How Penalties Develop
The journey to a manual penalty often begins in non-obvious ways, with compliance erosion happening gradually.
An e-commerce company might start with aggressive link-building strategies that accumulate unchecked spam links over the years.
A publisher engages in commercial partnerships involving sponsored content, integrating these into their main site structure.
A SaaS business expands into new markets with low-quality location pages.
Lead generation companies scale supplemental SEO content without thorough editorial oversight, simply adhering to industry standards.
Though these tactics might initially boost visibility and revenue, they often fall out of line with Google’s quality standards over time.
Why Historical Violations Still Matter
Manual spam actions are disruptive partly because old policy violations can persist without being flagged for years. Google doesn’t forget historical footprints in its search system, meaning unresolved past SEO practices can become today’s liabilities.
Practices like paid placements, commercial guest posting, or directory spam from years ago can remain risks until they’re addressed, creating vulnerabilities that must not be ignored.
Reputation Abuse and Publisher Liability
When a trustworthy brand allows unsupervised content from third parties, the site’s credibility might suffer. Once a manual spam action hits, the entire site can lose visibility—even the genuinely valuable sections suffer.
Recovery from such penalties is not simple or cheap. It often demands structural changes and more stringent editorial and technical controls, as I can attest from my own experiences.
The Risks of Scaled Content
Google is now more vigilant about large-scale publishing systems that lack originality and value. I’m aware of how easily businesses, unintentionally, slide into creating repetitive, low-value content.
Most businesses don’t cross these lines deliberately. However, without ongoing reviews and updates, significant issues can fester under the radar.
Compliance Requires Ongoing Oversight
For me, regular compliance reviews are non-negotiable. It takes external expertise to assess true compliance comprehensively. Even powerful internal SEO teams can miss potential exposure points if left unchecked.
I’ve found that organizations integrating compliance into governance see considerable advantages. Regular audits and assessments can preempt violations and protect critical search traffic, especially during pivotal business moments.
In essence, prevention through regular audits is a more efficient and less painful approach than dealing with recovery after a penalty.
Today, as I explore updates from Microsoft, I’m excited to share that Bing Webmaster Tools is rolling out a preview of its new AI performance report enhancements. These include insights like Intents, Topics, Citation Share, and Compare, and they’re being introduced globally. After witnessing Microsoft’s demo in April, it’s thrilling to know these features are finally accessible to us.
Reflecting on their past roll-outs, Bing officially launched its AI performance report earlier in February, a bold move ahead of Google’s similar feature which wasn’t available in Search Console until June. Google’s delayed release felt quite rushed to many of us.
New Insights: Krishna Madhavan from Microsoft describes these updates as built to give publishers a clearer understanding of why their content surfaces, the broader subject areas they’re gaining visibility in, and how their presence compares with other sources over time.
Intent: The Intents feature now classifies grounding queries into categories such as Informational, Commercial, Navigational, and more. This provides deeper insight into the intent behind user queries, moving beyond just triggering citations to understanding broader query contexts.
An example given was an e-commerce publisher finding visibility in comparison-focused experiences or an educational publisher learning their content is popular in research-oriented interactions. These insights can guide us in refining content structure and depth.
Topics: Topics group related queries into thematic clusters, offering us a more organized way to understand visibility, similar to modern AI’s reasoning across themes rather than isolated keywords.
For instance, queries like “solar panels” and “solar energy efficiency” can all be part of a larger topic cluster such as Solar Energy. This thematic organization helps us align our content with how AI systems engage with our content.
During this preview phase, some labels might remain broad, especially for niche domains, but meaningful patterns are already emerging.
Citations: Citation Share now displays the percentage of citation visibility your site enjoys compared to others. It’s a directional metric to help us understand our evolving visibility over time, without ranking or quality scores.
Compare: We can now compare citation changes over time. This feature overlays previous data onto current reports, helping us observe citation activity, which can be influenced by various factors like AI model updates, user demand shifts, and more.
Why This Matters: Although we’re still awaiting click and click-through rate data, these growing AI performance insights are invaluable. I’m hopeful that such detailed data will become available to us from Microsoft or even Google one day.
