Table of Contents
Scrunch review
Scrunch has developed into one of the more substantial AI-visibility platforms, combining prompt monitoring and citation research with site diagnostics, agent traffic and infrastructure for serving AI agents directly. Its appeal is strongest for teams that want to act on AI-search findings, while refresh cadence, prompt economics and the governance implications of AXP deserve careful review.
Reviewed August 2026
Product record
Product scope
Scrunch goes considerably further than a conventional LLM rank tracker
Scrunch combines AI-answer monitoring with citation research and technical tools intended to change how a site is discovered and interpreted by AI systems. That gives the product a wider operating range than tools limited to share of voice and prompt rankings.
Prompt tracking extends into source attribution
Scrunch tracks presence, position, sentiment, competitors and citations across monitored prompts. Citation views can be inspected by domain and URL, including the prompts and AI environments associated with a cited source.
Monitoring findings can be translated into page-level changes
Scrunch analyzes website pages and produces recommendations intended to improve AI understanding and retrieval. Teams can move recommendations into their own CMS process or deploy approved changes through AXP.
AI crawler activity can be measured separately from human referrals
Server or CDN traffic can be sent into Scrunch through its custom API, allowing known AI crawlers to be classified from their user agents. This provides evidence that conventional referral analytics cannot capture.
AXP can serve optimized content directly to AI retrieval agents
Scrunch can use CDN-level middleware to deliver server-rendered, AI-oriented content under the same URL while human visitors continue to receive the standard webpage.
AI monitoring
The monitoring product is broad, but long-running prompts do not use a daily standard refresh
Scrunch currently documents monitoring across ChatGPT, Perplexity, Google AI Overviews, Google AI Mode, Gemini, Claude, Microsoft Copilot, Grok and Meta AI. Access to individual environments can vary by package.
Citation research is one of the stronger reasons to use Scrunch
The Citations section lets teams move beyond whether a brand appeared and inspect which domains or individual URLs contributed to the answer. Scrunch also provides an Influence Score based on citation frequency and distribution.
Reputation Insider would treat that score as a product-specific comparative metric rather than an independent measure of source authority. The underlying citation record is more useful than the label attached to it.
Standard cadence becomes slower after the first two weeks
Scrunch's product documentation says newly created prompts are collected daily for their first two weeks. Established prompts then move to a standard 72-hour refresh cycle.
Individual prompts can be manually refreshed on demand without an additional charge.
The standard cadence is adequate for trend monitoring, benchmarking and ongoing AI-search work. A reputation team following a fast-moving incident would need to use manual refreshes selectively or choose a product designed around more frequent automated collection.
Prompt economics
Package allowances can disappear faster than the headline prompt number suggests
Scrunch supports prompt templates across AI platforms, geographies, languages and personas. That flexibility is useful, but buyers need to understand how each monitored variation consumes capacity.
One business question monitored across five AI platforms in three countries produces 15 monitored platform-country combinations before additional languages or personas are introduced.
“350 prompts” should not automatically be read as 350 distinct business questions
Each prompt can expand into several monitored variants. Scrunch documentation also says each AI platform on which a prompt runs consumes a separate unit of the relevant package allowance.
International companies should model their actual monitoring matrix before selecting a plan. A modest number of buyer questions can become a large monitoring estate once competitors, countries, personas and AI platforms are added.
Geography and personas
Country targeting and persona simulation make the dataset more useful, with limits on how literally the results should be interpreted
Geographic monitoring is country-level
Scrunch currently supports dozens of country and territory settings for brand configuration and says individual prompts can be run from any country. City- and state-level geolocation are not currently supported, which limits some local-reputation use cases.
Personas modify the prompt context sent to the model
Scrunch can incorporate a persona description when collecting responses so that the same underlying question can be tested from different professional or consumer perspectives. These results are controlled simulations created by Scrunch rather than observations of real user conversations inside ChatGPT or other AI products.
Agent Experience Platform
AXP gives Scrunch a capability that changes the governance requirements of AI-search optimization
Scrunch's AXP documentation describes CDN-level delivery that can provide AI retrieval agents with server-rendered HTML, structured summaries, entity definitions, clarified claims and approved content enhancements.
Standard webpage
The user continues to receive the normal page and its normal front-end experience.
Agent-oriented representation
Supported AI agents can receive an alternate server-rendered representation under the same URL, including approved contextual additions intended to improve machine interpretation.
Scrunch states that AXP does not redirect Googlebot or Bingbot and characterizes the delivery method as distinct from cloaking. Reputation Insider would focus procurement attention on content governance. If summaries, entity descriptions or clarified claims shown to AI agents differ substantively from the human-facing page, legal, communications and compliance teams need an approval record for both representations. The requirement becomes particularly important for financial services, healthcare, regulated products and listed companies.
Agent traffic and site action
Scrunch connects AI visibility with evidence about crawler behavior and website changes
This is one of the clearest differences between Scrunch and monitoring products that end with a dashboard.
Crawler activity can be separated from AI referral visits
Scrunch can classify known AI agents from server or CDN logs. That helps teams distinguish a human clicking through from an AI answer from a crawler retrieving pages for AI systems.
Recommendations can feed a conventional CMS workflow or AXP
Site Diagnostics identifies page-level changes and can route approved recommendations into implementation. Scrunch also supports recurring application of approved recommendations, which makes ownership and change control particularly important on regulated sites.
Buyer fit
Who should consider Scrunch
Teams that want to connect AI monitoring with changes to the underlying website
SEO, AEO/GEO, digital communications and reputation teams that need prompt-level monitoring, citation research and technical site diagnostics are the strongest fit.
Scrunch becomes more compelling when citation sources matter as much as visibility itself, particularly when a team wants to understand which pages AI systems repeatedly rely on.
