AI and MCP

Categorise all Uber rides at once with an AI assistant

How to find every Uber ride in your records and categorise them in one go with an AI assistant, plus a rule so future rides sort themselves.

Editorial illustration for Categorise all Uber rides at once with an AI assistant

To categorise every Uber ride at once, search your transactions for the merchant, check that the matches really are rides, then apply one category to all of them. If your expense app is connected to an AI assistant and the scope allows changes, you can ask for it in a sentence. You can also make the fix permanent with a rule, so next month's rides land in the right place by themselves.

Why do rides end up in the wrong category?

Statements show merchant text, not your intention. A ride might appear as an Uber trip line, a payment-processor line, or a wallet top-up. Automatic categorisation guesses from the merchant and your history, and a guess may land in 'Uncategorised' or in something vague like 'Shopping'. One bulk fix beats opening fifty rows.

Step by step

  1. Search the merchant name and look at all the variants that come up.
  2. Remove anything that is not a ride: a food order from the same brand, a gift card, a refund.
  3. Pick the category, for example Travel, and a sub-category if you use one.
  4. Apply it to the whole set, by hand or by asking the assistant.
  5. Make a rule so future rides follow, then re-search to confirm nothing was missed.

Worked example

Say a search returns twelve rides over the quarter, average ₹380, so 12 × 380 = ₹4,560. Nine of them, ₹3,300, are 'Uncategorised' and three, ₹1,260, sit in 'Shopping'. Check: 3,300 + 1,260 = ₹4,560. After the bulk change, the Travel category should rise by exactly ₹4,560, Shopping should fall by ₹1,260 and Uncategorised by ₹3,300. Those three movements are your proof.

BeforeRidesAmount
Uncategorised9₹3,300
Shopping3₹1,260
After: all in Travel12₹4,560

What should you check before confirming?

  • Total count of rides matches what the search showed.
  • No refunds or reversals were flipped into the wrong direction.
  • Rides that were reimbursed by an employer are marked the way you track them.
  • Foreign-currency rides keep their own currency.

Should you use an assistant or a rule?

An assistant is good for a one-off clean-up of history. A rule is better for the future. Categorisation runs in stages: your own rules first, then your history and similar merchants, and only after that an AI model, if AI is switched on. SpendixAI categorises in that order, so a rule is the cheapest and most predictable route. Corrections you make also become history, and you can add merchant aliases when one brand appears under several names.

What if you use rides for business?

Create a separate category or sub-category such as Client travel so you can report on it. Categories can carry settings such as a tax-deductible flag and an accounting code, but whether any ride qualifies for anything is for your CA to say. Your job is to keep the records clean and keep receipts attached.

What can go wrong with a bulk change?

The usual slip is a search that is too loose. Searching for a short word can pull in a restaurant or a shop with the same letters in its name. Read the matched merchant text on every row before you apply anything. The second slip is a date range that quietly cuts off the first or last week of a quarter, so set the range yourself instead of accepting a default. The third is forgetting reversals: a cancelled ride that was refunded shows up as a credit, and moving it into Travel as if it were an expense will overstate the total. If you are unsure about a row, leave it out of the bulk change and fix it by hand.

Next time

After a month, filter by Travel and sort by amount. A ride that is three times the usual fare deserves a look. The rule handled the routine; your eyes handle the exceptions.

SPENDIXAI JOURNAL

Useful ideas start with a clearer picture of your own spending.

Explore your reports