You built something. It works. People are using it. You're further along than most people who only talk about building an app.
Now you need the grown-up features: a way to actually make money, a way to talk to your users, and a way to know what's actually happening inside your app. AI tools make these integrations far easier than they used to be, but you still need to know what to ask for and how to set each one up correctly.
Part 1: taking money
If you're not taking payments, you don't have a business, you have an expensive hobby.
Stripe is the standard for most apps: 2.9% plus 30 cents per transaction, roughly two to three days to set up properly, and it works for one-time payments, subscriptions, and complex pricing alike. PayPal is the fastest to start, same fee structure, up and running in about 30 minutes, and works because everyone already has an account. Lemon Squeezy runs 5% plus 50 cents per transaction, takes about an hour to set up, and is worth the higher fee for digital products since it handles global tax compliance for you. For most people reading this, start with Stripe.
To add Stripe for a one-time payment, tell the AI: "Add Stripe payment integration to the checkout page. When a user clicks 'Buy Now,' redirect them to a Stripe Checkout session for a one-time payment of $[amount]. After successful payment, redirect them to /success and send a confirmation email. Store the payment status and transaction ID in the database. Use Stripe test mode for now."
For subscriptions: "Add Stripe subscription integration with three tiers: Basic, Pro, and Premium, at [your prices]. Create a pricing page with cards for each tier. When a user selects a plan, create a Stripe Checkout session for that subscription. After successful payment, update the user record with subscription tier and status. Include a webhook endpoint to handle subscription updates and cancellations."
After that, sign up at stripe.com for your own API keys, add them as environment variables, and test with Stripe's own test card (4242 4242 4242 4242) before switching to live keys. Test payments thoroughly in test mode before anyone can pay you for real.
Part 2: talking to your users
Welcome emails, password resets, notifications, receipts, you need all of it.
Resend is the easiest starting point for transactional email: a free tier of 3,000 emails a month, a developer-first API, and about 20 minutes to set up. SendGrid is the older industry standard, more powerful for marketing email, and takes an hour or two to configure properly. Mailgun is developer-friendly with solid documentation and reliable delivery. Start with Resend unless you already know you need SendGrid's marketing features.
For a welcome email: "Add Resend email integration. When a new user signs up, send a welcome email with their name, a brief explanation of what to do next, and a link to the getting-started guide. Use our brand colors in the template. Store the Resend API key in environment variables."
For transactional email: "Set up email notifications for password reset requests with a secure token link, purchase confirmations with order details, and a weekly digest of user activity. Each email should use our brand template and include an unsubscribe link. Log all sent emails in the database for debugging."
A few things that matter more than they sound like they should: an unsubscribe link isn't optional, it's the law. Send from your own domain, not a free gmail address. Keep the first emails simple, since elaborate designs tend to break in email clients. Send a few to yourself first and check the spam folder before you trust it.
Part 3: understanding your users
If you're not tracking what people actually do in your app, you're flying blind on who's using it, where they drop off, and which features get touched at all.
Plausible is the simplest, privacy-focused option: $9 a month for up to 10,000 monthly pageviews, no cookies, GDPR compliant, and about 5 minutes to set up. PostHog is built for deeper product analytics, with a free tier around 1 million events a month, feature flags, and session recordings, at the cost of more complexity. Google Analytics is free and comprehensive but usually overkill for an MVP, and it carries its own privacy considerations in some regions. Use Plausible if you want simple, PostHog if you want to track specific user actions.
For basic page tracking: "Add Plausible analytics to track page views across the app. Set up custom events for button clicks on the pricing page, sign-up completions, and checkout initiations. Store the Plausible domain in environment variables."
For deeper product analytics: "Integrate PostHog to track user behavior: signup, feature used, purchase completed, friend invited. Include user properties like account type and signup date. Store the API key in environment variables."
Worth tracking: sign-ups, page views on key pages, clicks on your important calls to action, which features actually get used, where people abandon checkout, and how they move through onboarding. Not worth tracking, and not worth the privacy risk: exact mouse movements, personal data beyond what you need, every single click, or form input captured character by character.
The order to build these in
Add analytics first, in week one, so you can measure everything else you're about to change. Add email in week two and test it with welcome messages and password resets. Add payments last, in week three, once everything else works, and test extensively before going live.
Where this goes wrong
Hardcoding API keys directly in your code instead of using environment variables. Skipping test mode on payments. Forgetting to handle a failed payment. Leaving out the unsubscribe link. Tracking so much that the data drowns any signal. Not checking that emails actually land in an inbox instead of spam. Using production keys while you're still building.
Keep your keys in your platform's environment variable settings, never in the code itself, and never committed to GitHub. Every platform, Lovable, Bolt, Replit, Vercel, Netlify, has a place for this.
The bottom line
You can't run a real business without these three integrations, and with AI-powered tools, you can add all three in under a week. Once they're set up, they mostly run themselves. Pick your tools, write clear prompts, test everything twice, and ship it. Your MVP just became a real product.