Credit dossiers and files traveling by email: the portal that closes deals (instead of chasing attachments)

The attachment that is always missing

Anyone working with files knows this moment by heart. The file is almost ready. One document is missing — the updated chamber of commerce certificate, the latest payslip, the legible and uncut identity card, the form signed in the right spot. You write to the customer. The customer replies two days later, attaching the wrong document. You write again. The customer sends a crooked photo taken with their phone, illegible. You write again. Meanwhile five days have passed, the file is stalled, and you keep it in your head because no system tells you “this one is still missing one thing”.

This is the real work of those managing a credit dossier, a financing application, a mortgage, a request for a grant or a tender: it’s not the “intelligent” part of the job — the evaluation, the structure of the operation — it’s the hunt for attachments. A job of reminding, chasing, overlapping emails, documents arriving in the wrong format, folders on the computer named “Smith_final_2_THISONE.pdf”. And every day the file remains stalled due to a missing attachment is a day you don’t collect, the customer gets nervous, and the operation risks failing because deadlines expire or something changes.

The point I want to show you is that the time of the preliminary investigation is margin. Not in an abstract sense: if every file remains stalled five days longer than necessary hunting for attachments, you close fewer of them a year, with the same staff. The speed with which you collect the right, complete, legible documents is directly the number of files you manage to bring home. This article is for those making a living from documentary investigations and spending half their time chasing attachments via email. Let’s see how a portal with checklist, upload and status turns the hunt into a process, where AI helps (and where it mustn’t poke its nose), and why continuing to do it via email or on a shared Drive is not just inefficient: it’s a risk you are taking.

The math: how much chasing attachments costs

Let’s put down some numbers, stated as estimates but reproducible on your business, because “time is margin” remains a slogan until you translate it into euros. The cost hides in two items: the hours lost chasing and the days the file remains stalled.

The hours. Every file, between writing to the customer for documents, receiving the wrong attachment, requesting again, renaming and archiving files, checking what’s missing, reminding: let’s conservatively put 3-4 hours of pure collecting and chasing work, spread over the life of the file. If you manage, say, 200 a year, that’s 600-800 hours a year of one person spent not evaluating, but hunting for the attachment. At a fully loaded cost of 25 €/hour, that makes 15,000-20,000 € a year of time you could shift to the part generating value.

Item	Estimate
Collection/chasing hours per file	~3-4 h
Files managed per year	~200
Hours/year hunting for attachments	~600-800 h
Cost of that time	~15,000-20,000 €
Days of stall per file (waiting for documents)	5-10 days
Effect: fewer files closed per year	the true cost

The days. This is the item weighing most and appearing nowhere. If every file remains stalled 5-10 days longer than necessary hunting for attachments, and you have a limited capacity of files manageable in parallel, those dead days are files you do not open, operations you do not close, or that fail because deadlines expire while waiting for a document. It’s not a direct cost: it’s unearned revenue, and it’s almost always bigger than the hours. Reducing collection time doesn’t “save a few hours”: it increases how many files you bring home in the year with the same staff. That’s why, in this profession, the time of the preliminary investigation is literally margin.

Why email and the shared Drive are the problem

Let’s pause on the two tools with which today, almost always, documents are collected: email and the shared folder. They seem convenient. They are the root of the problem, and in the case of sensitive documents like credit ones, they are also a serious risk.

Email doesn’t keep the status. A file is made of ten, twenty documents arriving at different times. In an email inbox, they are twenty scattered messages, with attachments all named “document.pdf”, to be searched, downloaded, renamed, put in a folder. There is nowhere telling you, at a glance, “of this file I have 15 documents out of 18, these three are missing”. The status lives in your head and in a mental check-list you lose on Friday evening. Multiply by fifty open files and you have the structural chaos of the profession.

