Speak your customers' language

A multilingual app works completely in every language it offers, not just on the home screen. Menus, buttons, error messages, notifications, emails, invoices, WhatsApp updates, help pages and support all follow the language the customer chose.

Translation is part of it, but most of the work is in the build: keeping every piece of text out of the code so it can be translated, designing screens that hold Hindi, Tamil or German without breaking, writing dates and amounts the local way, and making search, voice input and support understand the same languages as the screens.

Done from the start, adding a language later is mostly new text. Done as an afterthought, every new language becomes a project of its own.

People decide faster and trust more when they read in the language they think in. Many customers in India are comfortable with English for a price or a button, but want Hindi, Gujarati, Marathi or Tamil when something matters: a medical instruction, the terms of a loan, a problem with a payment.

Language also widens who can use your product at all. Farmers, drivers, shopkeepers and older customers who avoid English apps will use one that speaks to them, especially if they can talk to it instead of typing.

It cuts support work too. Customers who understand the screen don’t need to call and ask what it means.

For lenders, clinics and insurers, it is also a matter of fairness. A borrower signing a loan agreement or a patient reading a prescription should understand every word of it.

Start with the languages your customers already use. Your support calls, WhatsApp messages, delivery addresses and sales regions usually make this clear. A business in Gujarat may need Gujarati and Hindi beside English; a transport company in Tamil Nadu, Tamil; a brand selling across India, Hindi first and then the largest regional languages.

We also look at where you plan to grow, and at which parts of the product matter most. Sometimes the customer app needs three languages while the staff dashboard stays in English, or the WhatsApp messages need more languages than the app.

The first languages don’t limit you. Built properly, a fifth language is no harder to add than the second.

From your side, we need three things: whatever you know about your customers’ languages, a list of the words that make up your brand and products, and someone for each language, inside your team or outside it, who can say whether a screen sounds right.

Every language changes the shape of a screen. Some run far longer than English. Scripts such as Devanagari, Gujarati and Tamil need more line height for the marks above and below their letters, and languages such as Urdu run right to left.

We design every screen for its longest version from the first sketch: buttons that grow, labels that wrap cleanly and no text baked into images. Screens are tested with real translations, never English stand-ins, on low-cost phones as well as new ones.

Fonts matter as much as words. We choose fonts that cover every script you need and look consistent side by side, and ship them with the app, so text appears the same on every phone.

Good translation reads as if it was written in that language. We keep all text in one place, with a note on where each line appears, so translators see the context rather than a list of fragments.

AI produces a quick first draft for every language, following a glossary of your terms: product names, legal phrases and words that should never be translated, such as UPI or your brand name. A fluent reviewer, from your team or arranged for you, then approves each language before release. Medical, financial and legal wording gets a specialist review.

When the English text changes, the matching translations are flagged for update, so the languages never drift apart.

Many people speak more easily than they type, especially in their own language on a small keyboard. Voice input lets them say what they want and see it written down for them.

Speech recognition for Indian languages is now good enough to fill in real forms. In an agritech app we built near Nashik, farmers list their produce by speaking Marathi; the app turns their words into the crop, the quantity and the harvest date, then reads it back for them to confirm. Services such as Bhashini, Google Gemini and OpenAI Whisper make this possible across many languages.

Voice works best for short, structured input with a clear confirmation step, and we design it that way. Background noise in a field or a market, and the accents and dialects of the people who will use the app, are part of the test: we try speech services on real recordings from those places before choosing one.

Search has to match the way people actually type. A customer may write a product name in the local script, in English letters or in a mix, with spellings that vary from person to person. We set up search to treat these as the same, so the customer finds the product however they write it.

Support follows the same language. Chat, WhatsApp and AI assistants reply in the language the customer wrote in, including mixed Hindi and English, and hand over to staff who speak it.

A customer’s language has to reach beyond the app. The order confirmation on WhatsApp, the SMS with a one-time password, the invoice PDF, the email receipt and the push notification are often what customers read most, and they are the easiest to forget.

Each of these goes out in the language the customer chose. WhatsApp templates are written and approved in each language. PDFs such as invoices and prescriptions use fonts that carry every script, so Hindi or Tamil prints as clearly as English. Push notifications are kept short enough to fit on a lock screen in every language.

Staff screens can stay in one language while showing customer-facing text in the customer’s. A support agent who reads English can still see exactly which message a Gujarati-speaking customer received.

Language comes with local habits. In India, amounts are grouped in lakhs and crores; in much of Europe, the comma and the full stop swap places in numbers. Dates are written in different orders, names are structured differently, and Indian addresses lean on landmarks and pincodes.

We handle these for each language and region, so prices, dates, phone numbers and addresses look right to each customer, and documents such as invoices and receipts follow local conventions.

Every language is tested by people who speak it. Before launch, customers or staff fluent in each language go through the main journeys, and we fix anything that reads oddly, breaks a layout or sounds too formal for your brand.

Testing also happens where the app is used. For a transport company in Coimbatore, we built the driver app in Tamil and English with large buttons, and it saves everything on the phone when highways have no signal, then sends it when the network returns.

Automated checks run on every release: no screen with missing translations, no text cut off, and no language quietly falling back to English. After launch, we keep the languages in step with new features, add languages when you are ready, and improve voice and search from what customers actually say and type.