Findability in the age of AI: How to make your research discoverable

In the era of generative AI, getting your research or policy communications seen requires more than just publishing high-quality work — or relying on traditional web visibility tactics like SEO. That’s because the way people find and consume information is changing. Increasingly, users are turning to AI tools — from Google’s new AI-generated overviews to ChatGPT and Perplexity — to answer their questions, rather than clicking through to a list of search results.

These platforms can summarise and respond directly to queries, often without ever displaying or linking to the source website. If your content isn’t easily understood or surfaced by these AI tools, it may not reach your audience at all. That’s why organisations now need to think beyond standard optimisation — and embrace a new layer of visibility: GEO (Generative Engine Optimisation). GEO focuses on how well AI systems can interpret, trust and present your content — and ultimately, whether your voice is part of the conversation.

Beyond SEO: Why traditional tactics aren’t enough

For years, search engine optimisation (SEO) has been the cornerstone of online content discoverability – think keywords, metadata and backlinks. Those still matter, but they’re no longer sufficient on their own.

Search behaviour has changed:

Nearly 60–70% of Google searches now end without a click to any website, as users get what they need from a snippet or AI-generated answer.
Google’s AI Overview feature appeared in roughly 20% of searches in the US and UK by early 2025 (exposureninja.com ).
Tools like ChatGPT and Perplexity are being used by hundreds of millions globally to answer questions directly.

For think tanks, research organisations and policy publishers, this matters. It’s not just about ranking #1 anymore; it’s about being included in AI-driven answers. If your work isn’t picked up by these gatekeepers, it might not be seen at all.

How generative AI changes the game

Generative AI systems don’t just match queries to pages – they read, synthesise and generate responses. AI chatbots and assistants act as gatekeepers, deciding what information to present (blumango.be ).

This has mixed implications:

Your organisation might be cited as a credible source, boosting your authority.
Or your research might be summarised without attribution or misrepresented due to poor structuring or outdated content.

What AI prefers:

Content that is well-structured and easy to interpret.
Pages that include summaries or FAQs.
Reliable, up-to-date information from trusted sources.

If your content lives only in a dense 100-page PDF or lacks on-page clarity, AI may skip over it entirely.

The GEO basics: How to make your research AI-discoverable

Improving AI-driven “findability” means adapting classic SEO principles and layering on new techniques.

1. Structure your content for humans and machines

Use clear headings, concise summaries and logical page structures.
Add plain-language intros to long reports.
Consider creating machine-readable files like llm.txt to guide AI crawlers to key points.

2. Use schema and metadata to help machines understand

Implement Schema.org markup for authorship, publication dates and document types.
Use Open Graph and X card metadata for improved parsing by AI and social tools.
Add language tags and canonical URLs to prevent misinterpretation by multilingual or duplicate content.

Consider publishing a machine-readable summary file such as llm.txt at the root of your domain. These emerging conventions are designed specifically for generative AI crawlers, offering them a guide to high-value or priority content, key datasets, or summaries — helping your most important work be understood and surfaced more reliably in AI-generated answers.

3. Demonstrate authority and build trust

Include author bios, institutional affiliations and source citations.
Cross-link to other reputable sites and publications.
Encourage citations of your work in external outlets and Wikipedia.

4. Keep information up to date

Review older content regularly and prune or update it.
When publishing a new edition of a report, highlight what’s changed.
Ensure staff bios, dates, and policy positions reflect your current stance.

5. Prioritise accuracy and depth over clickbait

Aim to be the definitive, comprehensive source in your topic area.
Avoid factual errors, outdated stats or unclear claims.
Rich, well-evidenced content is more likely to be used by AI in generated answers.

6. Technical GEO: Don’t accidentally block AI

You might be preventing discoverability without realising it:

Some services (like Cloudflare) allow users to block AI crawlers — but this isn’t default behaviour. At Soapbox, we typically advise against full blocking. Instead, we help clients mitigate unhuman-like bot behaviour (e.g. excessive filters or rapid-fire queries) without preventing legitimate AI crawlers from indexing content.
Block bad bots, but ensure both AI crawlers and AI-powered search tools are allowed to access your public content — especially if your goal is citation and visibility.
Check your site’s robots.txt and security settings to ensure AI bots can crawl the content you want seen.

Why this matters: Influence, accuracy and equity

In policy and research, visibility shapes impact. If your work isn’t surfaced by AI assistants in response to key questions, you’re absent from the public conversation.

Poorly structured content can also be misquoted or simplified in ways that undermine your reputation. On the flip side, clear, well-tagged content increases your odds of being cited fairly and accurately.

Finally, there’s a competitive lens: organisations that adapt to GEO early will be more visible in AI-curated results than those that delay. Staying findable means staying relevant.

Next steps: Preparing your site for GEO

Here’s how to get started:

Audit your AI visibility: Try searching for your key topics in tools like ChatGPT, Perplexity, Google’s AI Overviews (in Search Labs, where available), or You.com.

At the moment, AI tools remain mostly closed-box when it comes to analytics. However, platforms like SEMRush and SimilarWeb can help simulate AI queries and track which sources are being cited in generated answers. These tools don’t offer full transparency but do provide directional insights — especially when monitored over time.

Beyond that, monitor traffic sources in your analytics. Referrals from OpenAI user agents or other AI-branded domains can help gauge how on-demand AI tools are engaging with your content.

Refactor your content: Add summaries, use schema and write with AI parsing in mind.
Check your crawl access: Make sure AI bots can actually reach the content you want seen.
Train your team: Build GEO considerations into your publishing workflows and style guides.
Ask for help: Soapbox offers discovery audits, AI visibility reviews and implementation support for research and policy teams. Contact us to find out more.

The future of discoverability

We’ve adapted before – to social, to mobile, to algorithmic news feeds. Now, it’s time to adapt to AI. That doesn’t mean chasing every trend. It means doubling down on clarity, authority and strategic content design.

Let’s make sure that when someone asks an AI about the topics you work on, your voice is part of the answer.

At Soapbox, we help research and policy teams make their content more discoverable — and more influential — in the age of AI. Whether you need an audit of how your site appears in generative search, guidance on GEO best practice, or hands-on support improving your content structure and metadata, we’re here to help.

Get in touch to future-proof your visibility.