Checkout in the age of AI
Impact and opportunity predictions for non-developers
Enabling AI development tools to get payment right in the first shot is not just delightful it’s essential.
I recently read a colleague’s social media post regarding the advantages and disadvantages of powerful AI coding tools for the vast majority of non-developers in the world. It triggered a common reaction in the comment thread which was essentially, “AI will give you 95% of what you ask for fast. Settle on that and ship it otherwise you are in for a world of pain”.
After testing a few solutions directly, it is easy to see this wise advice checks out. So, what does that mean for this wide world of non-developers using AI to program payments and checkout?
Asking bolt.new, Copilot, Cursor and others to implement a Stripe checkout or Ethers backed wallet will give you all the scaffolding and UI but leave out the 5% that will require your tool belt to realize a completed transaction:
Creating accounts, verifying identity, generating API Keys or RPC urls, establishing a connection to a destination account for funds…
Further, if we are concerned with future proofing for AI, are we sure AI agents will be able to transact with our payment gateway if and when it is done?
Ok so, now what?
Spendcodes are Payment QR Codes that developers can rely on. The content is a programmatic standard that AI agents can use.
While the use case is still very early it is reasonable to believe:
A non-developer can ask to AI build 95% of a new app or site
And then ask AI to include an npm package for a <Spendcode />
AI then prompts to generate or paste a destination address
Once address is set, agents and users can now scan to send USDC
100% √
Have questions? Want to see how Spendcodes works for you?