SEO Consulting 
SEO audit 
Uncover market and competitor insights 
Increase Sales Qualified Leads (SQL) 
For CMOs 
For demand generation leaders 
For technology startups 
Enterprise SEO 
Web Search Optimization 
Local SEO 
PPC Advertising 
CRM Consulting 
Website Design & Branding 
Website Development 
Web Support 
Web pricing 
Content Writing 
Editorial Calendar 
All resources 
Case studies 
Marketing fundamentals 
Worksheets 
Essential SEO checklists 
Case studies 
The team 
Careers 
The search landscape in 2025 is changing without a doubt.
How worried should you actually be about your hard-earned organic rankings?
Should you be investing in content creation? Accessibility compliance? Core web vital improvements?
Is there a way to stay on top of AI overviews? How about being recommended by answer engines like Gemini, Claude, ChatGPT and Perplexity?
How fast are AI engines being adopted for search, and what’s the difference between AI engine-referred users and organic users?
To answer these and many more questions, we’ve analyzed the state of AI search of 20 SaaS companies we work with. This will give you some pretty cool insights as to how to approach SEO, or should I say GEO (generative engine optimization).
🧠 This post will be updated once at the end of every quarter to ensure the accuracy of the data and advice.
The basic logic of our approach
- We’ve selected x20 SaaS companies from various industries, such as cybersecurity, document management, whiteboard software, identity verification, and learning management systems.
- The time period we’ve analyzed is all of 2024 and will continue to add to this study 2025 numbers quarterly as they roll in.
- Why we did this: to ensure we can adjust SEO tactic direction in an AI-first world.
- Additional notes:
- This study focuses mainly on the impact of AI search/answer engines. However, we also did research on Google AI overviews (for some of our data), so do not confuse the two concepts.
- We’ve selected a company portfolio with the lowest amount of analytics variation and data changes to ensure data integrity (e.g. companies that do not change their event tracking, have not gone through massive data migrations, have not done major rebrands, did not engage in SEO bad practices and so on).
The impact of generative search: Questions and answers
| Metrics | Average |
|---|---|
| How fast are AI tools being adopted? | AI Traffic is growing MoM at an average of 45% although the absolute number of sessions is still significantly smaller than Google Search. Organic Traffic is growing MoM at an average of 1.03% although the absolute number of sessions is significantly higher than AI traffic. The data: + AI sessions make 0.3% of total traffic. + Organic sessions make 38.11% of total traffic. |
| What % of sessions are from AI search vs organic search? | Organic contributes a considerable amount more traffic to each website, with the average being 38.11%. AI traffic accounts for a small fraction of the traffic organic brings in despite the rapid growth. The data: + 0.8% AI sessions (from organic traffic) + 0.3% AI sessions (from total traffic) + 38.11% Organic sessions (from total traffic) |
| What % of sessions are converted from AI search? | AI conversation rate is 0.54% less than the organic conversion rate. Though this only represents the aggregate, specific AI tools have considerably higher conversion rates. For example, a whiteboard and process modeling collaboration company’s AI conversions are at 3,783, while organic conversions are at 398,524. AI accounts for 0.9% of organic conversions. The data: + AI conversion rate: 3.96% + Organic conversion rate: 4.31% *Note that you need to track other ai tools as well to get a complete picture of your conversion rates: chatgpt.com, Perplexity, Edgepilot, Edgeservices, copilot.microsoft.com, openai.com, gemini.google.com, nimble.ai, iask.ai, claude.ai, Aitastic.app, bnngpt.com, writesonic.com, Copy.ai |
| Has organic CTR decreased with the increase of AIO and AI Search Engines? | Yes but ONLY for keywords where the AIO is present. When an AIO is introduced CTR for that particular term decreases. This insight is in formation, we will update you shortly. |
| Session duration (engagement) | AI referral traffic has 17.11% longer session durations, with an average of 58 seconds more time spent per session. The data: + Average AI Session -00:06:26 + Average Organic Session -00:05:28 + Average AI Session vs. Average Organic Session (difference) - 58 seconds or 17.11% AI sessions might be longer but it does not necessarily equate to conversion. As an analog example, roughly 86% of those AI-influenced sessions are landing on top-of-funnel (TOF) content, primarily blog posts. This suggests that users engaging with AI are largely seeking informational content, such as understanding project management techniques, rather than being further down the funnel and ready to convert. They're primarily in the 'information gathering' stage. |
| Bounce rate (content resonance) | The organic traffic bounce rate is 5.71% less than AI bounce rates. This implies that organic users find on-page content more resonating than AI citation-follow-through users. 💡 This is a different finding than our reported numbers from October 24, when AI traffic displayed longer engagement behaviour. The data: + Organic bounce rate: 37.72% + AI bounce rate: 43.43% |
