There is a measurable gap between brands that LLMs mention and brands that LLMs actively recommend. Most SaaS companies have not started measuring it. The few that have are quietly winning the AI search game.
This post defines Brand Recommendation Rate (BRR), gives the formula, and breaks down the seven reasons most SaaS companies score below 12%.
What is Brand Recommendation Rate (BRR)?
Brand Recommendation Rate is the percentage of buyer-intent prompts, across major LLMs and AI search engines, where your brand appears as a recommended option in the answer.
The formula:
BRR = (Prompts where AI engines recommend your brand / Total buyer-intent prompts tested) × 100
A buyer-intent prompt is any question a real prospect would ask an LLM before purchasing. Examples for a SaaS company:
- “What is the best [your category] software for [your ICP]?”
- “Which [category] tool should I use if I need [feature]?”
- “What do you recommend for [job to be done]?”
- “Compare top [category] platforms”
- “Alternatives to [your largest competitor]”
Run those prompts across ChatGPT, Claude, Perplexity, Gemini, and Google’s AI Mode. Count the responses where your brand shows up as a recommended option. Divide by the total prompts tested. That is your BRR.
How BRR differs from the metrics you already track
BRR is not NPS. NPS measures whether your existing customers would recommend you. BRR measures whether AI engines do, to people who are not yet customers.
BRR is not citation rate. Citation rate counts how often AI engines link to your content. You can have a high citation rate (LLMs link your blog) and a low BRR (LLMs do not recommend your product when asked). These are different things and the conflation is costing SaaS teams money.
BRR is not brand mention rate. Mention rate counts every appearance of your brand name, including incidental ones. BRR only counts recommendations in buyer-intent contexts.
The hierarchy looks like this:
- Mention: “Tools in this space include [your brand], among others.”
- Citation: “[Your article] explains the category well.”
- Recommendation: “I’d recommend [your brand] for [specific use case].”
Recommendation is the only one that converts. The other two are noise unless they are feeding the recommendation layer.
Why BRR matters more in 2026 than rankings did in 2020
In 2020, ranking #1 for “best CRM for startups” meant something. Real clicks. Real pipeline. Real attribution back to a keyword.
In 2026, your future customers are not asking Google “best CRM for startups.” They are asking ChatGPT “what CRM should I use for a Series A B2B startup with a 4 person sales team in the US?”
The answer to that question is not a ranked list. It is a curated recommendation. The LLM picks 2 to 4 brands and recommends them. Everyone else is invisible in that conversation, regardless of how well they rank on Google.
Your BRR is the measurable version of this. If your BRR is 5%, you are recommended in 1 of every 20 buyer-intent prompts in your category. If a competitor’s BRR is 35%, they are recommended seven times more often than you. The future revenue gap that creates is enormous and it compounds, that’s why AI SaaS Marketing Agency is important.
Why your SaaS BRR is low
In our audits across SaaS categories, most companies score between 2% and 12% on BRR. The causes cluster into seven patterns. Most companies have at least four of them at once.
1. You are absent from the platforms LLMs use as ground truth
LLMs are trained on and reference high-trust public corpora: Reddit, YouTube, top-tier publications, Wikipedia, G2, Capterra, GitHub, Stack Overflow, and a small set of editorial sites. If your brand is absent from those, the LLM has no reason to surface you when a user asks for recommendations.
Most SaaS companies invest heavily in their owned blog and ignore everything else. Your blog could be brilliant. The LLM weight on your owned content is a fraction of what it gives to a four year old Reddit thread where three users compare your category. If you are not in that Reddit thread, you are not in the AI’s mental model of the category.
2. You have no proprietary data or named methodology
LLMs prefer to recommend brands that have something specific to attribute. A study, a benchmark, a framework you named, a public dataset you maintain.
If a user asks “what should I use to track product analytics,” an LLM is more likely to recommend the company that publishes an annual “State of Product Analytics” report than the equivalent company that publishes 50 listicles a year. The report becomes the artifact LLMs cite, and the brand becomes the entity associated with category authority.
If you have no proprietary research, no named frameworks, and no public data, you are competing for recommendation slots with one hand tied behind your back.
3. Your category positioning is fuzzy
LLMs categorize brands. Your category language has to be tight enough that the LLM puts you in the right bucket without confusion.
We see this constantly. A company sells revenue intelligence software but their website calls them “an AI-powered platform for modern sales teams.” The LLM has no idea where to place them. So when a user asks “what is the best revenue intelligence software,” the LLM recommends the brands whose category is unambiguous.
If your homepage hero does not tell an LLM what category you compete in, you forfeit category recommendations.
4. Your third-party presence is weak
LLMs use review aggregators as evidence of legitimacy. G2, Capterra, TrustRadius, Product Hunt, and a few category-specific sites carry disproportionate weight. Not just for the star ratings, but for the depth of language in the reviews.
