TL;DR:
This is a real-world example of an AI Search Visibility Audit (also known as a Generative Engine Optimisation or GEO audit), run using AttributeIQ, a B2B marketing attribution product developed by our team, as the test case. The methodology and findings are real. We’re publishing it in full so other brands can see exactly what an audit like this involves, and use it as a benchmark for evaluating their own AI search presence.
How We Ran This LLM Visibility Audit
To make the findings representative of how AI can influence a real buying journey, we kept the test deliberately controlled and benchmarked AttributeIQ against the competitors a buyer is most likely to consider.
| Element | Detail |
|---|---|
| Buyer prompts tested | 15 high-intent commercial prompts covering discovery, comparison, pricing, and validation. |
| AI platforms tested | ChatGPT, Perplexity, Gemini, Claude |
| Test conditions | Clean sessions, zero historical context, default web-browsing capabilities enabled. |
| Competitors benchmarked | Dreamdata, HockeyStack, Ruler Analytics |
| Evaluation criteria | Brand inclusion rate, factual accuracy, positioning sentiment, source citation authority. |
Executive Summary: AttributeIQ’s AI Search Performance Benchmarks
This section provides a high-level readout of where AttributeIQ currently stands across major LLMs and the immediate commercial opportunity available by fixing the gaps.
1. Multi-Platform AI Visibility Snapshot
Across the 15 high-intent buyer prompts tested, AttributeIQ achieved a favorable or factually accurate result on just 2 prompts (13% overall visibility rate).
| AI Platform | Favourable / Accurate Prompts | Visibility Rate | Position | Executive Readout |
|---|---|---|---|---|
| ChatGPT | 1/15 | 7% | 🥇 Strongest | Most consistent representation, but weak across competitive comparison queries |
| Perplexity | 1/15 | 7% | 🥇 Strongest | Shows AttributeIQ occasionally, but relies heavily on sparse third-party data |
| Gemini | 0/15 | 0% | - | Zero favorable representation across all tested B2B buyer prompts |
| Claude | 0/15 | 0% | - | Brand is rarely cited or confidently recommended for marketing attribution |
| Overall | 2/15 | 13% | - | AttributeIQ is favourably represented in only a small share of high-intent prompts |
2. 90-Day AI Visibility & Growth Opportunity
With favourable or accurate representation currently limited to 2 of 15 prompts, the 90-day targets below set out what a meaningful improvement could look like across buyer discovery, competitive evaluation, and independent proof.
| Metric | Current State | 90-Day Target | Commercial Implication |
|---|---|---|---|
| Favourable representation in buyer searches | 2/15 favourable | 8-10/15 | Creates more opportunities for AttributeIQ to enter the shortlist when buyers use AI to research attribution software. |
| Pricing accuracy across AI platforms | 0/4 correct | 4/4 correct | Prevents incorrect price expectations from carrying into the buying process and sales conversation. |
| Competitive comparisons with AttributeIQ | 0/3 controlled | 3/3 controlled | Gives AI reliable first-party information to explain AttributeIQ’s differences rather than relying primarily on competitor narratives. |
| Presence when buyers seek alternatives | Rarely surfaced | Consistent inclusion | Creates a route into searches where buyers have already identified a competitor or existing solution and are actively considering alternatives. |
| Access to product and attribution methodology | Weak/ inconsistent |
Clear and verifiable | Makes it easier for AI to accurately explain how AttributeIQ works, supporting confidence among buyers evaluating the product in more detail. |
| Independent reviews and third-party evidence | Incomplete | Complete and current | Provides stronger external validation when prospects move beyond product claims and look for evidence they can trust. |
3. Priority Actions by Commercial Impact
Rather than simply listing what needs to be fixed, we prioritised the actions by the weight of their commercial consequences, addressing the issues that could most directly affect buyer decisions first.
| Priority | Action | Commercial Consequence | Owner | Effort | Timing |
|---|---|---|---|---|---|
| 1 | Correct AttributeIQ pricing across AI-facing and third-party sources, ensuring the £89 / £149 / £299 plans are consistently represented. | Highest immediate risk. Prospects researching AttributeIQ may be given an outdated, lower price and enter the buying process with the wrong budget expectation. | Marketing + Engineering | Low | Week 1 |
| 2 | Build substantive AttributeIQ-side comparisons for Dreamdata, HockeyStack and Ruler Analytics | Competitors currently have more material for AI systems to draw from when generating comparisons, increasing the likelihood that AttributeIQ is omitted, mischaracterised or positioned as the weaker option. | Marketing | Medium | Weeks 1-4 |
| 3 | Make core product documentation reliably accessible | Existing product knowledge is harder to verify than it should be. This increases reliance on generic descriptions or third-party sources and makes accurate product recommendations less likely | Engineering | Low-Medium | Weeks 1-2 |
| 4 | Complete and strengthen G2, HubSpot Marketplace and other relevant third-party profiles | AI-generated recommendations are influenced by the amount and quality of independent evidence available. Thin or incomplete profiles leave competitors with a stronger external evidence base | Marketing + Partnerships | Medium | Weeks 3-8 |
| 5 | Strengthen category and buyer-fit positioning | AttributeIQ is less consistently associated with broader searches around attribution software and specific buyer needs, limiting discovery beyond direct brand or product searches. | Marketing | Medium | Weeks 4-8 |
15 AI Search Prompts Used to Evaluate AttributeIQ’s Visibility
These 15 prompts form the core of the audit, taking AttributeIQ through the questions a buyer is likely to ask when researching the category, comparing vendors, checking fit, pricing and credibility.
