Agentic Search Optimisation: Everything You Need to Know About ASO

Your next enterprise customer might not Google you. They might not visit your website, read your case studies, or book a demo before deciding you're worth considering. According to G2's March 2026 Buyer Behaviour Report, 51% of B2B software buyers now begin their research with an AI chatbot - not Google. That's up from 29% just eleven months earlier. Sixty-nine percent chose a different vendor than the one they originally planned, based on what the AI told them. One in three bought from a company they'd never heard of before.

If you’re feeling a bit overwhelmed by that information, same. 

There's now a new discipline: Agentic Search Optimisation or ASO. It's the practice of making your brand discoverable, trustworthy, and selectable by AI agents that are doing the vendor research your buyers used to do themselves. They’re not skimming your blog or browsing your G2 profile… they’re… buying. Here's what's happening, and what to do about it.

Key Takeaways

• AI buys software: 51% of B2B buyers now research with AI chatbots first and 69% end up choosing a vendor they didn't originally plan to.

• Shortlist shrinks: AI agents compress the vendor consideration process into a single answer. You're either in it or you're not.

• ASO ≠ AEO: Being cited in an AI answer is different from being selected by an AI agent acting on a buyer's behalf. Both matter but they're not the same thing.

• SEO stays foundational: Without good SEO, ASO is useless. 

• Most brands are invisible: 44% of B2B SaaS companies score below 50 on AI Presence Score despite having active marketing teams and real budgets.

What is Agentic Search Optimisation (ASO)?

ASO is the practice of making your brand discoverable, parseable, and trustworthy to AI agents that compare options and act on a buyer's behalf.

Act? Yes.

Search Engine Optimisation (SEO) gets you found by humans using a search engine. Answer Engine Optimisation (AEO) or Generative Engine Optimisation (GEO) gets you cited when an AI answers a question. ASO gets you selected when an AI agent is running a vendor comparison, building a shortlist, or … and we are already here … completing a transaction without a human in the loop.

Gartner predicts that by 2028, 90% of B2B buying will be AI agent-intermediated, representing over $15 trillion in spend through AI agent exchanges. I don’t know about you, but that makes me feel a little queasy. 

ASO became a formally named discipline in April 2026, when Adobe completed its acquisition of Semrush and put the acronym into mainstream marketing vocabulary. If you're wondering whether it's too early to care about this, Adobe's own analytics show AI traffic to retail sites jumped 269% year over year in March 2026. The wait and see window has closed.

The optimisation gang and where ASO fits

Here's the thing about all the acronyms that have colonised B2B marketing this year: SEO, AEO, GEO, ASO. They're layers, not competing disciplines. Here’s what you need to know about each:

Download the infographic here.

• SEO - be found: You rank in traditional search. Users click blue links. This has been going on since the late nineties and it still works. Anyone telling you SEO is dead is probably selling a GEO course.

• AEO (Answer Engine Optimisation) - be the answer: When a buyer asks ChatGPT which project management tool is best for a remote team of twelve, AEO is what gets your name cited in the response. According to Forrester, 89% of B2B buyers have adopted generative AI as a central source for self-directed research throughout the entire buying process. If you're not being cited, you're not in the running.

• GEO (Generative Engine Optimisation) - be visible across platforms: Similar to AEO but broader. It's about showing up consistently whether the buyer is on ChatGPT, Perplexity, Claude, or Google's AI Mode.

• ASO (Agentic Search Optimisation) - be chosen when no human is in the loop: The agent is now choosing your product. Semrush describes ASO as the broader umbrella covering not just answer inclusion, but how your brand shows up within agent-driven actions, including agentic commerce.

The agentic buyer journey: What it looks like

The traditional B2B SaaS buying process involved a lot of Googling, a lot of comparison sites, a multi-tab browser situation, and a lot of back-and-forth with procurement. That process hasn't disappeared but it's changing.

Here's the agentic version: A director of marketing at a mid-market tech company asks ChatGPT: "What are the best B2B content automation platforms for a team of four, budget under £2,000 a month?" ChatGPT returns three options with brief descriptions. She doesn't click any links. She asks follow-up questions directly in the chat, builds confidence, shortlists, then books a demo.

G2's March 2026 data found that 83% of buyers feel more confident in their final choice after AI chatbot research, and 4 out of 5 say it accelerated their decision. The brands that don't appear in that compressed process simply don't get considered.