Email makes the wrong documents arrive. When you ask “send me the payslip” via email, the customer sends what they think is the payslip: maybe the annual certification, maybe an old one, maybe an illegible photo. There is no check at the time of upload, so you notice it later, when you open it, and the round of requests starts again. Every round is lost days.

The shared Drive with credit documents is negligence. Here I must be clear, without playing the lawyer. The documents of a credit file — payslips, identity documents, chamber of commerce certificates, income data, sometimes health data — are among the most sensitive in existence. Putting them in a folder shared via link, where anyone with the link enters, where you don’t know who saw what, where files stay for years without anyone deleting them, where an employee leaving takes them away: it’s not “a cheap solution”, it’s exposing yourself to serious damage. If tomorrow a customer asks you “who had access to my documents?”, with a shared Drive you don’t have an answer. And “I don’t know” is the answer you cannot afford to give on such delicate data. It’s not a matter of formal compliance: it’s basic professional diligence. Whoever manages credit dossiers on a shared Drive is running a risk they haven’t evaluated — towards customers, towards lenders, towards themselves.

The theme of the document that must be managed as a flow with a status, and not as an attachment going around, is the core of the whole guide to documents, files and administrative flows: as long as documents “travel” instead of “staying” in a place with rules, chaos is guaranteed.

The portal: checklist, upload, status, expired

Let’s look at the solution, concretely, because it’s simpler than it seems and it changes lives. A portal for documentary investigation flips the logic: instead of chasing documents, you give the customer a place where they bring them, guided, and you see in real time where it stands. Four elements.

The checklist. Every type of file has its list of required documents. The portal knows it: when you open a file, it knows that for that type those twelve documents are needed. The customer logs in and sees their checklist: “these documents are needed, you uploaded 5, 7 are missing, here they are”. They don’t have to figure anything out, they don’t have to guess: the list tells them. And you see the same list from your side, always updated. The mental hunt becomes a progress bar.

Guided upload with check. The customer uploads documents from the portal, and the portal can perform checks on the spot: is the file legible? is it the right type? is it in the accepted format? Not everything can be checked automatically, but even just blocking the empty file, the crooked 3-megabyte photo, the wrong format, eliminates half the rounds of requests. The customer uploads, the portal says “ok” or “this is no good, do it again”, immediately — not two days later when you open it.

The status, for you and for the customer. Every file has a visible status: waiting for documents, under verification, complete, under investigation at the lender, approved, needs integration. You see it for all files together (which are stalled, which are ready, which to remind); the customer sees it for theirs (where it stands, what’s still missing, what they must do). This single element eliminates most of the “where do we stand?” phone calls and your reminders, because the answer is in the portal.

The deadlines. Documents expire (a chamber of commerce certificate has a validity, a payslip must be recent, an offer has a term). The portal knows when a document is about to expire or has expired, and flags it before it becomes a problem — instead of discovering during the resolution phase that the certificate uploaded two months ago is no longer valid and you have to start over.

The upload checklist: what the customer sees

Let’s make concrete what the customer sees when they log in, because it’s the part that makes the difference between a collaborating customer and one who gets lost:

• The list of required documents, with the status of each next to it: to be uploaded / uploaded / verified / to be redone.
• For each document, a short explanation: not “CU”, but “Single Certification: the document your employer gives you at the beginning of the year, the one with last year’s income”. Because the customer doesn’t speak your jargon, and every ambiguity is an extra round of requests.
• A simple upload button, accepting photos from the phone (because the customer will use their phone) and straightening/cleaning them if needed.
• Immediate feedback: “received, looks fine” or “this file is illegible/empty/wrong type, try again”.
• A progress bar: “8 documents out of 12, you’re almost there”. Because the sense of progress pushes the customer to complete.

This guided experience is what turns a confused customer into a customer who uploads everything in one evening. It’s the same UX principle of restricted areas where the experience is the product: if uploading is easy and clear, the customer does it; if it’s confusing, they stop at the third document and you have to chase them.

