A Climbing Gym Owner’s 30-Day AI Practice (and Why He Put It on Crypto Rails)
A Climbing Gym Owner’s 30-Day AI Practice (and Why He Put It on Crypto Rails)
By Kaori | March 2026
In early March 2026, 香蕉攀岩 founder Qian Xiaolei pulled his core team—scattered across China—into Shanghai for an intensive AI workshop. He didn’t want to wait.
Qian runs 21 climbing gyms across 7 cities, with 200+ employees. By national standards, that’s still a small business. But in China’s climbing industry, it’s the largest chain by store count. Over the past decade, he evolved from an internet media and PR entrepreneur into an offline operator who spends his days on site selection, construction timelines, and performance reviews—the kind of operational gravity that quietly consumes every fast-growing company.
Qian’s point was simple: these repetitive decisions are exactly what AI should compress.
What surprised his team wasn’t just how quickly AI could reduce operational workload—but how naturally their “30-day AI practice” extended into blockchain operations: stablecoin settlement, on-chain reconciliation, wallet security, and even a tokenized loyalty program. In 2025 and 2026, the fastest adopters aren’t always crypto-native startups—sometimes they’re founders in the most physical, offline businesses, using AI to finally make crypto usable at scale.
Why AI + Crypto Became a Practical Stack in 2025–2026
The last cycle taught everyone the same lesson: crypto isn’t only about price—it’s about rails.
- Stablecoins became the most practical bridge between traditional operations and on-chain settlement, even as policymakers focus on managing related risks. The IMF has repeatedly emphasized both the potential benefits and the need for strong safeguards (reference: IMF on stablecoins and payments).
- Regulators also pushed harder on compliance expectations for virtual asset activity, especially around implementation of the Travel Rule (reference: FATF targeted update on virtual assets (2025) and FATF best practices on Travel Rule supervision).
- At the infrastructure layer, mainstream finance accelerated tokenization narratives—often described as a path toward more programmable settlement (reference: BIS on a tokenised “unified ledger” blueprint (Annual Economic Report 2025)).
- Meanwhile, attackers increasingly targeted individuals and businesses, making wallet hygiene non-negotiable (reference: Chainalysis 2025 Crypto Crime Report (PDF)).
AI sits on top of all of this like a force multiplier: it turns policy PDFs into checklists, transaction histories into exception reports, and messy merchant workflows into step-by-step playbooks.
That’s exactly how Qian’s 30-day experiment unfolded.
The 30-Day AI Practice: From Store Ops to On-Chain Ops
Days 1–3: Build an “Ops Brain” Before You Automate Anything
Qian’s team started with a rule: no AI tool shopping until the process map is clear.
They collected the internal “tribal knowledge” that normally lives in group chats and veteran managers’ heads:
- store opening checklist (lease, permits, contractor milestones)
- pricing and promotions logic
- staff scheduling heuristics
- membership disputes and refund policies
- equipment procurement and maintenance cadence
Then they turned it into a searchable internal knowledge base and used AI to standardize outputs:
- “Write the best possible reply to a refund request under Policy X”
- “Generate a weekly store manager report template”
- “Summarize construction delays and propose next actions”
Crypto angle (starting early): they added one more knowledge category—digital asset handling rules—even before they used crypto in production:
- who can create payment addresses
- who can approve transfers
- how to store backup phrases
- what counts as a suspicious request
This is where many teams fail: they treat crypto security like a “later” problem. In 2025–2026, it’s a Day 1 problem.
Days 4–10: Choose One Revenue Workflow Where Stablecoins Actually Help
A climbing gym is not a trading desk. So Qian’s team avoided “accept every coin” thinking and focused on one narrow, realistic use case:
stablecoin payments for:
- membership top-ups (especially from globally mobile customers)
- event registration deposits
- refunds that benefit from fast settlement and clear audit trails
AI’s role here was not to write smart contracts. It was to reduce operational friction:
- generate customer-facing FAQs (“What is a stablecoin?”, “How do refunds work?”)
- create internal scripts for front-desk staff
- draft a lightweight risk policy: limits, escalation paths, and a “do not accept” list for high-risk scenarios
The outcome by Day 10 wasn’t “full crypto adoption.” It was a single pilot with clear boundaries:
- small amounts
- clear refund rules
- a defined treasury sweep schedule
- a documented approval chain
That’s how offline businesses survive change: pilot, measure, expand.
Days 11–17: Reconciliation Becomes the Real “Killer App”
If you’ve ever operated 21 physical locations, you know the real pain isn’t taking payment—it’s reconciling it.