Tim Sanders, Chief Innovation Officer at G2, said it better than most: "The Yellow Pages compressed the market into the big book. Google compressed it into the first page of results. Now, AI chatbots are compressing it into a single answer."

AEO vs ASO: The difference

They're not the same thing, and treating them as interchangeable is a mistake.

Features	Answer Engine Optimisation (AEO)	Agentic Search Optimisation (ASO)
Primary goal	To be cited as a trustworthy source in an AI's conversational response.	To be selected or acted upon by an autonomous agent.
The user	A human reading a generated answer and making a decision.	An AI agent performing a task (e.g., booking, procurement, shortlisting).
Key metric	Citation frequency and recommendation authority.	Conversion and inclusion in an autonomous shortlist.
Technical focus	Structured data, clear answers, and E-E-A-T (Trust).	Model Context Protocol (MCP), low token counts, and parseable data.
Optimisation factor	Being the most "trusted" answer to a query.	Being the most "efficient" option for an agent to process and act on.
Failure state	The AI provides an answer without mentioning your brand.	The agent truncates or skips your page because it's too bloated or slow.

AEO is about being cited when an AI answers a question. The user is reading the response and they're making the decision. Your job is to be the source the AI trusts enough to quote.

ASO is about being selected when an AI agent is acting autonomously. The agent is reading and then doing something with it. Booking a demo, running a procurement comparison, and adding a vendor to a shortlist without asking anyone first.

Search Engine Land draws the line here: AEO is about being the recommendation. ASO is about being chosen when no human is in the loop.

The technical difference matters too. ASO requires thinking about things like the Model Context Protocol (MCP), which allows agents to pull from your content and systems directly. 

If your website is slow, JavaScript-heavy, and structured in a way that AI bots can't parse, you will be invisible to agents even if humans love your site.

Addy Osmani, Google Cloud AI's director of engineering, published guidance in April 2026 that makes this concrete: AI agents fetch, parse, and act on pages differently than humans do. Token count is now an optimisation factor. Bloated pages get truncated or skipped. An agent may never reach your content if the key information is buried half way down your page.

The visibility problem most SaaS companies don't know they have

DerivateX's 2026 AI Visibility Benchmark Report analysed 50 B2B SaaS companies across ChatGPT, Perplexity, Claude, and Gemini (1,400 buyer-intent prompts in total). The average AI Presence Score was 56.9 out of 100. Forty-four percent of companies scored below 50.

The gap between the highest-scoring company (89 out of 100) and the lowest (2 out of 100) was 87 points despite both operating in established software categories. The difference was how their content was structured, distributed, and cited across the web.

And only 22% of marketers are currently tracking AI visibility at all, according to Loganix's April 2026 analysis. The problem is real and widespread. Most teams just don't know they have it.

5 ways to optimise for agentic search

Here are the main things that make a difference when it comes to optimising for agentic search (ASO).

Strategy	Key Actions	Why It Matters
1. Solidify SEO Foundations	Fix technical issues (speed, thin content), secure backlinks, and ensure overall site health.	AI agents use the same trust and authority signals that traditional search engines like Google rely on.
2. Build a Citation Portfolio	Secure guest posts, podcast features, press mentions, and high-authority reviews (G2, Capterra).	AI models build "confidence" in your brand based on third-party mentions in sources they already trust.
3. Prioritise Direct Answers	Place plain-language answers to common questions at the top of the page; keep quick-start content under 15,000 tokens.	Agents have limited context windows and will skip pages that bury answers under lengthy introductions.
4. Implement Deep Schema	Use FAQ, HowTo, and SoftwareApplication schema across all primary landing pages.	Schema acts as a direct translation layer, telling agents exactly what your product does without guesswork.
5. Track AI Visibility	Audit brand appearance in ChatGPT, Claude, and Perplexity; use tools like Semrush or Conductor for scale.	You need to benchmark where you appear (and where you don't) relative to your competitors to refine your strategy.

1. Sort your SEO first

ASO is built on SEO foundations. If your technical SEO is a mess (slow load speeds, missing schema, thin content, weak backlinks) no amount of AI-specific work will save you. 

The agents that do vendor research rely on the same signals of trust and authority that Google has always used. Get your SEO sorted first before tackling ASO - or even GEO/AEO.

2. Get cited in places AI trusts

AI models build confidence in your brand based on where you've been mentioned. Guest articles in trade publications, podcast appearances, press mentions, listings in high-authority directories, or Reddit threads where people recommend tools like yours. 

