If there are few documents and they share the same layout, manual entry may be enough. If documents number in the hundreds, their layouts differ and new ones keep arriving, reading them with AI and turning them into a table makes more sense; but suspicious fields must be checked by a person.
In many companies, information sits not in tables but in folders: invoices, certificates, technical data sheets, contracts. One day this information needs to be gathered into a single table, for an audit, an inventory or a simple question such as "which supplier's document is about to expire?" At that point there are two options: open the documents one by one and enter the data by hand, or have AI read them.
The real cost of manual entry
Manual entry looks simple: open the document, find the relevant field, write it into the table. The difficulty appears as the work repeats. When documents do not share a layout, the information you are looking for is in a different place in each one; one supplier puts the date at the top right, another on the last page. With fatigue come skipped rows, mistyped figures and copying errors. And the work is never done once: as new documents arrive, the table has to be updated again.
What does AI do in this work?
An AI-powered tool reads the documents; whether they are PDFs or scans usually makes no difference. It finds the required fields even in documents with different layouts and writes each document as a row in the table. It flags any field it is not sure about, so it is clear what needs checking. We explain how this kind of tool works on the Document to table page.
The main gain of this approach is consistency more than speed: every document is read by the same rule and written into the same columns. New documents are added to the same table; the table stops being a one-off exercise and becomes a record that stays up to date.
Where is care needed?
Reading documents with AI does not remove checking; it changes where checking happens. The main points to watch:
- Suspicious fields. Illegible scans, handwriting or unusual layouts increase the chance of errors. The tool should flag these fields and a person should check them.
- Sample checks. On the first run, randomly selected documents should be compared with the table one by one; this shows whether the rule has been set up correctly.
- Field definitions. Which date is "date": issue date or validity date? If columns are not clearly defined from the start, both the tool and people will write different things.
- Risk of fabrication. Information not in the document must not be filled in by guesswork; an empty field should stay empty and be flagged. We discuss this risk in the article what happens if AI makes things up.
When does it make sense to keep doing it by hand?
Not every document task is suited to automation. If there are few documents, the job is one-off and all documents share the same layout, manual entry may still be the quickest route. If the time spent on setup and checking exceeds the manual work, automation makes no sense. On the other hand, if documents number in the hundreds, layouts differ and new documents arrive regularly, continuing by hand means spending the same effort again every month.
What to prepare before you start
- Which columns the table will have, and a definition of each column.
- A few samples of each document type.
- Where the documents are kept and where new documents will come from.
- Who will do the checking.
If this data needs to be transferred into a system once the table is ready, we explain how to do that separately in the article bulk-importing an Excel list into a system.
Frequently asked questions
Can scanned (image) PDFs be read too?
Usually yes. The lower the scan quality, the higher the chance of errors; so the tool should flag fields it is not sure about, and a person should check those fields.
Can a table produced with AI be trusted as is?
It should not be trusted without checks. A sample check should be done on the first run, and flagged suspicious fields should be reviewed one by one. Checking does not disappear, but it focuses on the suspicious fields rather than all documents.
After how many documents does automation make sense?
There is no fixed number. The number of documents, how different their layouts are and how often new documents arrive are considered together. If the job is one-off and there are few documents, manual entry may be quicker.
Is there a job that keeps repeating at your company? Tell us about it and we will build the tool that does it.
Tell us what you needOther articles
Which questions can an AI customer assistant answer?
An assistant answers well the questions that can be answered with the information the company gives it: prices, opening hours, address, scope of services, appointments and return terms. Questions that require a decision, negotiation or individual review should be passed to a person without guesswork.
25 September 2026Read →How many days does an automation project take?
Duration is set not by how much code is written but by how organized the data is and who needs to approve, and when. Even a simple task can stretch out as more approvers get involved.
23 September 2026Read →Does a small company need an AI budget?
The budget is not set by the size of the company but by how many hours a specific task wastes. A small company can also start with a single repeating task and a measurable budget.
23 September 2026Read →Who benefits from an appointment system?
An appointment system can help when customers need to reserve a time slot and arranging those slots by phone interrupts your work. First clarify service durations, shared resources and exceptions.
22 September 2026Read →What if AI makes things up?
A fluent answer is not verified information. A business tool needs more than well-written output: it needs a workflow that prevents actions based on incorrect or incomplete information.
22 September 2026Read →Can a spreadsheet with thousands of rows be entered into a system in bulk?
In most cases, yes. The hard part is not the entry itself but getting the data into the shape the system expects and verifying that every row was really saved. If the system offers no bulk route, the work goes through the screen, and that depends on the rules of that system.
21 September 2026Read →Why does preparing a quote take so long?
Writing the quote does not take long; the time goes into hunting for the price, copying and correcting an old quote, and rebuilding the layout every time. Once the price list sits in one place and the rules are written down, a quote comes out in minutes.
21 September 2026Read →Which jobs are suited to automation, and which are not?
The job best suited to automation is one that repeats often, whose rules can be written down and whose data is already digital. Work that needs judgement, is full of exceptions or has its data on paper needs another step first.
18 September 2026Read →What if the AI assistant on your site gives a wrong answer?
The risk is real, but manageable. An assistant's wrong answer usually comes from guessing about something it does not know. The solution is to draw a clear line from the start around what the assistant knows and does not know, and to have it hand over to a person when it does not know.
18 September 2026Read →Is ranking in Google the same as being mentioned by AI?
Partly. Both draw on the same foundation, but they do not reward the same thing: Google ranks a page, while AI reads several sources, writes an answer and decides for itself who to mention in it.
17 September 2026Read →Why doesn't ChatGPT mention my company?
The reason is rarely “the AI does not like us”. Usually it is a technical block, scattered company information, or the company not being mentioned anywhere else online. All three can be fixed.
17 September 2026Read →What should your site have so AI describes you correctly?
When AI describes a company, it uses the clearest sentences it finds on the site. The list below goes page by page through where those sentences should sit, how they should be written and the markup machines will read.
17 September 2026Read →What is GEO? Being visible in AI answers
GEO (Generative Engine Optimization) is the work of making sure that when ChatGPT, Gemini, Perplexity and Google's AI Overviews answer a question, they find your business, describe it correctly and cite it as a source. It does not replace SEO; it builds on top of it.
17 September 2026Read →Will an AI tool learn my company's data?
Short answer: not in a tool running through a commercial API. But that answer depends on how the setup was done, so it is worth knowing what goes where.
16 September 2026Read →When does a spreadsheet stop being enough?
Excel is not a bad tool; it handles most jobs and more. The problem is carrying on with it for months after the point where it stopped being the right fit.
16 September 2026Read →Off-the-shelf software or custom-built?
The right answer depends on how specific the work is to you. Work that everyone does the same way is cheaper off the shelf; work you do your own way is cheaper built for you.
16 September 2026Read →