The world of SEO has changed – fast and forever. Optimising for AI search is a whole new world – and whilst there is certainly some overlap with old school SEO, there are also lots of areas requiring a fresh approach and a mindset shift.
As a B2B software focused AI SEO agency, here is an overview of some of the things we look at when improving a B2B software brand’s AI answer visibility.
1. Deeply Understand Your Audience
Before you write a single word of content and before you start tracking prompts or trying to optimise anything, you need to know who you’re optimising for.
Really this is true of any marketing strategy – it’s definitely not new because of AI SEO. But it is often overlooked, and now more important than ever as part of a GEO strategy.
But in the world of SEO, we had keyword volumes. In the world of GEO, we don’t.
Which means we believe everything must start with deep buyer intelligence.
At FirstMotion, we’ve built a platform focused on B2B software buyer journeys designed for exactly this:
- Map your ICPs, their personas, their roles in the buying unit
- Understand the stages of their journey, use cases, and evaluation triggers
- Analyse competitor reviews to semantically understand buyer language
- Enrich the data using intent data, millions of B2B buyer data points and AI to deeply understand buyer behaviour
No prompt strategy is complete without deep, enriched buyer context.
2. Mine Prompts
Only once you have a deep audience and buyer behaviour understanding can you begin to ‘mine’ prompts that might be being used at the various stages of the buyer journey. Our platform takes all of the audience intelligence data and uses AI to help with this process.
Our methodology at FirstMotion is to align prompts by B2B buyer journey stages, building Prompt Matrix that maps prompts across ICPs, personas and stages.
By default we use the following buyer journey stages, but customise this for each client, ICP and decision making unit. Here is an example of a General Counsel in a mid size tech company exploring contract lifecycle management software solutions:
| Buyer Journey Stage | Prompt Example |
|---|---|
| Problem Identification | How do other GCs in mid-sized companies handle version control and contract visibility across departments without relying on email chains and manual tracking? |
| Solution Exploration | Can you compare leading contract management tools like Ironclad, LinkSquares, and ContractWorks in terms of ease of use, legal team adoption, and integration with Google Workspace and Salesforce? |
| Requirements Building | What features should a mid-sized tech company’s legal department look for in a CLM solution if we want to automate clause fallback, ensure version history, and allow cross-functional access (e.g. sales, finance, procurement)? |
| Supplier Evaluation | Based on reviews from legal teams in mid-sized SaaS companies, how does Ironclad compare to LinkSquares in terms of implementation speed, usability, and legal team satisfaction? |
Notice that prompts are much longer and more contextual than ‘SEO keywords’ – and whilst the above are just examples, we would expect to see even longer prompts on a real client engagement.
3. Analyse Prompt Visibility
Once you’ve ‘mined’ your prompts, the question becomes: are we showing up in generative answers currently for these prompts?
This is where prompt visibility tools like Peec come in. Peec is our analytics and visibility monitoring partner at FirstMotion.
We use Peec to:
- Track brand’s visibility across AI tools like ChatGPT, Perplexity, and Google AI Overviews
- Compare visibility against competitors for the same prompts
- Monitor how visibility shifts over time
A key mindset shift for marketers to make here that there are no “rankings” in generative search – there are only probabilities.
4. Run Influence Intelligence
If your brand isn’t showing up, you don’t just need to know who is – you need to know why,
LLMs pull from sources they trust. And increasingly, that’s not your homepage.
Using Peec we are able to analyse citations and sources – the content that is influencing the answers tools like ChatGPT may give users.
Here is an example from one of the prompts relating to the contract lifecycle management software example above, showing the specific URLs the various models have cited and how frequently:
As we specialise in B2B software, we often see sources like Gartner Magic Quadrants, Forrester Wave reports, LinkedIn Pulse, G2 and Capterra reviews and niche forums like Quora or Reddit being referenced by LLMs. If you would like to know more about how these sources might be influencing your own AI SEO visibility, get in touch.
5. Create a Prompt Aligned Content Strategy
Aligning content with your prompt matrix and based on your influence intelligence is worthy of its own post that would be thousands of words long, so we’re not going to go deep into the process here.
In the GEO era, the most effective content is shaped by the actual language and context your buyers use when interacting with AI assistants. So here are a few things to think about for now:
- Stay niche and really break down content by ICP, persona and buyer stage – journey context is key
- Don’t just mirror the prompt – answer the users intent with scannable, citable, contextual content
- Think about how content can be easily parsed and summarised by LLMs and if summaries, clear headers, lists and schema markup might help
- Focus on semantic richness over keyword stuffing
- Ensure content remains fresh and up to date with your own citations and references
- Think about your offsite content strategy and where content could be republished
6. Define an Offsite Strategy
Our research has shown that, for certain types of prompt at certain buyer journey stages, LLMs often prefer third party citations over your own site’s content. This is one of the most large and complex areas of a generative engine or AI search optimisation strategy.
Branded mentions in locations beyond your own website are a factor we’re tracking carefully.
So some areas to think about might include:
- G2, Capterra, TrustRadius – high-quality reviews and clear positioning
- Reddit, Quora, LinkedIn Pulse – authentic content and commentary
- Analyst reports from Gartner and Forrester
- Integration directories, analyst blogs, product roundups
Tactics:
- Review campaigns on trusted platforms
- Create content for third party guest publications
- Comment and contribute to the discussion where LLMs learn in places like Reddit and Quora
- Think about content republishing and repurposing, and if it makes sense to build authority in other places beyond your own website
5. Monitor Visibility Continuously
GEO isn’t static and models are updating regularly. Prompt phrasing shifts. Citations change. And monitoring visibility is therefore both more complex and more important than in the old school days of SEO, when Google was really the only search engine you had to worry about.
You need ongoing visibility monitoring to:
- Track how often your brand is showing up for your selected prompts
- See which models favour your content
- Spot shifts in competitors’ inclusion rates
- Identify which content or source changes lead to visibility changes
7. Don’t Abandon SEO Fundamentals
This isn’t a binary choice between SEO and GEO. In lots of our research we see positive correlation between brands with historically good SEO also performing well in AI search.
But hopefully it’s also clear from our post that there are lots of new things to think about and some mindset shifts that need to happen in the new era of AI search and generative engine optimisation. Maybe it’s time to build the business case to invest in AI SEO.
One important note: OpenAI has a partnership with Bing through it’s relationship with Microsoft – meaning when a ChatGPT prompt triggers a web search, it relies on Bing’s index. So whilst your previous SEO strategy might have only focused on Google, it may be worth thinking about how your site is indexed in Bing too.
If you’re a B2B software brand looking for help understanding the shifting landscape of SEO and AI search, we would love to talk.