What I automate / Spreadsheets and back-office
The workbook that runs the business
Most Indian businesses run on a spreadsheet that has quietly become critical. It works. It also has a single maintainer, no version history worth the name, formulas nested six deep, and a monthly ritual of cleaning that everyone has stopped questioning.
The risk is rarely the spreadsheet itself — it is that the logic lives in one person's head and one file's formulas. When that person is on leave, or leaves, the business discovers how much was undocumented.
Automating this does not mean throwing the spreadsheet away. It usually means moving the repetitive parts — the cleaning, the matching, the consolidation, the checks — into a job that runs on its own, with a name and a log, while the sheet stays as the thing people look at.
You probably know this is you if
- The same cleaning is done every month: date formats, stray spaces, duplicate rows, wrong codes
- Two lists are matched by eye, or with a VLOOKUP somebody rebuilds each time
- Branch or client files are combined by hand, and one is always missing
- A key workbook has exactly one person who understands it
- An error was found long after it cost something
The jobs in this area
Reconciling two lists by eye
Matched automatically on the rules you already apply in your head; only genuine exceptions reach a person.
Cleaning the same mess every month
Whatever you fix by hand — date formats, stray spaces, duplicate rows, wrong codes — done the same way every time.
Consolidating files from many branches
Files collected, checked for the right shape, combined and totalled, with the one that did not arrive flagged by name.
A workbook only one person understands
The logic moved out of nested formulas into a job with a name and a log, so the business does not stop when that person is on leave.
Errors found only after they cost something
Validation that runs before the data is used, not after: totals that must balance, values that must exist, rows that must be unique.
How a build like this goes
- 1
Write down the rules the person applies without thinking. This is the real work, and it is why I want to watch the job being done rather than be told about it.
- 2
Automate the cleaning first, because it is unambiguous and it makes everything after it more reliable.
- 3
Then the matching, with a deliberate decision about tolerance — what counts as the same transaction when the amounts differ by two rupees and the dates by a day.
- 4
Then validation that runs before the data is used rather than after: totals that must balance, references that must exist, rows that must be unique.
A worked example
Bank statement against the billing system
- Before
- Two days a month, matching line by line, with a tail of unexplained differences carried forward.
- After
- Matched automatically on agreed rules. Only genuine exceptions reach a person — typically a handful rather than hundreds.
Two days a month is roughly 192 hours a year. At a loaded ₹300/hour that is about ₹58,000 a year against a module-band build. It pays back inside a year on hours alone, and the carried-forward differences usually stop accumulating, which is worth more than the hours.
Those are illustrative figures, not a quote and not a promise. Your own numbers are the only ones that matter — the calculator does the same arithmetic on them.
What this normally costs
A job in this area is usually a connected module: ₹60,000 to ₹1,80,000, two to four weeks. The exact number comes from the audit, in writing, before anything is built.
The audit itself is ₹9,999 and comes off the first build in full. If it turns out this is not worth automating for you, that is what the report will say.
Questions about this kind of work
Do we have to stop using Excel?
No, and usually you should not. People know it and trust it. The automation does the repetitive part and can put the result straight back into a sheet.
Our data is genuinely messy. Is that a problem?
It is the normal starting condition, and the demo on this site is built on deliberately messy data for exactly that reason. What matters is whether the mess follows rules — it almost always does, even when nobody has written them down.
What if the rules have exceptions?
Exceptions are fine as long as they are nameable. The automation handles the rule and routes the exception to a person, which is a better outcome than a person handling both and getting bored during the first.
Tell me about your version of this
Tell me the job you are tired of
One call, half an hour. You will get a straight answer about whether it is worth automating — including 'no' if that is the answer.