Agencies and larger in-house teams can also make good use of country targeting, personas and multi-model tracking if they budget prompt capacity around the full monitoring matrix.
Companies that only need a simple AI visibility score
A company tracking a small number of branded questions can buy simpler AI-monitoring software without paying for AXP, site diagnostics and agent-traffic infrastructure.
Scrunch is also a weaker fit for reputation teams that require fully automated high-frequency monitoring during a fast-moving crisis, since established prompts normally refresh every 72 hours.
Local organizations requiring city-level or state-level prompt testing should verify alternatives because Scrunch's current geographic monitoring is country-based.
Pricing
The current pricing page is unusually transparent for the category, although Scrunch's own documentation is not fully synchronized
Scrunch currently publishes Starter and Growth prices alongside a custom Enterprise plan. The latest pricing page is the clearest current source for package allowances.
The current annual Starter package includes the core monitoring product with a meaningful prompt allowance.
- 350 custom prompts
- 1,000 industry prompts
- 3 personas
- 5 site audits
- 3 users
Growth approximately doubles the custom prompt allowance and expands several account limits.
- 700 custom prompts
- 2,500 industry prompts
- 5 personas
- 10 site audits
- 5 users
Enterprise pricing is negotiated around larger monitoring estates and additional requirements. Model access, AXP, API requirements and commercial allowances should be documented in the Order Form.
The latest pricing page advertises 350 custom prompts on Starter and a seven-day trial with no credit card required. A separate official FAQ still describes a $250 Core plan with 125 unique prompts and four AI platforms, and says the trial requires a credit card and converts automatically unless cancelled. The newer pricing page appears to reflect the current commercial structure, but Reputation Insider would confirm prompt allowances, AI-platform access and trial conversion terms before purchase.
What the company currently documents
These figures come from Scrunch-owned product and Help Center materials. Commercial availability can vary by subscription.
Pros
- Citation-level research provides substantially more evidence than a simple AI share-of-voice dashboard.
- Agent Traffic and Site Diagnostics connect monitoring with the technical behavior of AI crawlers and the pages they retrieve.
- AXP gives sophisticated teams a direct mechanism for improving the representation delivered to AI retrieval agents.
- Public Starter and Growth pricing makes initial commercial evaluation easier than at many enterprise AI-search vendors.
- Country targeting, personas and multi-model monitoring provide considerable flexibility for structured reputation research.
Cons
- Established prompts normally refresh every 72 hours, limiting the product's usefulness for fully automated crisis monitoring.
- Multi-model and multi-country monitoring can consume package allowances much faster than the headline prompt number suggests.
- Current official pricing and trial documentation contains material inconsistencies.
- City- and state-level geographic testing are not currently supported.
- AXP introduces content-governance requirements that become significant when agent-facing claims differ from the standard webpage.
Contract and data terms
The public legal terms appear older than the current product and deserve particular attention after the Sitecore acquisition
Scrunch's publicly available Terms of Service are dated March 6, 2025 and refer to Excelsa Labs Inc. dba Scrunch AI. Enterprise agreements may replace or modify these website terms.
The public terms permit use of customer Content to develop and improve Scrunch services. Customers can request exclusion from model training by email. Reputation Insider would put that exclusion directly into an enterprise agreement where confidential material may enter the platform.
Current public terms allow Scrunch to delete commercial-account data 30 days after the subscription ends. Teams relying on Scrunch for historical reputation records should establish an export process before termination.
The public terms cap aggregate liability at the greater of fees paid during the preceding 12 months or $1,000, subject to the detailed contractual exclusions and applicable law.
The public terms grant Scrunch permission to use a business customer's name and logo in marketing materials. Organizations with strict publicity controls may want to remove or qualify that permission contractually.
The website terms contain language limiting use of the service to collecting customer-sentiment information. That wording sits awkwardly beside the current product's AI-search optimization, AXP and agent-traffic functionality and suggests the public legal documentation has not fully caught up with product development.
Sitecore acquired Scrunch in June 2026. Buyers should rely on the agreement presented during procurement rather than assuming the older website terms describe every obligation of the current Sitecore-owned product.
Independent user evidence
Early marketplace evidence is favorable, although the independent review corpus remains modest
Scrunch has enough current user feedback to identify recurring product themes, but considerably less external evidence than established monitoring platforms such as Meltwater or Birdeye.
Where users are most positive
Current reviews frequently praise the interface, custom prompt control, citation research, API capability, support and the ability to connect AI visibility with agent-traffic data.
Where criticism becomes useful
Negative comments mention UI inconsistencies, sentiment limitations, integration friction, historical-data depth and gaps in reporting. Some users also describe difficulty understanding which AI models are available at different commercial tiers.
Capterra does not yet provide a meaningful comparison set
Scrunch AI has a Capterra listing, but it does not currently contain a large enough review corpus to support a useful independent rating comparison.
Marketplace identity requires care
G2 also contains an older product named Scrunch associated with influencer marketing. That historical listing is separate from the current Scrunch AI product and its rating should not be used when evaluating today's platform.
Reputation Insider verdict
Scrunch is one of the more convincing specialist platforms for teams that want to move from AI visibility into action
Citation research, prompt control, Agent Traffic and Site Diagnostics give Scrunch considerably more operational depth than a standard LLM visibility tracker. AXP extends that proposition further by allowing approved agent-facing content to be served directly to AI retrieval systems. Sitecore's acquisition gives the product a logical place inside a larger web-content stack.
The reservations are specific. A 72-hour standard refresh limits unattended monitoring during fast-moving reputation events, prompt allowances require careful modeling across markets and AI platforms, and public commercial documentation remains inconsistent. AXP also creates a governance requirement that buyers should address before technical deployment.
