How AI Utilisation Within Blockchain Data Analytics Creates a Paradox.

With Openclaw, Claude Code, GPT Codex as well as MCPs from Dune, Herd, and Allium Labs, 2026 has started on such a high note, and we are still in March.

Every blockchain data analyst or engineer, my self included, is now officially living in a period of time where we know for sure AI systems can write SQL queries (regardless of type) better than we can and faster than we can.

That’s because debugging data issues is way more easier now, data strategies/structures can be executed even cleaner and problem solving is even more straight forward for data engineers. But does that mean anyone can become a data analyst now?

Not even close. In fact, AI has made the fundamentals more important, not less. Joel was quite clear on this topic:

Not just him, Chris also shared a perfect example of why the fundamentals are more right now for any data analyst:

Read that again. Two weeks. One metric. With AI helping. They still had setbacks.

I believe the problem with AI when it comes to data analytics isn't query generation. The problem is knowing what you're actually trying to measure and how to calculate it correctly.

AI can't define business goals. It can't tell you which metrics matter for your specific situation. And it will absolutely hallucinate table names and give you confident answers that are completely wrong if you don't know enough to catch the errors.

How do you know enough to catch or even prevent errors? Know the fundamentals and master them. It’ll also help you when building dashboards too, with AI, verbosity is another problem, you’d, with AI, churn out a 10 chart dashboard when you could have well acheived the same results with just 2-3 charts

Less Is More (But Few People Believe It)

This is one theme that kept coming up throughout our conversation on the episode: the best dashboards are often the smallest ones. Joel shared something brutally honest about his own journey:

What did he replace those metrics with? Revenue analysis. Liquidation data. Supply cap utilisation. Reserve history. The kind of metrics that risk managers and protocol operators actually need to make decisions. Chris gave an example of Alex, a DeFi analyst and Researcher who’s known to niche down on Blockchain Payments; especially Visa and Polygon payments data:

How is that possible? Because Alex's charts answer important questions, and the answers change frequently enough to stay relevant. Three well-designed charts tracking the right things beat fifty charts tracking everything.

The principle is simple: every chart on your dashboard should answer a specific question in your hypothesis that will drives a specific decision. If it doesn't, cut it.

However if you leave this process to AI systems, you’d eventually produce one of the most common words related to AI Data outputs: Slop Data

So if vanity metrics are taken out, what should teams and Analyst focus on instead?

The answer depends on your specific situation, which is exactly the point. There's no universal "good metrics" list that works for every protocol. But there is a universal framework:

For protocols:

• What decision are you trying to make?

• What would change your behavior?

• What leading indicators predict the outcomes you care about?

• What metrics would make you concerned enough to change strategy?

For analysts:

• What questions does this protocol's team actually ask?

• What patterns would surprise them?

• What risks are they blind to?

• What opportunities are they missing?

Joel's work in risk analysis is the perfect example. He's not tracking how many users a lending protocol has. He's tracking concentration risk, liquidation thresholds, reserve ratios, and asset correlation.

These aren't sexy or easy metrics. They don't trend on Twitter. But they're the difference between a protocol that survives volatility and one that gets wrecked. However many protocol. still obsess on the Vanity metrics. Or Metrics/Data Charts that do not move the needle.

And here's what became clear throughout this episode: there's a massive gap between where most Web3 teams are with data and where they could be.

The 80% still chasing vanity metrics aren't stupid. They're just playing a game where surface-level numbers get engagement on Twitter, and that engagement feels like progress.

Breaking out of that pattern requires intentionality and a longer-term view. For aspiring analysts, this gap is an opportunity:

• Learning to ask the right questions,

• Build focused dashboards, and

• Communicate insights clearly puts you in rare company.

Most people can generate queries now thanks to AI. Very few can define what's worth measuring in the first place. For protocols, i think this should be a wake-up call too.

The teams that figure out data-driven decision making while their competitors are still celebrating meaningless milestones on Twitter will have a compounding advantage that becomes impossible to close.

Funny enough, when you look at this issue properly. The data is available on-chain to be queried, we might as well do it the right way and get the right results.

Continue the Conversation

Watch and Listen to the full podcast episode on YouTube.

Watch and Listen to the full podcast episode on Spotify.

If you’re new here, subscribe so you don’t miss out.