The short answer
How should software teams divide budget between GEO and SEO?
Do not start from a percentage split. Start from tasks, because most of the work sold as GEO is also SEO: pages that can be indexed, unique content, accurate facts and genuine third-party coverage. Google's guide (2026a) says optimising for generative AI search is "still SEO" for its own AI features. Fund the shared work first, then fund the genuinely separate tasks (sampling AI answers, community and video presence) according to where your buyers actually ask. We found no evidence for any fixed ratio, including the 70/30 split this article used to recommend.
What do SEO and GEO mean for a budget?
SEO covers the work that makes pages findable, understood and chosen in search results: crawlability, content, links and reporting. GEO is the label for work aimed at being named or cited in AI answers. The two overlap heavily. In Google’s own framing, work aimed at its AI features is SEO. For other engines (ChatGPT, Perplexity, Gemini, Claude), there is no equivalent guide, so the evidence is third-party. See our engine evidence grid.
Which tasks overlap?
Use the calculator below. Edit the monthly costs and tags to match your plan. The default tags are our judgement and the default costs are placeholders.
SEO and GEO task overlap calculator
Edit the monthly costs and tags to match your plan. Default costs are placeholders, and the default tags are our judgement. Nothing leaves your browser.
| Task | Monthly cost | Tag |
|---|
- Total
- -
- Shared (helps classic search and AI answers)
- -
- AI-specific
- -
- SEO-only
- -
- Optional or unproven
- -
- Share of budget that is shared work
- -
Text version of the tag logic.
| Task | Our default tag | Reason |
|---|---|---|
| Crawlability, indexation and page experience | Shared | Google says a page must be indexed and snippet-eligible to appear in its AI features (Google, 2026a) |
| Unique, non-commodity content | Shared | Google says this will likely matter more than anything else in its guide (Google, 2026a) |
| Consistent product and company facts | Shared | Reduces conflicting information in classic results and AI answers |
| Genuine third-party mentions, reviews and coverage | Shared | Ahrefs found branded web mentions correlate with AI mentions (0.656 to 0.709, correlation only) (Linehan, 2025b) |
| Structured data for rich results | SEO-only | Google says it is not required for generative AI search (Google, 2026a) |
| Community and video presence, with transcripts | AI-specific | One practitioner reports Reddit, YouTube and forums dominate ChatGPT citations in his studies; YouTube mentions correlate about 0.737 in Ahrefs’ data (Chmurkowski, 2026; Linehan, 2025b) |
| AI answer sampling and reporting | AI-specific | No third-party tool sees Google’s internal systems, so you sample (Google, 2026a) |
| llms.txt and similar files | Optional or unproven | Google Search ignores them (Google, 2026a) |
| Classic rank tracking and Search Console reporting | SEO-only | Standard reporting |
With the placeholder costs in the calculator, shared tasks come to about 72% of the budget. That is an illustration of the method, not a recommended ratio.
When does the separate AI work deserve more money?
These are our decision rules. Claim strength: opinion.
- Your buyers ask AI tools. If sales conversations or first-touch surveys show people arriving from AI answers, fund sampling and community presence earlier. Over roughly the last six months, the best-informed inbound prospects at our agency mostly said they arrived via ChatGPT or Gemini. That is our own experience, observed and not measured, and not a benchmark.
- Your category is crowded with comparison and review content. Genuine reviews and third-party coverage carry more weight in shortlists.
- You are regulated or cautious. Accuracy and consistency matter more than volume, so lean on shared work and corrections.
- You are early-stage in a mature category. Shared work comes first, because AI features rely on the same index as search.
What does the evidence say about traffic?
Pew’s analysis (Chapekis and Lieb, 2025) of 68,879 Google searches (900 US adults, March 2025) found a click on a standard result in 8% of visits when an AI summary appeared and 15% when it did not, a click on a link inside the summary in 1% of visits, and sessions ending on the results page 26% of the time with a summary versus 16% without.
Claim strength: measured by Pew, US only, March 2025. The implication for a budget is narrow: report mentions, visits and conversions together, because visibility and traffic can diverge.
How should you measure both?
- SEO: impressions, rankings, clicks and conversions in Search Console and analytics.
- Google AI features: the Generative AI performance report in Search Console.
- Other engines: repeated sampling of a fixed prompt set, with the appearance rate and its uncertainty. See Tracking AI Visibility.
- Business: qualified leads, revenue, assisted conversions and customer acquisition cost.
What mistakes should you avoid?
- Treating GEO as a replacement for SEO. Google’s AI features rely on its core systems.
- Buying volume. Google warns that creating pages for every fan-out query mainly to manipulate AI responses violates its scaled content abuse policy (Google, 2026b).
- Manufacturing mentions. We pursue natural mentions only where they fit, and we do not manufacture them. Google’s guide says inauthentic mentions are not as helpful as they seem (Google, 2026a).
- Locking in a ratio without measurement.
What does a 90-day plan look like?
Our suggestion, not a measured result.
- Days 1 to 30: audit indexation, content and consistency of facts. Build a prompt set of real buyer questions (our prompt set builder helps) and sample it weekly.
- Days 31 to 60: fix the shared work first. Publish at least one non-commodity page per core topic.
- Days 61 to 90: compare the shared work and the AI-specific tasks by their outcomes (qualified leads, mentions), then reset the budget.
Key takeaways
- A split such as 70/30 or 50/50 has no published evidence behind it, and we found none. Treat any ratio as an assumption you set and revisit.
- Google says GEO and AEO are still SEO for AI Overviews and AI Mode, and that no special files, markup or chunking are needed.
- Most tasks have a shared benefit. Only a few are specific to AI answers, such as sampling answers across engines and community or video presence.
- Pew's study of 68,879 Google searches found clicks on a standard result fell from 15% to 8% when an AI summary appeared (Chapekis and Lieb, 2025). Clicks are a poorer measure of AI visibility, so budget for measurement.
- Fund by task, label each one shared, AI-specific, SEO-only or optional, and review quarterly.
References
- Chapekis, A. and Lieb, A. (2025) 'Google users are less likely to click on links when an AI summary appears in the results', Pew Research Center, 22 July. Available at: https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/ (Accessed: 22 October 2026).
- Chmurkowski, M. (2026) 'How to Build Brand Visibility in AI' [Podcast], Globalne Horyzonty. Interviewed by R. Moszkowcow (recorded July 2026), 18 September. Available at: https://moszkowcow.com/podcast/maciej-chmurkowski-widocznosc-w-llm/ (Accessed: 22 October 2026).
- Google (2026a) 'Optimizing your website for generative AI features on Google Search', Google Search Central. Available at: https://developers.google.com/search/docs/fundamentals/ai-optimization-guide (Accessed: 20 October 2026).
- Google (2026b) 'Spam policies for Google web search', Google Search Central, last updated 28 August. Available at: https://developers.google.com/search/docs/essentials/spam-policies (Accessed: 20 October 2026).
- Linehan, L. (2025b) 'Top brand visibility factors in ChatGPT, AI Mode, and AI Overviews (75k brands studied)', Ahrefs Blog, 12 December (updated 12 August 2026). Available at: https://ahrefs.com/blog/ai-brand-visibility-correlations/ (Accessed: 22 October 2026).