I realized that many web pages effectively address initial search queries, but often fall short when it comes to guiding the user toward their final decision. This is where the concept of next-question intent becomes crucial. It’s a tool that not only aids users but also aligns with AI systems for enhanced content utility and visibility.
In the world of GEO, much of the discussion revolves around how AI systems discover, extract, and suggest content. While these aspects are essential, I’ve learned that what truly determines visibility is the substantive content these systems find once they’ve reached my pages.
Next-question intent isn’t just about answering the initial query. It’s about whether my page provides enough depth for the user to take their next step, be it selecting a product or making a decision.
Often, a user’s first search is just a starting point. Key decisions hinge on follow-up questions and considerations that must be addressed.
By crafting content that anticipates these subsequent inquiries, I equip AI systems with rich materials to synthesize, compare, and recommend.
From Results to Narratives: Traditional Search vs. AI Search
Traditional search was once about offering a suite of links for users to peruse and decipher. Now, AI search focuses on delivering synthesized responses, pulling information from multiple sources.
This shift emphasizes the need for my content to provide comprehensive information that can help build AI-generated answers. Next-question intent is vital here.
While search intent asks what the user wants to do, next-question intent goes further. It asks what the user will need to know next to trust, compare, or decide.
In this AI-driven environment, content must support a complete answer pathway, far beyond the initial query.
The initial search often serves as just the beginning, an entry point. True decision-making occurs through follow-ups and specific concerns that arise thereafter.
Take the query “best CRM software for small business” as an example. It opens the door, but the true selection journey starts with follow-up questions.
Which platform is easiest for a two-person team?
Which integrates best with QuickBooks?
Which one works for a business without a formal sales department?
Which one is best for a local service company rather than a software startup?
Which one won’t frustrate owners or interns with tech complexity?
These aren’t ancillary. They define the decision-making path.
Otherwise well-structured content may falter if it fails to engage at this level, leaving AI systems with less context to assemble an answer, thereby reducing visibility.
Next-Question Intent is Not Just a Writing Exercise
As I’ve delved into content creation, it’s clear that next-question intent goes beyond simply writing better content—it ensures my pages support the next steps in a user’s decision-making process.
Practically speaking, it means crafting answer-ready content that addresses initial user needs, foresees additional decision layers, and provides concrete, verifiable information.
Visibility in AI search isn’t just about where I rank. It’s about citations and whether my brand becomes a trusted source in context-rich settings.
To achieve this, my content must offer enough substance for systems to understand what my brand does, whom it serves, when it’s useful, why it’s trustworthy, and how it fares against alternatives.
Where Good Content Goes Thin
While I often find that brands have content that’s accurate and keyword-optimized, it still might not suffice in the AI search environment.
AI systems require clarity and context to determine what I offer, who benefits from it, when it’s applicable, and why claims are valid.
This depth is where many pages fall short.
A service claim like “customized marketing strategies” begs the question: customized how?
A product claim like “safe for families” prompts: safe for which family members?
A software claim like “built for small businesses” asks: which type of business?
General claims offer little for people and even less for AI systems to utilize. Specific, structured, evidence-backed content serves a far better purpose.
I’ve witnessed AI tools become indispensable in automating complex processes that traditionally demanded a lot of manual effort. However, I’ve also seen them used without any real benefit just because they are available.
That’s why I prefer focusing on AI applications that save time and address genuine challenges.
Recently, I was tasked with aligning the SEO architecture for over a dozen websites across three separate businesses, eight regional domains, and numerous languages, including three English dialects, Italian, Japanese, Spanish, Thai, French, and Korean.
Mapping thousands of URLs to create seamless hreflang XML sitemaps traditionally required specialized software or extensive spreadsheet work. Instead, I used Google Gemini to develop a custom Python script to handle the heavy lifting.
Here’s how an initial prompt evolved into a fully customized automation tool and what it taught me about utilizing AI for technical SEO.
Where AI Delivers the Most Value
I leverage AI primarily for practical, time-saving tasks, including:
Generating regex patterns when I need quick solutions without researching syntax from scratch.
Creating complex spreadsheet formulas for reporting workflows that depend on manual data exports.
Speeding up research and planning for projects requiring competitive analysis across business lines.