| # | Buyer Prompt | Result | Priority |
|---|---|---|---|
| 1 | What is the best marketing attribution software for B2B SaaS? | AttributeIQ appeared in some responses, but never as the leading recommendation. The platforms that did surface it generally provided limited detail on why it would be a strong choice for a B2B SaaS buyer. | High |
| 2 | AttributeIQ vs Dreamdata | Dreamdata had substantially more information available for AI to draw from, particularly from its own website and supporting content. The resulting comparisons therefore leaned heavily on Dreamdata’s positioning. | High |
| 3 | AttributeIQ vs HockeyStack | AI could identify both products as relevant to the category, but there was limited independent information available to support a detailed comparison of their capabilities, positioning or relative strengths. | High |
| 4 | AttributeIQ vs Ruler Analytics | AttributeIQ was frequently absent when the two products were compared. Where it did appear, AI had limited information with which to explain the differences between the platforms. | High |
| 5 | How much does AttributeIQ cost? | Every platform quoted a price below what AttributeIQ actually charges (£89 / £149 / £299), and they all landed on roughly the same wrong number. | Critical |
| 6 | Does AttributeIQ integrate with HubSpot? | AI generally understood that AttributeIQ integrates with HubSpot and could communicate this without significant ambiguity. | Strength |
| 7 | Is AttributeIQ good for multi-touch attribution? | AI recognised AttributeIQ as relevant to multi-touch attribution, but provided limited detail about how AttributeIQ actually approaches the problem or what distinguishes it from other platforms. | Medium |
| 8 | What are the reviews of AttributeIQ? | Responses had limited customer evidence with which to support a confident assessment of the product. | High |
| 9 | What are the alternatives to AttributeIQ? | AI could identify several alternatives, but the responses gave little detail on how the products differed or which option might make the most sense for a particular buyer. | Medium |
| 10 | What is a more affordable alternative to Dreamdata? | AttributeIQ appeared rarely, despite being relevant to the underlying requirement. | High |
| 11 | What attribution tool is right for a Series B SaaS company? | AttributeIQ was not strongly associated with this specific buyer context. | Medium |
| 12 | AttributeIQ vs HubSpot attribution? | AI had considerably more material available to explain the existing HubSpot option than to make the case for using a dedicated attribution platform. | Medium |
| 13 | How does AttributeIQ calculate pipeline attribution? | AI struggled to provide a clear explanation of the methodology and could not confidently verify the underlying information. | High |
| 14 | What is the best tool for proving marketing ROI to the board? | AttributeIQ barely came up here. The responses leaned towards platforms that were more strongly associated with reporting marketing performance and demonstrating ROI to senior leadership. | High |
| 15 | What is AttributeIQ’s G2 rating? | AI could not give a confident answer on AttributeIQ’s G2 rating, with the available information either missing or difficult to verify. | High |
AI Visibility Findings, Commercial Impact and Recommended Actions
Below, we’re going to drill down into what each finding means and how these specific gaps affect AttributeIQ’s pipeline and revenue.
1. Pricing Hallucination: Finding 5
What we found
All four AI platforms gave AttributeIQ a lower price than the current £89 / £149 / £299 plans. The fact that they all landed on roughly the same figure suggests the old price is still being repeated somewhere online.
Commercial Risk
A buyer can walk into a sales conversation already believing AttributeIQ costs less than it does. Once the actual price comes up, it raises a question about which information they can trust.
Action
Make the AttributeIQ pricing page the canonical source for current pricing, with the £89 / £149 / £299 plans clearly stated. Then update or correct every third-party listing carrying the old price and add structured pricing information so AI systems have a clear source to reference.
👤 Marketing + Engineering ⏱ Week 1
2. No Owned Comparison Content: Findings 2-4
What we found
There is very little AttributeIQ content designed to answer direct competitor questions. For Dreamdata, HockeyStack and Ruler Analytics, the available comparison material is therefore weighted towards what those companies publish themselves.
Commercial risk
These searches often happen when a buyer has already narrowed the field. If AttributeIQ is not able to clearly explain its differences at this point, it has less influence over the shortlist and less chance of shaping the final preference.