| Sessions | ChatGPT remains the AI referral source that is attracting the most traffic with Perplexity closely following behind. (These numbers do not include Google’s AI overview search results) The data: + ChatGPT: 882.47 + Gemini: 491.42 + CoPilot: 211.84 + Preplexity: 759.7 Assumption Apart from ChatGPT being the first player in AI, according to tomsguide.com, ChatGPT won for having natural language, problem-solving, knowledge retrieval, and conversion which likely contributes to its higher user engagement and conversion rates. |
| Conversions | ChatGPT is ahead in terms of total traffic and qualified traffic. Even though it attracts the most traffic it also has the largest proportion of users that are willing to convert. The data: + ChatGPT: 83.47 + Gemini: 55.42 + CoPilot: 9.16 + Preplexity: 26.29 |
| Conversion rate | ChatGPT and Gemini attract significantly more qualified traffic than Copilot and Perplexity despite having very similar bounce rates The data: + ChatGPT: 12.31% + Gemini: 7.98% + CoPilot: 4.64% + Preplexity: 4% |
| AIOs vs SERP features | Over 90% of AIO results, also trigger a traditional SERP feature. |
| AIO Difficulty and Volume | On average B2B AIO keywords have 38% keyword difficulty and 354 search volume (which makes sense, since most of them are informational keywords) The data: + 97,000 keywords analyzed across multiple companies |
| How many images should a URL have to rank in AI results? | 65% of ranking pages contain at least 1 image. The average number of images per ranking page is 7 implying that AI prefers richer content. The data: + 85.71% of ranking listicles have an average of 11 images. + 71.43% of ranking landing pages have an average of 5 images. + 65.69% of ranking blog posts have an average of 4 images. |
| What type of tone do AI tools prefer? | The most common tone used (as defined by Gemini) was “precise” at 42%, and ‘’formal’’ at 25.11% of all AI overview+ Gemini pages. Definition of Tone Precise - Simple, direct, clear, precise, informative, instructional, and factual. Formal - Professional and educational. The data: + 44.87% of ranking blog posts have a precise tone. + 24.05% of ranking blog posts have a formal tone. |
| How many times does the core keyword need to be mentioned to be ranked by AI results? | Pages that rank in AI search results mention the core keyword and its variations an average of 7 times (within the headings and body text). The data: + On average, ranking blog posts have at least 8 keyword mentions. + On average, ranking landing pages have at least 7 keyword mentions. + On average, ranking listicles have at least 9 keyword mentions. |
| Does the mention of the keyword on the URL impact AI appearance? | No, mentioning the keyword in the URL does not impact AI appearance since 90% of the links do not include the keyword in the URL. The data: + 63% of ranking blog posts do not mention the keyword in the URL. + 87% of ranking landing pages do not mention the keyword in the URL. + 75% of ranking listicles do not mention the keyword in the URL. |
| Does having a data point on the page impact AI appearance? | The presence of data points does not significantly correlate with ranking in AI search results. The data: + 48% of ranking blog posts have at least 1 data point. + 33% of ranking landing pages have at least 1 data point. + 35% of ranking listicles have at least 1 data point. Best Practice: Our analysis suggests that data points may not significantly impact AI search results. However, the observed difference is sufficient to recommend including at least one data point as a best practice. |
| What types of content are more likely to get into AI Overview and other AI search results based on keyword intent? | Blog posts with informational keyword intent are more likely to appear in AI overviews and other AI search results, representing 86% of the total ranking pages. The data: + For informational keyword intent, ranking blog posts represent 89%. + For commercial keyword intent, ranking blog posts represent 61%. + For navigational keyword intent, ranking blog posts represent 40%, sharing the same percentage with ranking landing pages. + For commercial/informational keyword intent, ranking blog posts represent 81%. |
| Which types of content perform best? | Blog posts are the most effective content type on AI overview and other AI search results. The data: + 85.6% of the ranking pages are blog posts. + 10.7% of the ranking pages are landing pages. + 3.7% of the ranking pages are listicles. Best practice: Despite Listicles performing lower in our analysis, this type of content can enhance SEO strategy by providing an effective and clear content format. |
| Is there a relationship between the number of backlinks and the ability to get into AI search results? | Yes, 90% of the analyzed ranking pages have backlinks. However, only 26% of the ranking pages have backlinks with a high domain authority (SEMrush) of 30 and above. The data suggest that having high domain backlinks is not a prerequisite to appear in AI overview and other AI search results. The data: + 75% of ranking blog posts have no high-authority backlinks. + 71% of ranking landing pages have no high-authority backlinks. + 78% of ranking listicles have no high-authority backlinks. |
SEO vs GEO Trends (with examples)
Below are some 2024 traffic and conversion trends for a few B2B websites. Although we are seeing a considerable amount of growth in AI traffic, it is clear that organic still outclasses AI in terms of sheer volume.