The UX of the anxious customer

There is an emotional aspect, in this profession, that must be designed and not ignored: the customer of a credit file is anxious. They are asking for a mortgage for the house, a loan for the company, a grant on which a project depends. For them it’s not “a file”: it’s an important, often stressful thing, over which they have no control. And anxiety, if not managed, is discharged onto you in the form of phone calls, worried emails, requests for updates.

A well-made portal also works on this. Showing the customer the status — “your file is under verification”, “we have everything, it has passed to the lender”, “we’re only missing this” — reduces anxiety because it gives them back some control and visibility. They don’t have to call you to know where it stands: they see it. An informed customer is a calmer and less pressing customer, and your day changes. Conversely, silence generates anxiety, and anxiety generates work for you.

Be careful not to confuse this with “flooding the customer with notifications”. The point is not to bombard them: it’s giving them a reliable place where, when worried, they find the answer. A few right communications at the right times (document received, file complete, integration needed), plus an always consultable status. It’s respect for the customer and less work for you, together.

Permissions: customer, office, partner

Here comes the piece the shared Drive cannot do and which is the technical reason why a real portal is needed: permissions. In a credit file there are different actors, and everyone must see exactly what concerns them — no more, no less.

The customer sees their file: their checklist, their documents, their status. They do not see other files, obviously, and they do not see your internal processing notes. They upload and consult, period.

The office (you and your collaborators) sees the files it manages, with permissions you can differentiate: who can only upload, who can verify, who can change status, who sees sensitive data and who doesn’t. A junior collaborator maybe collects documents but doesn’t access evaluations; the manager sees everything.

The partner — the bank, the financial institution, the lending body, the colleague you collaborate with on the file — sees only what they need, when they need it, and only for the files concerning them. You don’t send them a pack of sensitive documents via email: you give them tracked access to the file, with what they must see.

This who-sees-what matrix is not a luxury: on such sensitive documents it’s the minimum of diligence. And it requires a system managing identity and permissions seriously — the same discipline holding up every portal with restricted areas, applied here to particularly delicate data, where getting who-sees-what wrong is not an inconvenience: it’s damage.

The audit: who saw what

Linked to permissions is the audit, i.e. traceability: who uploaded that document, when, who saw it, who downloaded it, who changed the file’s status. In a world of sensitive documents, being able to answer the question “who had access to what?” is not bureaucracy: it’s your protection.

Think about what having an access log means. If a customer contests, you have the documented history of the file. If a collaborator leaves, you know what they saw and you can revoke their access with a click (instead of hoping they haven’t copied the shared folder). If there were ever a problem, you have the traceability of what happened. It is exactly what a shared Drive does not give you: there, once the file is in the folder, you no longer know anything about who touches it. The audit transforms “I trust and hope” into “I know and can prove” — which on credit data is the difference between professionalism and carelessness.

The AI boundary: completeness check, not resolution

We come to the question that always arrives today, and which here is particularly delicate: where is artificial intelligence in a dossier portal? The correct answer is clear-cut, and protecting it is in your interest: AI checks completeness, it doesn’t decide the file.

Where AI helps, and a lot: in reading a newly uploaded document and understanding if it’s the right type (“does this really look like a payslip? is it from the last month? is it legible?”), in flagging when something is missing or when a piece of data doesn’t add up between two documents, in extracting data from documents to pre-fill the file (instead of manually re-typing income, expiration, numbers), in answering the customer on trivial questions (“what document is the CU?”). All this saves hours and reduces rounds of requests — it’s the boring job of completeness checking, and AI does it very well as an assistant.

