Jellyfish vs Xponent21
In AI Marketing Agencies, Jellyfish holds position 3 and Xponent21 position 2. They list no industry in common. Their published stacks have no tool in common. Jellyfish quotes on request; Xponent21 publishes a starting price of $10,000.
Side by side
| Fact | Jellyfish | Xponent21 |
|---|---|---|
| HQ | London, United Kingdom | Richmond, United States |
| Founded | 2005 | 2015 |
| Team | 200+ | 11-50 |
| Pricing | Undisclosed | From $10k |
| Starting price | Quotes on request | $10,000 |
| Pricing model | Not published | Project |
| Stack | Pencil (AI asset creation), Share of Model (tracks brand perception in LLMs), Now Next Soon (media plan simulation), J+ Bidding, Social Agents, Switchboard, J+ Scribe | CARL Intelligence, llms.txt, Schema.org markup, WordPress |
| Industries | Ecommerce, Financial Services, SaaS | Home Services, Healthcare |
| AI Marketing Agencies | #3 of 22 | #2 of 22 |
| Third-party ratings | ||
| Best for | Large advertisers that want media, creative and measurement from one global agency and need brand representation inside language models tracked as a reportable metric. | A contractor or mid-market firm that wants to be the name an AI assistant gives when a buyer asks for a recommendation. |
Ratings are published by the sources named and are shown as each source displays them. AI Agency Radar does not combine them into a score of its own, and positions come from the published criteria alone.
Our verdict on Jellyfish
For a large advertiser, Jellyfish is the option here that most convincingly does AI work rather than AI positioning: Share of Model is a reportable metric that did not exist before language models, and seven named AI products inside a four line service menu is a real allocation of engineering rather than a page on a website. The trade is control and cost visibility.
The main caution: Publishes no pricing at any tier, so an engagement cannot be sized before contact.
Our verdict on Xponent21
Xponent21 fits a contractor or mid-market firm that wants to be the answer an AI assistant gives, and wants the work described rather than gestured at.
The main caution: One verified Clutch review at 4.5. The second rating is eleven DesignRush reviews at an unbroken 5.0, which says less than a mixed record on a larger base.