When a business decides to automate, the first idea is usually the same: "we need an app." Its own app, with the logo on the customer's phone, so they can order, book or pay from there.

It sounds modern. In practice, it's almost always the most expensive option and the least used. Your customer already has the apps they open every day on their phone, and yours would have to earn a place among them. The useful question isn't "what app should we build?" but where is my customer already, and how do I serve them better there?

This is the list we use to answer it, with an honest rating of how well each idea works for a mid-sized company in Mexico, and real cases from systems we built.

WhatsApp to answer, book and sell: 10/10

It's the channel where your customer already is. There's nothing to download, no password to remember, and the conversation stays in the same place where they talk to their family and their other suppliers.

Automating here doesn't mean a "press 1" menu. It means an AI assistant answers the usual questions (hours, prices, location, availability), books straight into the real calendar and hands off to a person whatever actually needs one.

Real case: CentroMed 360, an urgent care clinic in Puebla, handled families on WhatsApp by hand, and at peak hours messaging became a bottleneck. With its WhatsApp assistant, the system has handled almost 7,000 appointments, resolved more than 4,000 messages with AI and cut waiting time by 90%.

Automatic reminders and follow-ups: 9/10

Much of the money that gets lost isn't from a lack of customers but from a lack of follow-up: the forgotten appointment, the quote nobody mentioned again, the customer who never came back for their next service.

An automatic message at the right moment, on the channel the customer already reads, recovers a lot of that. A day before the appointment. Three days after an unanswered quote. A month after the last service. It doesn't require an app; it requires the system to know what happened with each customer.

A tracking link instead of phone calls: 9/10

When a customer leaves something in your hands (a car at the shop, an order in production, a paperwork process), their question is always the same: "how is mine going?" And every time they ask by phone, someone on your team stops working to answer.

Real case: at Marcas Puebla, an auto repair shop, every service is documented with photos and the customer gets a link to follow their repair from their phone, without installing anything. In the first months there were 3,566 tracking checks made by the customers themselves: questions that used to come in by phone.

AI collection calls: 8/10

For large portfolios, the channel where customers actually answer is still the phone. The problem is that nobody has a team to call thousands of accounts every month.

Real case: Vive la Baja, a residential land developer in Baja California, had a portfolio of tens of thousands of customers that was impossible to follow up by hand. A voice AI has made almost 57,000 calls, held more than 9,000 real conversations and cut collection costs by 70%.

We give it an 8 and not a 10 only because it applies to businesses with volume: if your portfolio is fifty customers, a human call is still better.

A web catalog with orders that land in WhatsApp: 8/10

If you sell products, your customer needs to see what's available, what it costs and order without friction. A web catalog that opens from any link and sends the finished order to WhatsApp covers 80% of what an app would do, without going through the app store and without asking the customer to create an account.

Reports that arrive on their own: 7/10

This one isn't for the customer, it's for you. If every week someone pulls data from several spreadsheets to build the same report, that work can be fully automated.

Real case: at PSP México, the weekly report of hours per project was built by hand. Today the system holds more than 17,700 time entries and the weekly reports are generated and sent on their own.

A web app for repeat customers: 6/10

There are cases where something of your own on the customer's phone does make sense: when they'll use it often, like a membership or a loyalty program. Even then, an app-store app is rarely needed. A web app opens from a link, can be saved to the home screen and updates without the customer doing anything.

Real case: Cafaith Pass is a coffee membership that connects more than 70 independent coffee shops in Puebla and Cholula with their subscribers. It works as a web app, with no app-store download, and has already logged more than 3,000 verified redemptions with a satisfaction score of 4.96/5.

A downloadable app for customers to order: 2/10

Here's the problem. For an ordering app to work, your customer has to find out it exists, look for it, download it, create an account, grant permissions, remember they have it and open it when they want to buy. Every step loses people.

And on the business side, an app-store app means publishing and maintaining two versions (iPhone and Android), passing Apple and Google reviews with every change, and paying for that maintenance even while orders keep coming in through WhatsApp. It's common to see ordering apps with few downloads while the business keeps serving customers by message.

It isn't a 0 because there are cases where it does make sense: when your customer will open it several times a week, when you need phone features the web doesn't handle well, or when you already have a large base of repeat customers. If you're not in one of those cases, start with the channels above.

How to decide what to automate first

Where do your customers message you today? That's the first channel to automate, not a new one.
What question does your team answer most times a day? That's the first automatic answer.
Where does follow-up get lost? Forgotten appointments, unanswered quotes, customers who don't come back. That's where the first reminder goes.
What report do you build by hand every week? That one can generate itself.
Would your customer open your app more than once a week? If the honest answer is no, you don't need one yet.

Automation that works doesn't ask the customer to change their habits. It fits into the ones they already have. If you want to review what's worth automating first in your business, let's talk.