A whiteboard collaboration software
Chart of AI vs Organic traffic trend
Chart of AI vs Organic conversion trend
Hotel management software
Chart of AI vs Organic traffic trend
Chart of AI vs Organic conversion trend
An identity management platform
Chart of AI vs Organic session trend
Chart of AI vs Organic conversion trend
Business process modeling
Chart of AI vs Organic session trend
Chart of AI vs Organic conversion trend
Forming trends
Will AI answer engines overtake organic traffic and conversion numbers in 2025?
For SaaS, highly unlikely, AI referral traffic and AI overview-influenced traffic.
The data
- Although AI conversion is growing MoM at an average of 26%, the absolute number of conversions is significantly lower than Organic.
- Although AI traffic is growing MoM at an average of 45%, the absolute number of sessions is significantly lower than Google search.
- Although Organic conversion is growing MoM at an average of 4%, the absolute number of conversions is significantly higher than that of AI.
- Although Organic traffic is growing MoM at an average of 1%, the absolute number of sessions is significantly higher than AI traffic.
Hypothesis
Organic may be growing at a more relaxed pace (1% traffic, 4% conversions MoM), but it’s the absolute number of Organic sessions and conversions that seals the deal. Google search continues to dominate in sessions, and organic channels are the undisputed conversion champions. Our bet? AI will be a valuable supporting player, augmenting the customer journey, but organic will remain the MVP for 2025.
Are AI overviews and AI answer engines a separate category from organic search?
No. Although they are considered referral sources, we should still consider them organic channels since users still start with the intent to solve a problem.
How does ranking in AIOs impact CTR?
- Ranking on position #1: The CTR is 8%
- When an AI overview appears for a search phrase in which our website ranks in the top 3, the CTR goes down about 55% for that top 3 ranking page.
- On the contrary, If your website is not ranking in the top 10, you have very limited CTR, the AIO appearing as a source will get you additional traffic. The majority of the companies will benefit from AIOs:
- Example: out of a keyword base of 1,000 keywords, a company may be positioning for 150 keywords in the top 10. But with AI overviews, they will now start getting more traffic for the remainder 850 keywords which they were not able to position for organically.
- For keywords that do position for AI overviews (in which we have low rankings), it introduced new traffic, not augmented existing traffic.
- To trigger AI overviews: You need to use long tail keywords with a minimum of 4 words (long-tail keywords).
Is SEO transitioning to GEO?
Yes, although most SEO principles will apply to Generative Engine Optimization, we will likely transition to calling it GEO or the alike very soon.
Which AI tool is used most frequently to query for SaaS answers?
ChatGPT brings the most sessions and has the highest conversion rate by a significant margin.
Best practices for implementation to trigger AIOs and GEO results
Blog posts
- Precise and professional tone that is a human authority: Write with expertise while maintaining a conversational style. Include first-person perspectives and practical insights from direct experience, using phrases like “In my experience” or “Based on our implementation” to establish authentic credibility.
- Citations: Include references from reputable sources like academic papers, industry reports, and research studies to strengthen content credibility. Use recent citations when possible to show current knowledge of the subject matter.
- Quotes: Integrate relevant quotes from industry experts and thought leaders to add authority. Ensure quotes are properly attributed and support key points rather than standing alone without context.