A SaaS company with 12 G2 reviews is invisible to the LLM compared to a company with 400, even if the 12 are five stars. Volume of authentic third-party language is what the LLM uses to calibrate confidence in a recommendation.
This is solvable but slow. It also cannot be faked, because LLMs are getting better at detecting astroturf and downweighting suspicious clusters of reviews.
5. You have no point of view
LLMs recommend brands that have taken clear positions. A SaaS company that says “we believe X about how this category should work, and here is the framework we use to operationalize it” gives the LLM something to anchor to.
A SaaS company whose entire content output is “5 tips for choosing a vendor” and “what is [category] software” gives the LLM nothing to differentiate them from 30 competitors saying the same thing. So the LLM either does not recommend them, or recommends them as a generic option that loses the click-through.
Point of view is not a luxury anymore. It is how LLMs decide who to recommend.
6. Your founder and team are invisible online
LLMs use entity associations to build recommendations. If your founder has a strong public presence with category-relevant content on LinkedIn, podcasts, and YouTube, the LLM associates that founder with the category. That association strengthens the brand’s recommendation eligibility.
If your founder posts once a quarter and your team is anonymous, the LLM has no human signal to anchor the brand to. Brand trust in AI search has a human layer. Companies that hide their team are skipping a free trust signal.
7. You optimize for traffic, not for citation
This is the deepest issue. Most SaaS SEO content is written for clicks. Catchy headlines, hooks designed to win in search snippets, conclusions designed to push toward a trial.
That content does not get cited by LLMs. LLMs cite content that defines, explains, and quantifies. They cite content that reads like a reference, not a sales pitch. The blog posts that win recommendations are usually the ones that do not try to convert at all, because they earn the LLM’s trust by being useful first.
If your entire content strategy is conversion-optimized, your BRR will stay low.
How to raise your BRR
The companies we see making meaningful progress on this typically move from low single-digit BRR into the 15% to 25% range over 9 to 12 months. The pattern is consistent.
Start by measuring. Build a prompt set of 50 to 100 buyer-intent queries in your category. Run them across the major LLMs monthly. Record where your brand appears and where competitors do. That baseline is your starting BRR.
Audit your presence on the platforms LLMs trust. Reddit, YouTube, the top three publications in your category, G2, Capterra, and any niche communities where your buyers actually talk. Build a 6 month plan to populate those surfaces with genuine, useful presence.
Produce one piece of proprietary research per quarter. A survey, a public dataset, an annual report, a named methodology. Anything an LLM can cite as a specific artifact. This is the single highest-leverage action for raising BRR.
Tighten your category language. Your homepage, your meta descriptions, your About page, your G2 category placement. Use the exact phrases buyers and LLMs use to describe your category. Stop trying to invent new category names that the LLM does not recognize.
Make your team visible. Your founder, your head of product, your head of engineering. Get them on podcasts, posting on LinkedIn, writing under their own names. Brand trust in LLM answers correlates with human signal.
Audit your content for citation potential, not click potential. Add clear definitions. Add data. Add named frameworks. Make every post the kind of thing an LLM could quote without rewriting.
Then measure again at 30, 60, and 90 days. BRR moves slower than rankings do, but it compounds. The brands that establish BRR leadership in their category in 2026 will be very hard to dislodge in 2027 and beyond.
The bigger picture
Brand Recommendation Rate is not the only AI search metric worth tracking. Citation rate, mention rate, and traditional rankings still matter. But BRR is the closest measurable proxy for what actually moves revenue in an LLM-dominated buyer journey.
Most SaaS marketing teams will spend 2026 optimizing for metrics that no longer translate to pipeline. The ones who measure BRR, understand why their score is low, and systematically work to raise it are going to look like outliers two years from now.
They are not outliers. They just measured the right thing earlier.
FAQs
What is a good BRR score for a SaaS company?
Anything above 25% is exceptional and usually indicates category leadership in AI search. 15% to 25% is strong. 5% to 15% is the typical mid-market range. Below 5% means your brand is functionally invisible to LLMs in buyer-intent conversations.
How is BRR different from NPS?
NPS measures the likelihood that your existing customers will recommend you to others. BRR measures the likelihood that AI engines will recommend you to potential customers who are not yet in your funnel. NPS is a customer loyalty metric. BRR is an acquisition metric.
How do you measure BRR?
Define 50 to 100 buyer-intent prompts in your category. Run them across ChatGPT, Claude, Perplexity, Gemini, and Google’s AI Mode. Count the prompts where your brand appears as a recommended option. Divide by the total prompts tested. Multiply by 100.
How long does it take to raise BRR?
Most of the movement happens between months 3 and 12. The first 90 days establish the baseline and build the infrastructure (proprietary research, third-party presence, team visibility). Citation patterns in LLMs typically start shifting after 90 days and compound from there.
Can BRR be gamed?
No, and attempts to game it backfire quickly. LLMs are increasingly good at detecting astroturf, paid placements, and review manipulation. BRR rises through genuine category authority. There is no shortcut.
