Cobo: How are we using AI to drive transformation?

“Many companies talk about AI + Web3. But if they haven’t completed AI transformation internally, they’re just talking concepts externally.”

Author: alexzuo4, Investment & Custody VP @Cobo

Since the end of 2024, besides our core crypto custody and stablecoin payment services, Cobo has been exploring the integration of AI and blockchain.

Our earliest insight was the potential of MCP to bring standardized skills. In theory, if skills are sufficiently standardized, AI can invoke capabilities like plugins, making blockchain the most natural financial infrastructure for AI.

So we internally incubated an MCP application store. But it was quickly disproven.

At that time, AI barriers were still high, only mature engineers could proficiently invoke it. MCP was not standardized enough; each integration was time-consuming, costly, slow, and the implementation far less effective than expected.

But the AI team was established. It was expensive, hard to recruit, and impossible to easily dismantle.

So we decided to change direction. Since we can’t transform our clients’ worlds yet, let’s first transform ourselves.

First Issue: Security

As an asset custody company, Cobo handles extremely sensitive data and internal technical processes. We have strict data hierarchies internally. But without data or real business input, it’s impossible to develop our own Agent.

Initially, we considered deploying local models. But the reality was that local models didn’t meet the intelligence requirements. They could run, but were not user-friendly; they could answer questions, but weren’t smart enough.

In the end, we chose Claude and Gemini as the main models (with the option to apply for ZDR—Zero Data Retention clauses—to achieve the highest level of isolation).

But large models are just the underlying “brain” of the business. The real complexity lies in data and permissions.

We then developed a comprehensive internal knowledge base and Agent framework.

Internal knowledge base + Cobo self-developed agent system

The knowledge base manages internal company data hierarchies, assigning read permissions based on employee roles.

When Agents access the knowledge base, they inherit employee permissions, not a “god view.”

Details include:

• How to isolate network environments
• How to restrict cross-layer data flow
• How to control log retention for auditability
• How to prevent sensitive information leaks

These may not be glamorous, but they determine whether this system can run long-term. AI must not become a security vulnerability.

After the architecture was set up: the problem of no users

Even today, the company still faces a reality: many front-end business units are dismissive of AI.

If we only encourage usage, workflow changes won’t happen.

We later realized that we must start from company management.

First breakthrough: OKR Agent

Our first strongly promoted scenario was not customer service or coding.

It was OKRs.

We used AI to decompose company strategy, help set OKRs, track progress, and review milestones.

In other words, transforming company management from human-driven to a co-governance of silicon and carbon. This process is extremely uncomfortable for employees.

In the past, goals could be written more beautifully, and processes explained more reasonably. Now, weekly data is there, and excuses are fewer.

From that moment, goals are no longer just discussed in meetings but are continuously recorded in the system.

Strategy OKRs weekly monitor business progress

But it was also through performance management that everyone truly became familiar with AI. Because if you don’t participate, it will directly impact your salary.

From performance to business: full Agentization

Once OKRs were running smoothly, we began pushing internal service Agentization. We used competitions + bonuses to enforce each department to establish and operate their own relevant Agents.

Customer service creates customer service Agents. Legal develops contract support Agents. Sales deploys CRM Agents.

Finding the most sarcastic client agent

In total, over 100 Agents were launched.

We can’t precisely quantify the results of “silicon-carbon co-governance.”

But at least one clear change is evident:

Previously, when encountering problems, the first reaction was “should we hire another person?” Now, the first reaction is “can the system participate first?”

This is our understanding of silicon-carbon co-governance. It’s not AI replacing humans, but humans getting used to working alongside systems.

Lessons from this year’s journey

First, maintain healthy cash flow.

If the company’s cash flow isn’t healthy, this transformation can’t reach the end. AI isn’t a cost-saving tool; it’s an upfront investment for long-term structural upgrades. Thanks to Cobo’s main business, we still have healthy cash flow.

Second, top-down push is essential.

Organizations won’t change spontaneously. Without strong management leadership, this effort will naturally fail.

As is well known, Cobo’s founders are heavy AI enthusiasts. CTO Dr. Jiang started AI research during his postdoc at CMU in the early 2000s.

Third, mandatory usage is necessary.

If it’s only encouraged, AI will remain just for writing emails. Real process change must involve some “compulsion.”

Fourth, start with your own business.

Many companies talk about AI + Web3. But if they haven’t completed AI transformation internally, they are just talking concepts externally.

Looking back

We can’t fully quantify this transformation. The company is gradually shifting from “people-driven processes” to “goal-driven systems.”

If a truly “intelligent organization” emerges in the future, it won’t be a natural evolution. It will be pushed out through uncomfortable iterations.

Because with everyone’s participation, the company can better understand the real needs in the AI era.

This is also a byproduct of our internal transformation.

Recently, we launched Cobo Waas Skill. Cobo WaaS Skill is an integrated operational capability layer designed specifically for AI Coding Agents. Through structured knowledge, executable examples, and scenario orchestration, it enables Agents to accurately invoke WaaS APIs. We are upgrading wallet APIs into financial capability modules directly callable by AI Agents. Development cycles are shortened from weeks to dialogue-level.

This isn’t the result of a single product idea. It’s the natural spillover of our internal silicon-carbon co-governance.

We are still exploring.

But at least, today’s Cobo is no longer the same company as in 2024.