Where AI must not enter: in the decision. Whether the file is financeable, under what conditions, if the creditworthiness holds up, if the operation makes sense — that is professional evaluation (yours, or the lender’s), with precise responsibilities, and you cannot delegate it to a model that “usually” guesses right. An AI that resolves is a huge legal and ethical risk: it makes mistakes in an unexplainable way, it has no responsibility, and on a credit decision an error is damage to a real person. The rule, here, is the most important of all: AI prepares and checks, the human being decides. AI tells you “the dossier is complete and the data is this”; you make the decision. Whoever sells you “the AI that evaluates files” is selling you a problem with the label of efficiency. This boundary — intelligence where it helps to read and check, human responsibility where decisions are made — is the same applying in every process where AI meets decisions that count.

Integration with the lender

A dossier portal doesn’t live in isolation: downstream there is the lender — the bank, the financial institution, the public body, the fund. And the value increases greatly if the portal connects, where possible, with the lender’s systems: to send the complete file without repackaging everything by hand, to receive back the status (“under evaluation”, “approved”, “needs integration”), to close the circle. Where technical integration is not possible (often, with bodies and banks, systems are closed), the portal still gets a complete, orderly, verified dossier to the lender — which is already half the work, because a well-packaged file is resolved faster and comes back less often for integrations.

The point is that the portal does not replace the lender or your professionalism: it accelerates the collection and verification piece that today eats your time, so you arrive at the part that counts — the evaluation, the relationship with the lender — with the dossier already ready instead of halfway. Fewer days on collection means more closed files in the same time: again, preliminary investigation time equals margin.

A typical case: from hunting attachments to the process

A typical profile, architectural, no names. A business managing financing and grant files, with several files open in parallel and a couple of collaborators. The real work was half evaluation and half chasing: emails to customers for documents, wrong attachments, files on a shared folder with improbable names, and nowhere saying at a glance which files were complete and which stalled. Sensitive documents sat on a Drive with shared links, and to the question “who has access to them?” nobody knew the answer. Files slipped, customers called anxiously, and some operations failed due to expired terms while waiting for an attachment.

What was tackled. A portal with, for every type of file, its document checklist. The customer received an access, logged in, saw what was needed explained in their words, and uploaded — with an immediate check blocking illegible or wrong files. The office saw all files with their status and completion percentage, and knew exactly who to remind. Permissions separated customer, office and partner; every access was tracked. AI helped recognize uploaded documents and extract data to pre-fill, but the evaluation remained human.

Fully operational, the difference wasn’t “we have software”: it was that chasing attachments became a process. Customers uploaded on their own, guided, often in one evening; the office saw at a glance what was missing and reminded accurately instead of from memory; sensitive documents sat in a place with permissions and audit instead of on an open Drive; and files closed faster, because the bottleneck — document collection — had unblocked. More files closed in the same time, with less stress and less risk. The honest note: the portal didn’t speed up the evaluation part (that remains human work), it removed the dead days of collection — which however were most of the lost time.

Timeline, adoption and maintenance

How long does it take? The analysis part, as always, is what counts: mapping the types of files and their document checklists, the permissions among actors, the statuses of the flow. Having done this, the core is built — checklist, guided upload, status, permissions, audit — and tested on a few real cases before opening it to all customers. As always, weeks for a working first core, a few months for a complete system on a business with many file types.

Adoption has two sides. The customer willingly adopts if the portal makes their life easier than sending emails randomly — and usually it does, because a guided checklist is clearly better than “I don’t know what I need”. The office adopts if the portal removes chaos instead of adding bureaucracy: if uploading, verifying and changing status is fast, they use it; if it’s cumbersome, they go back to emails. The rule of every internal tool applies: it must be faster than the method it replaces.

Maintenance has a specificity: files change (new types, new required documents, new lender rules, new regulatory deadlines). The checklist and rules must be updated when the world of files changes, or the portal ages. You need to be able to update file types and required documents simply, and someone who answers when a deeper change is needed.

Where to start

If you recognize yourself in the problem, the first step is not choosing software: it’s putting two things down in black and white, and you can do it this week.