Action
Create three dedicated comparison pages: AttributeIQ vs Dreamdata, AttributeIQ vs HockeyStack and AttributeIQ vs Ruler Analytics. Cover pricing, integrations, attribution capabilities, methodology, ideal customer and the situations where each product is the better fit.
👤 Marketing ⏱ Week 3
3. Limited Visibility Around More Affordable Alternatives: Finding 10
What we found
AttributeIQ rarely appeared in prompts where affordability was part of the buying requirement, including searches framed around alternatives to established vendors.
Commercial risk
Price-sensitive buyers are actively looking for credible alternatives, but AttributeIQ is not consistently entering those conversations. That creates a missed opportunity to compete for buyers who may already see the value of the category but need a more accessible option.
Action
Expand the existing comparison content to cover the broader searches buyers use around alternatives, pricing and specific use cases. Include clear language around “Dreamdata alternative,” affordability and relevant SaaS use cases, supported by specific product and pricing information.
👤 Marketing ⏱ Weeks 1-3 (bundled with comparison pages)
4. Product Documentation Not Reliably Accessible to AI: Finding 13
What we found
AttributeIQ has documentation covering its methodology, but AI platforms were not consistently able to retrieve and use that information when answering questions about pipeline attribution.
Commercial risk
For a product built around attribution, buyers need to understand how the underlying methodology works. A vague or incomplete answer at this stage can make it harder for AttributeIQ to establish technical confidence during evaluation.
Action
Review the documentation pages covering pipeline and multi-touch attribution, remove any access or rendering issues affecting retrieval, and provide the core methodology in straightforward HTML that AI platforms can reliably access. Update the sitemap once the changes are live.
👤 Engineering ⏱ Weeks 1-2
5. Limited Third-Party Evidence: Findings 8, 15
What we found
AI had limited independent information to work with when we tested prompts around AttributeIQ reviews and its G2 rating. The HubSpot Marketplace listing also does not provide a complete picture of the product.
Commercial risk
Independent evidence becomes increasingly important when a buyer is deciding whether a vendor is credible enough to evaluate further. At present, there is less external proof available for AttributeIQ than there should be.
Action
Bring the G2 and HubSpot Marketplace profiles fully up to date, including product information, integrations, pricing and company details. Put a repeatable customer-review process in place and aim to build a consistent base of recent, verified reviews over the next 60 days.
👤 Marketing / Partnerships ⏱ Weeks 3-8
6. Weak Category and Buyer-Fit Positioning: Findings 1, 9, 11, 14
What we found
AttributeIQ was rarely surfaced in broader prompts such as “best attribution software” or “best tool for proving marketing ROI”, and there was limited evidence connecting the product to specific SaaS company stages.
Commercial risk
These searches can introduce AttributeIQ before a buyer has a shortlist. If the product is not clearly associated with the category or a buyer’s specific situation, competitors have a better chance of becoming the default options considered.
Action
Create a dedicated category page explaining where AttributeIQ fits within marketing attribution software, then add specific buyer-fit sections for SaaS companies at different stages. Support these with customer examples, use cases and clear product capabilities.
👤 Marketing ⏱ Weeks 4-8
7. Vague Answers on Core Capability: Finding 7
What we found
AI recognised AttributeIQ as relevant to multi-touch attribution, but struggled to explain how the product actually handles it or what makes its approach different.
Commercial risk
For buyers evaluating attribution platforms, this leaves an important product question unanswered and gives better-documented competitors more room to shape the decision.
Action
Create a dedicated multi-touch attribution page that clearly explains AttributeIQ’s methodology, models, data inputs and outputs. Link to it from the product and documentation pages so AI systems can consistently find and reference the same information.
👤 Product / Marketing ⏱ Weeks 4-6
8. Limited Differentiation Against Existing Tools: Finding 12
What we found
When AttributeIQ was compared with HubSpot’s attribution capabilities, AI had far more information available to explain what HubSpot can already do than why a buyer might choose a dedicated platform.
Commercial risk
For teams already using HubSpot, this can make the existing platform feel like the obvious starting point, even when their attribution requirements go beyond what they currently have.
Action
Create a clear AttributeIQ vs HubSpot attribution comparison covering capabilities, reporting depth, attribution models, pipeline visibility and where a dedicated platform becomes useful.
