DeFi Protocol AI Visibility: The First Comprehensive Benchmarks
Decentralized finance is one of the most dynamic and competitive sectors in crypto, yet most DeFi teams have no visibility into how AI assistants represent their protocols. This report establishes the first category-level benchmarks for DeFi AI visibility, giving protocol teams a quantitative baseline to measure against. The data covers 120 DeFi protocols across four subcategories — lending, DEX, yield optimization, and stablecoins — tracked across ChatGPT, Claude, Gemini, and Perplexity.
These benchmarks matter because an increasing share of DeFi users begin their research with AI. Our user survey data indicates that 31% of DeFi users under 35 now consult an AI assistant before interacting with a new protocol, up from 12% in 2024. For these users, if your protocol does not appear in AI responses, it effectively does not exist.
Lending Protocol Benchmarks
The lending category shows the starkest visibility gap between established and emerging protocols. Aave dominates AI mentions despite an increasingly competitive market.
| Protocol | ChatGPT Mention Rate | Claude Mention Rate | Gemini Mention Rate | Perplexity Mention Rate | Avg Mention Rate | TVL Rank |
|---|---|---|---|---|---|---|
| Aave | 88% | 82% | 84% | 91% | 86% | #1 |
| Compound | 61% | 54% | 58% | 42% | 54% | #4 |
| MakerDAO / Sky | 47% | 44% | 41% | 52% | 46% | #2 |
| Morpho | 18% | 22% | 14% | 48% | 26% | #3 |
| Euler | 12% | 15% | 9% | 38% | 19% | #6 |
| Radiant | 8% | 6% | 7% | 24% | 11% | #5 |
| Silo | 4% | 3% | 2% | 19% | 7% | #8 |
The most striking finding is the visibility-TVL mismatch. Compound ranks 4th by TVL but 2nd by AI mention rate, while Morpho ranks 3rd by TVL but 4th by AI mentions. This lag reflects AI training data recency: Compound's long history and extensive documentation give it persistent AI visibility that newer, larger competitors have not yet matched. Perplexity narrows the gap significantly — Morpho's mention rate on Perplexity (48%) is more than double its rate on other platforms, because Perplexity pulls current data that reflects Morpho's actual market position.
DEX Protocol Benchmarks
The decentralized exchange category is dominated by Uniswap, but the competitive landscape is more nuanced when examined by chain ecosystem.
| Protocol | ChatGPT | Claude | Gemini | Perplexity | Avg Mention Rate | Primary Chain |
|---|---|---|---|---|---|---|
| Uniswap | 92% | 87% | 89% | 94% | 91% | Ethereum / Multi |
| Curve | 44% | 41% | 38% | 47% | 43% | Ethereum / Multi |
| Jupiter | 32% | 38% | 28% | 61% | 40% | Solana |
| PancakeSwap | 38% | 29% | 34% | 36% | 34% | BNB Chain |
| Raydium | 14% | 18% | 11% | 42% | 21% | Solana |
| Aerodrome | 6% | 9% | 4% | 34% | 13% | Base |
| Trader Joe | 12% | 10% | 8% | 22% | 13% | Avalanche |
When users ask generic DEX questions ("best decentralized exchange"), Uniswap appears in 91% of responses. But when queries are chain-specific ("best DEX on Solana"), Jupiter's mention rate jumps to 78% — significantly higher than its overall average. This highlights an important strategic consideration: chain-specific AI visibility may be more achievable and more valuable for ecosystem-native DEXs than competing for generic DEX queries, which Uniswap dominates.
Yield and Stablecoin Benchmarks
Yield optimization and stablecoin categories round out the DeFi benchmark dataset, each with distinct visibility dynamics.
| Category | Protocol | Avg Mention Rate | Accuracy Score | Key Visibility Gap |
|---|---|---|---|---|
| Yield | Yearn | 52% | 71/100 | AI cites v1 vaults; v2/v3 architecture absent |
| Yield | Convex | 34% | 74/100 | Accurate but rarely recommended to beginners |
| Yield | Pendle | 28% | 68/100 | Yield tokenization mechanics poorly explained |
| Yield | EigenLayer | 41% | 62/100 | Restaking concept frequently mischaracterized |
| Stablecoin | USDC (Circle) | 89% | 92/100 | Minimal — strong entity authority |
| Stablecoin | USDT (Tether) | 84% | 78/100 | Regulatory concerns overstated vs. current status |
| Stablecoin | DAI / USDS | 54% | 69/100 | Rebrand to USDS not reflected; still called DAI |
| Stablecoin | Ethena (USDe) | 22% | 64/100 | Mechanism frequently confused with algorithmic stables |
The yield category has the lowest average accuracy scores in all of DeFi, averaging just 69/100. The core problem is complexity: yield strategies involve composable, multi-protocol interactions that AI models struggle to explain accurately. EigenLayer is a notable case — restaking is a novel primitive that did not exist in most AI training data, and models frequently mischaracterize it as staking or conflate it with liquid staking. For protocols building genuinely novel DeFi primitives, the AI education gap is a first-order business problem.
Stablecoins are the most accurately represented DeFi category, but rebranding creates acute visibility problems. The MakerDAO-to-Sky and DAI-to-USDS transitions are poorly reflected across all platforms except Perplexity, meaning millions of AI users are getting outdated information about one of DeFi's foundational protocols.
Benchmark Methodology
These benchmarks are based on 6,400 AI responses to 800 DeFi-specific prompts across ChatGPT, Claude, Gemini, and Perplexity, collected in February-March 2026. Mention rate represents the percentage of category-relevant queries where a protocol was named. Accuracy scores reflect manual fact-checking of protocol descriptions, TVL figures, feature sets, and security characterizations. TVL rankings are from DeFiLlama as of March 15, 2026. Benchmarks will be updated quarterly. Protocol teams interested in detailed per-protocol reports should contact the Presenc AI research team.