- Key takeaways + definition e.g. https://nordlayer.com/learn/network-security/network-security-management/ Start with a clear, accessible definition followed by practical takeaways that summarize main points. This structure helps AI tools identify and extract essential information effectively.
- Data points: Use recent statistics and metrics to provide concrete evidence, ensuring all data is properly sourced and presented with meaningful context rather than as isolated numbers.
- Contextual internal links: Place relevant internal links naturally within content to create depth and demonstrate comprehensive topic coverage. Links should connect to related articles that expand on mentioned concepts.
Listicles
- Write about the top software solutions in your space
Links
- Buy backlinks in listicles that are popular
- E.g. https://expertinsights.com/insights/the-top-application-security-posture-management-aspm-tools/
- The average spend per backlink across multiple SaaS industries is: $495 per link
Trends graph observations (TBD)
- Observation: there is an exponential increase in AI driven conversions and traffic, however, it is a very small subset (0.31%) of the total organic traffic and conversion numbers.
- Future: in Q2 of 2025 it will be more clear (since we will have more solid baseline) as to how much of search traffic will be augmented by AI traffic.
What is the source-medium if a URL is clicked on through a result that is ranking in an ai overview?
- Google has not disclosed this, most likely it will be a custom source-medium that will be labeled as “AI Overview”.
Are AI overviews a novelty/fad?
AIOs are here to stay, just like SERP features, their trigger criteria and layout will likely evolve over time and will purely be driven by relevancy of search intent vs content served. But, we theroize that just like like people have learned to ignore featured snippets and paid ads to seek relevant content, the aforementioned CTR rate that is 7% to 9% may slightly decrease as people adopt searching via answer engines, or, scroll to organic results.
What you should avoid
Scaled content abuse: Don’t flood your website with AI-generated content just to rank in search results. Google will catch and penalize sites that mass-produce low-quality pages that aren’t designed to help real users. Practices that will get you penalized:
- Using AI tools to generate many similar pages without adding real value
- Scraping content from other sites and rewriting it automatically
- Combining pieces from different pages without adding original insights
- Creating multiple websites to hide mass content production
- Publishing pages stuffed with keywords that don’t make sense to readers
Keep your content out of Google’s index if it isn’t genuinely created to help users. Focus on quality over quantity – your visitors (and Google) will thank you for it.
🔗 Read more here: https://developers.google.com/search/docs/essentials/spam-policies
How to Track Your AI Search Rankings
Thus far, our favorite tool remains SemRush which will show you which keywords trigger AIOs, there are two use cases for this:
- When you are researching: using the “Organic Research” tool you’ll be able to see keywords on any given website that have AIOs.
- When you are position tracking a project: in your position tracking project, your approve keyword based will show you which keywords are eligible for AIOs.
Appendix
Research notes
- Note that the percentages (%) used in this write up are meant for comparison purposes only to derive insights. To ensure we can do a trends analysis for the whole SaaS industry, we would need live access to 100 portcos +. But, for the purposes of this writeup the 20 SaaS companies suffice as a start.
- Note that Productive Shop has multiple initiatives running Q4 through Q1 to keep updating on AI trends. These numbers and this report will be updated with more information and sections.
About Google’s AI overviews (AIOs)
AI Overviews are a new type of search result that provides concise summaries of a topic directly on the search results page. Instead of sending users to a website for information, AIOs present key information and links to relevant sources right there.
Where do AI Overviews appear? They appear in Google Search results when the system determines that a generative response can be particularly helpful, such as when users are seeking quick overviews of complex topics.
What factors determine when AI Overviews are shown?
- User Intent: Google’s algorithms analyze the user’s search query to understand their intent. If the query suggests a need for a quick summary or an overview of a topic, an AIO might be displayed.
- Search Query Complexity: For complex topics with multiple facets, AI Overviews can be helpful in providing a structured and concise overview.
- Information Availability: Google’s systems assess the availability of high-quality information from a variety of reliable sources to generate accurate and informative AIOs.
- User Behavior: Google may also consider past user behavior, such as click-through rates on search results, to determine if an AIO would be a valuable addition to the search experience.
Tools included in this study
- chatgpt.com
- Perplexity
- Edgepilot
- Edgeservices
- copilot.microsoft.com
- openai.com
- gemini.google.com
- nimble.ai
- iask.ai
- claude.ai
- Aitastic.app
- bnngpt.com
- writesonic.com
- Copy.ai