👤 Engineering ⏱ Weeks 1-2
Measuring AI Search Impact on Customers and ARR
The next 90 days should show whether the recommended changes are turning better AI visibility into more qualified buyer opportunities, customers and ARR. We’ll track both the AI results and the commercial signals behind them.
| Metric | Current | 90-Day Target | How We Measure |
|---|---|---|---|
| Correct pricing stated by AI tools | 0 of 4 tools | 4 of 4 tools | Monthly re-test of Prompt 5 |
| Comparison prompts (2, 3, 4) with AttributeIQ favourably represented | 0 of 3 | 3 of 3 | Monthly re-test |
| Overall prompt score (favourable/correct out of 15) | 2 of 15 | 10+ of 15 | Monthly re-test of full prompt set |
| G2 review count | Below category minimum | Category-competitive volume | G2 dashboard |
| AttributeIQ appearing in “alternative to Dreamdata” prompts | Rare | Consistent | Monthly re-test |
| Additional customers influenced by AI search | Baseline to establish | 2+ per quarter | CRM + attribution review |
| AI-influenced ARR | Baseline to establish | £3,576–£7,176+ per quarter | CRM + subscription data |
If the recommended changes move AttributeIQ from 2 of 15 prompts to 8–10 favourable or accurate results, that creates a larger pool of potential buyers to compete for. At current Pro and Agency pricing of £149–£299/month, two additional customers per quarter would add £3,576–£7,176 in ARR each quarter, excluding expansion revenue.
Want the Same AI Visibility Audit Run on Your Product?
This is the exact AI visibility audit we run for other B2B SaaS companies, we simply used our own product because we could publish the findings honestly. If the gaps above raised questions about your own visibility across AI platforms, that’s a reasonable thing to want answered. Reach out if you’d like us to run one.
FAQs
How much control do SaaS companies have over what AI says about them?
You have more control over the inputs than the final answer. AI systems draw from your website, documentation, reviews, directories and other sources. You can improve the accuracy, consistency and availability of that information, but you cannot directly control how an AI platform weighs those sources or what it ultimately recommends.
Which sources have the biggest influence on how AI represents a SaaS brand?
There is no single source that controls an AI answer. Your website and product documentation provide first-party information, while review sites, marketplaces, directories, comparison pages and competitor content add independent context. The important question is whether these sources agree. Conflicting pricing, positioning or product details can leave AI with the wrong picture.
How do you choose the right prompts for an AI visibility audit?
Start with the questions your sales team hears before a deal closes. Test category searches, competitor comparisons, pricing, alternatives, product-fit questions, core capabilities and trust signals such as reviews. Include both branded and unbranded prompts. The goal is to recreate the decisions buyers make, not simply test whether your company name appears.
Which AI visibility problems are most likely to affect pipeline?
The biggest risks are usually problems that appear during active vendor evaluation: incorrect pricing, missing competitor comparisons, weak category visibility, unclear product capabilities and limited third-party proof.
How do you know whether an AI visibility gap is actually commercially important?
Look at where the gap appears in the buying journey and what happens if it remains unfixed. A missing mention on a broad discovery prompt matters less than incorrect pricing or a weak comparison during vendor selection. Prioritise gaps that can affect shortlist inclusion, perceived fit, or confidence before a sales conversation.
What should SaaS companies prioritise first when AI gets their product wrong?
Start with factual errors that could directly affect a buying decision. Correct pricing first, then address missing or weak competitor comparisons, inaccessible product information and incomplete third-party profiles. Once those are fixed, work on broader category positioning and buyer-fit content. This keeps the programme focused on commercial problems rather than producing content for its own sake.
Can AI visibility be improved without creating a large amount of new content?
Yes. Some of the highest-priority fixes may involve correcting existing information rather than publishing more pages. Update outdated pricing, strengthen existing product documentation, fix access issues, complete marketplace profiles and correct third-party listings.
How should AI visibility improvements be tied to pipeline and revenue?
Set up two tracking lines side by side: one for the specific prompts tied to commercial intent (pricing, comparisons, alternatives), re-tested on a fixed schedule, and one for actual pipeline metrics over the same window, demo bookings, trial signups, opportunities created. Add a source field or intake question asking new leads how they found you, so “ChatGPT” or “AI search” shows up as an option. When both lines move in the same direction across a few cycles, you have a defensible link.
What should a 90-day AI visibility programme realistically aim to change?
The first 90 days should aim to increase the number of high-intent buyer prompts where your brand is present, accurately represented and competitive. From there, measure whether that improvement is followed by more AI-influenced visits, signups, demos and qualified pipeline.
How often should an AI visibility audit be repeated?
For an active programme, monthly testing is enough to track movement without turning the process into constant monitoring. Re-run the same core prompts under comparable conditions, then review the results against the changes made. A broader audit every quarter can identify new gaps as competitors, product positioning, pricing and the wider information landscape change.
Author
Muiz Thomas in
Founder & CMO, GrowUp
Muiz is the founder of GrowUp, a B2B SaaS search marketing agency. His work has driven over £5M in qualified pipeline across construction tech, AI platforms and enterprise software. Outside of work, he’s probably rewatching a match he’s already seen the highlights of, or arguing about something that doesn’t matter nearly as much as he’s making it sound.
