AIQuescore

In the digital marketing world, the "Garbage In, Garbage Out" problem is reaching a breaking point. Companies spend billions on user surveys, only to receive low-quality, bot-generated, or mindless responses that skew critical business decisions. Meanwhile, users are often hesitant to share honest opinions due to mounting privacy concerns and opaque reward systems. Quescore is a decentralized platform built to restore this broken trust, ensuring that high-quality insights are met with fair, automated rewards.

By moving the entire survey lifecycle on-chain, Quescore eliminates "middleman bias" and introduces a transparent, merit-based ecosystem. We have integrated a sophisticated AI Quality Scoring system powered by 0G Compute and Sealed Inference. Instead of just checking if a form was filled, our AI evaluates logical consistency, response time distribution, and semantic depth for each answer. Because this occurs within a Sealed Inference environment (TEE), the AI can "verify" the quality of a response without the data being exposed in plaintext to the company or the provider. It is a true privacy-preserving approach to market research.

We’ve prioritized a seamless corporate experience by utilizing ENS (Ethereum Name Service) as a survey identifier. Instead of interacting with complex contract addresses, surveys are identified by human-readable names like nike-q3-survey.eth. This allows companies to manage survey assets and metadata—such as reward pools and deadlines—directly via ENS Text Records.

To ensure the system remains truly autonomous, we leverage KeeperHub for automated reward distribution. Once the deadline is reached, decentralized agents trigger the distribution logic, calculating rewards based on AI quality scores and executing USDC transfers without requiring manual intervention from the brand. Quescore isn't just about collecting data; it’s about incentivizing truth through the synergy of Ethereum, 0G’s compute, and automated execution.

In building Quescore, we adopted a three-layer modular architecture in which each technology works in close coordination to eliminate bottlenecks in decentralised applications. Firstly, in the Entry Layer, we utilise ENS not merely as a username but as a survey identifier; by enabling companies to store metadata—such as reward pools, deadlines and minimum scores—in ENS Text Records, we ensure that the front-end can instantly resolve complex configurations from human-readable names. Next, in the Processing Layer, we have fully implemented 0G technology. User responses are encrypted on the client side and stored in 0G Storage, whilst only the reference CID is recorded on-chain to keep the system lightweight. At the heart of this system, 0G Compute (Sealed Inference) executes the qwen3.6-plus model within a TEE (Trusted Execution Environment), allowing the AI to evaluate logical consistency and lexical diversity to calculate a quality score without exposing the response data in plain text. Finally, as the Execution Layer, a smart contract on Ethereum Sepolia manages the reward logic and verification of the AI’s attestation signature, and we have established a mechanism whereby companies can deposit USDC in a single transaction via the ERC-2612 permit. Furthermore, by integrating KeeperHub via the MCP server, we have automated reward distribution; when the deadline defined in ENS arrives, an autonomous agent automatically executes the payment, enabling rewards to be delivered reliably whilst reducing operational overhead.

As a technical innovation, we adopted a ‘hack’ combining Sealed Inference and TEE attestation as an alternative to zkML, which is difficult to execute due to computational costs. This allowed us to build custom logic within the contract to verify the signature of the execution proof generated at the hardware level, thereby achieving both high-precision scoring via powerful AI and on-chain reliability. Furthermore, by utilising ENS in an unconventional manner as a state management database for surveys, we have enabled flexible switching of survey publication settings on the front end without generating gas-intensive transactions. In this way, by synergising Ethereum’s transparency, zero-gas confidential inference, and KeeperHub’s automated execution, Quescore has become a next-generation market research platform that incentivises truthful responses.