The financial adviser who swapped spreadsheets for WhatsApp.

The facts. An independent financial adviser's practice, working with an average of 20 client families at a time · running on AI since late January 2026, when the first version of the app launched. The AI chat came after that first version, and the onboarding agent is planned · in production, with paying users since early September 2026.

• An experienced financial adviser works with an average of 20 client families at a time, mapping their commitments, assets, expenses and income into a budget they then live by. Edo Ceder, who runs the development studio Bumblebee, built her app with Claude Code and tells the story here.
• Every family used to have a spreadsheet template in Google Drive. She filled it in with them from their bank statements on calls of about two hours, and they kept it up to date by hand after that.
• Families now log their spending by WhatsApp. A typed message like "shopping 500", a voice message or a photo of a receipt goes into the budget, and the app replies with what's left in that category.
• The adviser saves about two hours a week on reviewing families' financial data before meetings. Families are also much happier reporting expenses by WhatsApp than in the old spreadsheet.
• Code does the maths and AI handles only what code can't. A message the app can't read, like "I went shopping yesterday", goes to an AI model, so the numbers stay exact and the AI bill stays small.
• A planned onboarding agent will read a family's bank and card statements and build a first budget with them in a chat. Edo expects it to cut onboarding from about three hours to 30 minutes, and families will be able to do it on their own.
• People pay for it. A paid tier launched in early September 2026, and its users include members of her free community, former clients and other advisers using it with their own clients.

The adviser helps families get control of their money, whether that's paying off debt or making ends meet. Her method starts with four lists, covering what they owe, what they own, what they spend and what they earn, then turns them into a budget and tracks the family against it.

She wanted three things from an app, in this order:

• Help her own clients.
• Give it to other advisers to use with their clients.
• Turn it into a standalone app that anyone could use without an adviser.

Edo thinks this kind of founder makes a better client than most startups.

"There's this type of founders who are experts in their field, and then they want to digitise that expertise."

— Edo Ceder

An app that supports the founder's own business has a better chance of succeeding.

A draft budget from the planned onboarding agent. In this demo the agent spots that eating out averages more than the family wants to spend, suggests a lower target, and updates the budget when they agree. A demo screen with a made-up family and made-up figures, supplied by Edo.

Before: a spreadsheet template that every family had to keep filling in

• Each family got a copy of her spreadsheet template in Google Drive, so she could see what was going on and help.
• She loaded their statements with them by hand. She went through all their bank statements with each family on calls of about two hours, from her own time.
• Families recorded every expense in the spreadsheet, which meant opening the template and finding the right row every time they spent money.
• The adviser checked the sheets herself to see who was over budget and where.
• She did the chasing. Prompts to plan the next week or review the month came from her.

Edo calls the spreadsheet "an overhead and a huge friction" for both sides.

What changed: she wanted families to notice their spending as it happened

The adviser wanted logging a daily expense to be easy, so that families would become more aware of what they spent. Edo offered to build an app that put WhatsApp first, since that was where her clients already were every day.

What Edo built: a WhatsApp channel for every day and a desktop app for the rest

• Each family has the app's WhatsApp number alongside a desktop app, which handles the bigger jobs like uploading statements.
• Statements are uploaded in the desktop app, where they are analysed with a mix of manual work and some AI and turned into the family's budget. The planned onboarding agent in workflow 2 will take this over.
• A family logs one or several expenses in one message. A short typed message, like "shopping 500", goes straight in with no AI involved, and anything else goes to an AI model that works out what they meant, so families don't have to learn a format.
• The app replies after every expense with how much is left in that category, or a warning that the category is over.
• Scheduled messages do the chasing. At the end of each week the app asks the family to plan the next one, and it does the same at the end of the month or when a category passes a set amount. One tap opens the plan to edit.

The AI layer came after the first version. With it, families can write freely, record a voice message or send a photo of a receipt, and the AI recognises the shop and the products and records the expense.

Edo kept the AI's job as small as possible.

"This is financial data, so you don't want AI handling everything."

— Edo Ceder

The code handles every message it can understand and the AI only sees the rest.

Result: a WhatsApp message instead of a spreadsheet row

• Families log an expense, or several at once, in one WhatsApp message, where they used to open a spreadsheet and find the right row.
• Families are much happier reporting expenses by WhatsApp than in the old spreadsheet.
• The adviser saves about two hours a week on reviewing families' financial data before meetings. This is an estimate, not a tracked figure.
• Advisers see a family's numbers more clearly than in the spreadsheet, and can ask the app questions about them before they talk to the family.
• Families pay to use it. The paid tier has been live since early September 2026.

Reality check: an app has to keep people accountable without a person

Any app that takes over part of a human adviser's job raises a question of accountability. A family that has paid to meet their adviser next week tries harder than a family with an app, and Edo raises this point himself.

Edo answers that the app reaches out first. It already sends weekly, monthly and over-budget messages, and the plan is for it to notice patterns and start conversations the way she does, like "it's getting close to the end of the month, and you've already spent this much in this category, want to talk about it?"

Your first step: write down the message you wish clients sent you

Pick one thing your clients send you regularly, like receipts, hours or orders. Write the one-line message format you'd want it in, like "shopping 500", and ask three clients to send it to you that way for a week.

If they keep it up, the format is ready for an app to read. Ten minutes is enough to write the format and send the three messages.

Security is the strongest objection to an app like this. The app holds families' bank statements, loans and spending, and an AI coding agent wrote much of the code.

Take this seriously if you are building anything that holds other people's data, whoever or whatever writes the code. Edo's approach has four parts.

Know the law first. Edo starts with the privacy rules where the app operates, then decides what data needs protecting and how.

Every change is reviewed before it ships. Changes go in as pull requests, two AI code reviewers check them, and each one is tested in its own feature copy of the app. A final regression check runs in a staging copy before anything goes live, and he reviews important changes personally.

The AI parts are checked against a published list of risks. He uses the OWASP list of security risks for AI systems, which covers attacks like a user tricking the app into revealing its instructions.

Each adviser sees only their own families. The database enforces it, and wider access, like the lead adviser seeing every family, is being checked against privacy rules before it is switched on.

"Humans make as many mistakes and as many bugs as AI."

— Edo Ceder

Both need the same tests and reviews.

Put everything you can in code, and give AI what's left. Code is faster, cheaper and always gets the sums right.

"AI participates only when it's really necessary."

— Edo Ceder

Stay in charge of the architecture. No written spec or set of rules lets you hand an AI the critical design decisions in a complicated system, he says, so you have to ask it the right questions.

Building for other people needs a release process.

"If you're building something for yourself, that's fine. You're not accountable to anybody, and you can suffer the bugs."

— Edo Ceder

Once other people use it, every change needs reviewing and testing before it reaches them.

Manage AI like a team.

"Once you work with these things, you're like a development manager. You have employees, and you need to make sure that the employees deliver properly."

— Edo Ceder

Case studies are published to help readers understand how other organisations approached a problem. They are not endorsements of a tool or a vendor, and one company's result is not a forecast of yours.

There's more of this story on the website, free to read. It includes a second workflow, the onboarding agent that builds a family's first budget. Create With members get the full case study, with two more powerful workflows in detail.

Continue reading →