Space and Time helps customers decode blockchain data with Microsoft Azure Open AI

AI is transforming modern business. From large language models (LLMs) to predictive analytics, AI technologies run on data. But what happens when that data has been tampered with? False assumptions, faulty analysis, and manipulated outcomes caused by bad data can erase all the gains that AI delivers.

Web3 startup Space and Time, backed by M12, Microsoft’s venture fund, aims to address this problem with a verifiable data indexer. Space and Time’s mission is to make tamperproof blockchain data accessible to users around the world, by indexing and decoding data from major blockchains so users can easily access it to separate off-chain data for deeper analyses.

With its Proof of SQL technology, Space and Time verifies the databases within its warehouse and that the queries running against them haven’t been tampered with. “Uniquely, our indexer is verifiable and powered by zero-knowledge technology, so that analysts and developers know they’re getting comprehensive, accurate, and tamperproof blockchain data,” says Stephen Hilton, Head of Solutions at Space and Time. “Proof of SQL effectively turns any centralized database into a verifiable database.”

Space and Time also aims to improve accessibility to blockchain data by allowing users to find the answers they need using natural language searches. An AI-powered chatbot within the tool is designed to quickly deliver results and generate queries, saving users time. With this next-generation technology, Space and Time will empower customers to find the data they need to train their AI models and ensure that malicious actors have not tampered with the information. “Verifiability has become a critical factor in business operations, especially those that secure financial value,” says Hilton.

Building trust in AI models with verifiable data

In AI, the old adage of “garbage in, garbage out” is fundamental. If the inputs of an AI model are unverified or biased, outputs can be compromised. And poor outputs can have wide-ranging implications, particularly for models being used for financial analysis or compliance purposes. By taking advantage of the industry-leading AI technology and enterprise-grade security of Microsoft Azure OpenAI Service, Space and Time has built a platform that helps customers work with AI responsibly.

“Reliable data is crucial, as AI relies on comprehensive and accurate datasets to train models that identify and predict fraudulent patterns and anomalies. Without this, the effectiveness of AI in monitoring and securing blockchain transactions may be compromised,” says Hilton. “Space and Time is built to verify the integrity not only of the indexed blockchain data, but of the actual analysis as well.”

Microsoft Partner FTI Technology is collaborating closely with Space and Time. The firm’s Blockchain & Digital Assets practice has been evaluating Space and Time for how the technology can support verifiability when experts are analyzing blockchain data. For FTI Technology, accuracy and defensibility of process are paramount. As a team of the world’s foremost specialists in blockchain and digital assets, FTI Technology is exploring use of AI models in two main use cases which Space and Time’s verifiable model and data may support. 

First, FTI Technology is testing how machine learning models can enhance expert-led analysis of blockchain data. Pattern recognition can help identify suspicious behavior, while predictive analytics will assess the risk of future transactions based on historical data. If the model is trained on inaccurate or incomplete blockchain data, it could result in false positives or overlooked suspicious activity. FTI Technology is exploring how to verify the integrity of not only the indexed blockchain data but also the analysis itself using Space and Time’s OpenAI-powered tools.

Making blockchain data accessible 

Requiring users to create complex queries to find and use the data they need from Space and Time’s warehouse went against the goals for the solution. Instead, Space and Time used OpenAI GPT 4.0 to create a chatbot named Houston. The chatbot responds to natural language queries such as “show me the latest block size on Sui” and runs them against the Space and Time data warehouse. Users can automatically publish their queries to dashboards and AI models without writing complex code. Houston can also create SQL or Python scripts, saving analysts time for more value-added work. 

FTI Technology has tested the AI-powered chatbot’s ability to simplify data retrieval and analysis. Initial findings have confirmed that Houston provides the potential to save time by expediting analysis and retrieval of blockchain data. 

To provide its Web3-native data warehouse, Space and Time turned to Microsoft Azure. Beyond Microsoft Azure OpenAI, the company’s architecture is built on Azure. Space and Time uses compute, networking, management and government, security, and Internet of Things (IoT) solutions from Azure, including Azure Storage, Azure Defender, Azure Virtual Machines, and Azure Monitor. “We’re excited to continue collaborating with Microsoft to enhance our AI SQL capabilities,” says Hilton. “Microsoft has been a world-class partner, providing Space and Time with enhanced development support and go-to-market acceleration.”

Space and Time is also listed on the Azure Marketplace, which is how FTI Technology initially engaged the solution. FTI Technology is continuing to examine how Space and Time, through Microsoft Azure, may support development of compliance auditing solutions and overall efficiency in expert-led blockchain data analysis use cases.

Find out more about Space and Time on X (Twitter), YouTube, and LinkedIn.