Building custom automation tools for recurring SEO and data-processing tasks.
The hreflang project I discuss here fits perfectly into the last category.
Mapping hreflang at Scale
The challenge was straightforward: accurately map thousands of URLs across multiple multilingual websites into cohesive hreflang XML sitemaps.
I chose not to tackle this manually. Instead, Google Gemini helped me build a custom Python solution.
Here’s a walkthrough of how the process unfolded.
Phase 1: Asking for an Approach, Not Just a Script
One common pitfall of using generative AI for coding is asking it to sprint before understanding the course. Typing, “Write a Python script to create an hreflang sitemap,” will yield generic code prone to breaking with real-world data.
Instead, I started by asking for an approach. I detailed the scenario: multiple regional domains, organic growth over several years leading to mismatched URL slugs, translated subfolders, and appended revision years.
Gemini suggested a multi-step, data-driven approach:
Crawl the websites to collect live URLs and their metadata.
Use Python in Google Colab to process the raw data.
Run an exact match cluster to group identical slugs.
Use an advanced semantic AI model (like SentenceTransformers) to fuzzy match translated pages based on their titles and normalized URLs.
Phase 2: Crawling and Data Collection
Following the recommended strategy, I used a crawler to spider all regional websites to generate a unified CSV file with live URLs, status codes, title tags, and H1s. Screaming Frog proved ideal for this task.
The quality of AI output relates directly to the quality of your crawl data, so make sure it’s robust.
An AI script can miss an obvious “exact match” if a target URL is a 404 or a 301 redirect. Before feeding data into the script, filter your CSV to include only indexable content.
Google Colab offers a free, cloud-based Jupyter notebook environment for coding, bypassing local installations or environment variable issues. I used Google Drive to access it. The free version sufficed for this project.
After uploading the CSV to Colab, Gemini provided an initial Python script that utilized a domain-mapping routine to assign language codes, clean the URLs, and generate an XML tree. The initial results required refinement.
Phase 4: The Iteration (Where the Real Work Happens)
If you expect AI to produce a flawless script on the first try, you’ll be disappointed. Like an intern, AI requires oversight. The true value lies in iteration.
After running the initial script, several unmatched URLs left orphaned pages rather than grouping them with international counterparts. Here’s how I iteratively guided AI through the complexities of human-managed websites.
The Directory Flattening Problem
The U.S. site had recently reorganized its blog into topical folders, unlike the Mexican and Italian sites. I presented these mismatches to Gemini, leading to a script adjustment that flattened directories, allowing slugs to align.
The Aggressive Semantic Trap
Concept traps we implemented were initially strict. A UK article about manufacturing wouldn’t match its Italian counterpart due to a slightly different title. By loosening these traps for general industries and enforcing them for critical terms, the AI became adept at delivering better matches.
The Translated Slug Epiphany
The pivotal insight arrived when examining Mexican blog orphans. A Spanish URL /detras-de-escenas-historias... matched the English /behind-the-scenes-stories..., which I pointed out to Gemini. As a result, Gemini updated the script to create a “Combined Semantic Signature,” dynamically translating slugs and efficiently bridging language gaps.
This project reinforced a simple truth: AI excels as a collaborator rather than a shortcut.
Be the strategist, let AI be the coder: Rather than demanding a finished product, discuss architecture and logic first, treating AI as a junior developer needing guidance.
Provide concrete examples: Don’t simply state, “It’s broken.” Give specific failed URL examples or mismatches to help AI refine its logic.
Embrace the iterative loop: Run the code, identify issues, and iterate. Each iteration enhances the tool’s intelligence.
Leverage Google Colab: You don’t need to be a Python guru to apply Python in SEO. Colab bridges the gap, providing access to complex data science libraries in your browser.
In the end, I had a fully customized Python script capable of processing a massive CSV to generate a cross-referenced hreflang XML sitemap in minutes.
Though AI isn’t replacing technical SEOs, those who collaborate with AI to build scalable tools will have a significant edge.
I recently delved into a fascinating study on Google Discover headline formats, looking at a staggering 3.4 million articles. The results were eye-opening and showed that a simple headline rewrite often doesn’t yield the expected lift.