G2 and Capterra reviews are especially powerful… 45% of B2B buyers say citations from software review sites are the most confidence-inspiring signal in an AI-generated response.

Doing this digital PR isn’t about vanity metrics (anymore). You’re now building the citation portfolio that AI agents actually use to make decisions.

3. Answer questions directly, near the top of the page

AI agents have limited context windows. They can't always read your entire 3,000-word thought leadership piece. Addy Osmani's April 2026 guidance recommends keeping quick-start content under 15,000 tokens. 

Your answer to the buyer's question needs to be near the top of the page (specific, plain-language, and structured so it can be extracted cleanly).

If your page buries the answer under three introductory paragraphs and a stock photo of a handshake, an agent will skip you.

4. Use schema markup properly

FAQ schema, HowTo schema, and SoftwareApplication schema tell agents exactly what your content contains and what your product does. Most B2B SaaS sites either have no schema or badly implemented schema. 

They’re easy to create and implement so get those added to all of your primary pages as soon as possible. Here’s the generator and guide I used to create my schema markups.

5. Track your AI visibility

You can't improve what you don't measure. Start manually by asking ChatGPT, Perplexity, Claude, and Gemini your top category queries and see if your brand appears. 

Note where you show up, where you don't, and what your competitors are doing that you're not. For ongoing tracking at scale, tools like Semrush's AI Visibility Toolkit, Conductor's AgentStack, and Siteimprove's AEO Insights are all worth evaluating.

Do something. Start small. Because most of your competitors haven’t even started.

Is SEO dead, then?

No.

Anyone telling you otherwise is either panicking or selling a course.

Gartner predicted traditional search volume will drop 25% by 2026 due to AI chatbots and virtual agents. It’s a drop, yeah, but it’s not a death. Things have changed, is all. 

Think about it practically: an AI agent crawling the web to research vendors still needs pages to crawl. It needs backlinks to assess authority. It needs structured, indexed content to pull from. SEO creates the infrastructure that ASO and AEO run on. Without it, there's nothing for the agents to find.

The brands winning in AI search right now are the ones who had great SEO to begin with and built a strategy where all three layers (SEO, AEO, ASO) work together. 

Frequently asked questions 

What is Agentic Search Optimisation (ASO)?

ASO is the practice of making your brand discoverable and selectable by AI agents that act on a user's behalf: running vendor comparisons, building shortlists, and in some cases completing purchases. 

Unlike AEO (which is about being cited in AI answers), ASO is about being chosen when the agent is making (not informing) a decision. The term became a formally recognised discipline in April 2026.

How is ASO different from AEO?

AEO (Answer Engine Optimisation) gets your brand cited when an AI answers a question. The human reads the response and decides. ASO is the next layer and it's about being selected by an AI agent operating autonomously, without waiting for human input at each step. AEO is ‘be the answer.’ ASO is ‘be chosen when no one is watching.’

How is ASO different from SEO?

SEO gets you found in traditional search results. ASO gets you selected by AI agents evaluating vendors on a buyer's behalf. They share the same foundations like technical health, domain authority, structured content, but the audience is different. SEO optimises for humans clicking links while ASO optimises for machines acting on behalf of humans.

Is SEO still worth investing in for B2B SaaS?

Yes. SEO is still worth investing in for B2B SaaS. SEO is the foundation everything else is built on. AI agents still rely on backlinks, indexed content, and technical signals to assess trust and authority. Abandoning SEO to focus on GEO or ASO is like pulling the load-bearing wall out to refit the kitchen. Make sure your SEO is strong before moving onto GEO or ASO.

How do I know if my B2B SaaS brand is visible to AI agents?

Start manually by asking ChatGPT, Perplexity, Claude, and Gemini your top category queries. These are the questions a buyer would ask when researching a tool like yours. If your brand doesn't appear, you have an AI visibility problem. For more reliable, ongoing tracking, look at Semrush's AI Visibility Toolkit or Conductor's AgentStack.

What's the quickest win for improving AI visibility?

Get cited in authoritative third-party sources. G2 and Capterra reviews, guest articles in trade publications, podcast appearances, and press mentions all contribute to the citation portfolio that AI models use to build confidence in your brand. Review site citations carry particular weight as 45% of B2B buyers say they're the most confidence-inspiring signal in an AI-generated response, according to G2's 2026 Buyer Behaviour Report.

Ready to audit your brand's AI visibility? Book an Introductory Content Audit and find out where you're showing up and where you're not.