Qian’s team used AI to turn on-chain activity into a daily accounting assistant:
- pull wallet activity
- label transactions by store / cashier / campaign
- flag exceptions (wrong amount, wrong memo, unexpected inflow)
- generate a daily “needs human review” queue
Why it matters: the blockchain is already an immutable ledger. The missing piece for most businesses is interpretation. AI provides that interpretation layer—turning raw transactions into operational truth.
At this stage they also started watching two industry signals that shaped their decisions:
- Account abstraction (smart-wallet UX) is gradually making crypto interactions less error-prone for ordinary users (reference: Ethereum.org on account abstraction and ERC-4337).
- But wallet compromise and social engineering remain a dominant risk, pushing teams toward stronger internal controls rather than “move fast” habits (reference: Chainalysis 2025 Crypto Crime Report (PDF)).
So they treated reconciliation and security as a single system—not two separate projects.
Days 18–24: Treasury and Custody—Design Like You’ll Be Targeted
Once a pilot starts working, funds accumulate. That’s the moment when “wallet setup” becomes corporate infrastructure.
Qian’s model was a simple three-layer structure:
- Front desk / cashier layer: minimal funds, daily sweep
- Operations wallet: enough float for refunds and settlements
- Cold storage: the business treasury, rarely moved
AI helped them write and enforce SOPs:
- role-based permissions (“who can initiate vs approve”)
- address whitelisting and out-of-band verification scripts
- incident runbooks (“if a cashier device is lost, what happens in 15 minutes?”)
- periodic drills (because recovery procedures that aren’t practiced don’t exist)
This is the point where a hardware wallet stops being a “crypto enthusiast gadget” and becomes basic corporate governance for self-custody.
If you’re building this kind of system, OneKey is a reasonable fit to consider because it’s designed around offline private key security and day-to-day usability for holding multi-chain assets—exactly what an operations team needs when the priority is reducing blast radius, not chasing complexity.
Days 25–30: Loyalty, Membership, and the “Tokenized” Idea That Actually Works
The most interesting part of Qian’s 30 days wasn’t payments—it was membership.
Climbing gyms already run on identity and belonging:
- monthly members vs casual visitors
- coaching packages
- competition communities
- referral loops
Instead of launching a speculative token, Qian’s team explored a tokenized loyalty program with strict utility:
- membership credential as a transferable (or non-transferable) on-chain badge, depending on the business rule
- token-gated perks for specific events (early booking windows, guest passes)
- “proof of participation” for competitions and courses
AI’s role again was operational:
- generate perk structures that don’t break margins
- simulate abuse scenarios (“what if badges are resold?”)
- draft clear customer disclosures in plain language
Meanwhile, the macro trend remained in their peripheral vision: tokenization is being discussed not only by crypto communities but also by central banks and market infrastructure institutions (reference: BIS on tokenisation and future financial market infrastructure). For a gym chain, that doesn’t mean “issue securities.” It means: learn how programmable membership might feel, before everyone else does.
The Playbook: If You’re an SMB Founder, Copy This Structure
Here’s the distilled version of Qian’s approach—useful for any operator integrating AI with on-chain treasury management:
- Start with a narrow workflow: one store, one payment type, one settlement rule
- Let AI write the documentation: staff scripts, FAQs, escalation paths
- Treat reconciliation as product: exception queues beat “monthly cleanup”
- Separate duties: initiator ≠ approver, even in a small team
- Assume scams will target you: train staff like it’s part of onboarding
- Move profits to cold storage: keep hot wallets small and boring
- Only then consider loyalty experiments: utility first, speculation never
For compliance-aware teams, keep a close eye on evolving expectations around virtual assets and supervision standards (reference: FATF targeted update on virtual assets (2025)).
Closing: AI Didn’t Replace the Operator—It Made the Operator Faster on Safer Rails
Qian didn’t go to Shanghai to become an AI influencer. He went because offline growth creates a specific kind of stress: too many decisions, too many stores, too many exceptions.
In 30 days, the most valuable outcome wasn’t “using AI.” It was building an operating system where:
- AI compresses repetitive work into checklists and dashboards
- stablecoins reduce settlement friction for specific scenarios
- on-chain records improve auditability
- self-custody and crypto security become standard operating procedure
If your business is starting to accept digital assets, or if you’re already holding them on the balance sheet, it’s worth treating custody like infrastructure. A OneKey hardware wallet can be a practical component of that infrastructure—especially when you want private keys kept offline and operations kept predictable.
In 2025–2026, the competitive edge isn’t adopting every new tool. It’s integrating the right tools into workflows that survive scale.