You might have come across these bold statements before:
Quote-led headlines outperform plain declarative ones by nearly 29%.
Question headlines underperform both, sometimes by 24%.
Format drives the result: Rewrite a statement as a quote, or add that magic word, and you should expect a real lift.
To put these claims to the test, I examined 1,674,518 English articles and 1,690,295 French articles from the 1492.vision Discover corpus. That’s quite a hefty sample size!
What I found was a deeper flaw than just numbers. It turns out that all three claims treat headline format as a leverage point for visibility. However, the data clearly shows that the impact of a headline’s format mainly reflects the publisher’s audience and the specific Discover surface used.
One striking analysis was Simpson’s paradox. An anomaly that, once noticed, appeared across the entire dataset.
Here’s what we’re really measuring:
Rather than clicks from Discover, our metric is hits per article: how often an article appears across the 1492.vision fleet. This serves as a proxy for visibility.
The dataset was limited to editorial articles, excluding platforms like YouTube because they have different headline norms. We’ll dive back into these at the end, as they bring more clarity than anything else.
Why is volume important? The crux of the argument depends on slicing this vast dataset by publisher, Discover surface, topic, and language while still keeping enough data in each segment for valid insights. This is where the real difference between numbers and insights, and between a genuine format effect and a statistical illusion, lies.
Here’s a sneak peek: when you pool all publishers together, a clear gradient appears with quote-led headlines leading the pack and statements trailing.
The frequently cited +29% is actually a conservative estimate for editorial pieces: quote-led headlines achieve a +37% lift in English and +48% in French. Even questions don’t lag behind as much as expected since they outperform statements to some extent (+7% EN, +16% FR).
Though claim 1 appears understated and claim 2 misguided at the aggregate level, these are the observations on which most headline advice leans. Let’s delve further to understand what the data is really revealing.
Let’s shift to the hidden aspects, starting with publishers. The raw comparison isn’t effectively between quotes and statements. It’s more about one set of publishers versus another because the publishers employing quotes often differ from those who don’t.
Some media, like celebrity-focused outlets, regional newspapers, and sites attuned to trending topics, gravitate towards quotes, and naturally earn more Discover hits compared to entities that prefer factual presentations.
This is a prime example of Simpson’s paradox: a strong trend at the aggregate level that fades or reverses when segmented into groups.
To focus on the format itself, publishers must each be their own baseline: comparing quotes with statements within the same publishing entities while controlling for audience and topic diversity.
So, the question is, how does each format fare on its own? Let me walk you through the rest of this journey as we unpack these layers.
While working with a major B2B SaaS account, I learned a valuable lesson about fundamentals through a €30,000 underspend. I initially tightened a target CPA to boost efficiency but neglected to monitor the consequences closely. This oversight resulted in the account falling short of its monthly budget target.
Underspending isn’t merely a media issue—it can affect future budget allocations. The unspent funds had to be returned to finance, making it challenging for the marketing team to justify similar investment levels in upcoming cycles.
The toughest part wasn’t the monetary mistake, but owning up to it. I had to personally explain the situation to the client, taking full responsibility without making excuses.
Even though the client was understanding, trust was undoubtedly shaken. To rebuild confidence, I introduced weekly budget pacing updates to demonstrate transparency and assure them that such an issue would not recur.
Throughout this experience, I realized the enduring importance of fundamentals like budget pacing, account monitoring, and conversion tracking. These basics are crucial, regardless of how advanced our advertising platforms become.
Reflecting on what I would do differently, I underestimated how impactful a target CPA change could be. Now, I consider any spend-related adjustment as a significant account change necessitating careful observation.
While I support using AI-powered tools, I’ve learned to be cautious about adopting every new feature indiscriminately. Balancing experimentation with human insight and strategic oversight is essential to me.
One of the industry’s recurring blind spots is conversion tracking, often flawed due to poor implementation. Accurate data is essential, as it drives optimization decisions and performance.
Building strong client relationships is also pivotal. When mistakes happen, open communication and honesty can be just as critical as producing solid performance results.
Ultimately, mistakes in PPC are unavoidable, but how we handle them defines our success. This experience reminded me that mastering the basics and maintaining trust lay the groundwork for enduring success.