The first is the map of file types. How many types of files do you really manage? For each, what documents are needed? Almost always this list is not fully written anywhere: it lives in the experience of those working on them, with a few “then it depends” and a few exceptions. Writing it down is an already useful exercise in itself, because it brings out inconsistencies (“do we ask for the chamber of commerce certificate for this type of file or not?”) and becomes the basis of the portal’s checklist. If you have many types, start from the two or three most frequent: they cover most of the volume.

The second is the map of who touches the documents. Who uploads, who verifies, who decides, who passes to the lender, which external partners enter. This gives you the permission matrix and makes you see, often for the first time, where today sensitive documents change hands without control — the forwarded email, the file in the open Drive, the USB stick. It’s a small audit of your current process, and it’s usually enlightening.

With these two maps in hand, the conversation with whoever should build the portal becomes concrete: not “I want an upload area”, but “I manage these types of files, with these documents, these actors, these points where I risk today”. From there the right thing is designed, starting from the most frequent file types, and expanding. And if from the math it emerges that your volume is small and the files simple, maybe you discover that email is still enough for you — and telling you so is part of the job. The portal makes sense when the hunt for attachments is truly your bottleneck, not for fashion.

Why a single hand is needed

As with every serious portal, a dossier system is three things together that must stem from the same head. The data: the files, the documents, the types, the statuses, the access log. The backend: the rules — which documents for which file, who sees what, when a document expires, when the file is complete, what AI can do and what not. The interface: the guided experience of the uploading customer and that of the verifying and reminding office.

On such sensitive documents, splitting these three pieces among different suppliers is particularly dangerous. The web agency making “the cute upload area” without truly managing permissions and audit gives you a shared Drive with lipstick: it looks like a portal, but underneath it doesn’t know who sees what. The technician doing “the connection” without thinking about the customer experience gives you a system no customer can use. The piece of permissions and audit — what protects you and the customers — is too intertwined with the experience and the data to be halfway delegated. Whoever designs the upload screen and whoever designs who-sees-what must be the same head, or the portal will be beautiful and insecure, or secure and unusable. A single direction is needed, holding together sensitive data, rules and interface — not three suppliers who, when there’s a problem on a confidential document, blame each other while the risk remains yours.

It’s for you if / it’s not for you if

It’s for you if: you manage files or dossiers requiring the collection of many documents from customers (credit, mortgages, financing, grants, tenders, complex files); you spend much of your time chasing attachments via email and chasing wrong or illegible documents; you manage sensitive documents and realize email and shared Drive are not up to the task; you have anxious customers bombarding you with update requests; the collection time is your true bottleneck.

It’s not for you if: you manage very few very simple files, with two documents crossed, where email is truly enough; you don’t handle sensitive documents and don’t have a volume problem; you are not willing to structure your files into types and checklists — because a portal requires putting order in the process, and if you prefer improvisation case by case, the portal won’t fit you.

Frequently asked questions

How is it better than a shared folder? In everything that counts on sensitive documents: it knows who sees what (permissions), tracks every access (audit), guides the customer with a checklist instead of letting them guess, checks files upon upload, and shows the status of every file. A shared Drive does none of this, and on credit data that absence is a risk, not a saving.

Is the shared Drive really so problematic? Yes, for credit documents. You don’t know who has access, you don’t track who sees or downloads, files stay forever, an employee who leaves takes them away, and to the question “who had access to the customer’s data?” you have no answer. It’s not a question of formal compliance: it’s basic diligence towards customers and towards yourself. On such sensitive data, “I trust and hope” is not a strategy.

Can AI evaluate files in my place? No, and beware of whoever proposes it. AI checks completeness, recognizes and extracts data from documents, flags what’s missing — precious work saving you hours. But the decision (financeable or not, under what conditions) is professional evaluation with precise responsibilities, and is not delegated to a model. AI prepares and checks, the human being decides.

Will the customer be able to use it? If it’s guided, yes — often better than via email. The checklist tells them exactly what’s needed, explained in their words; the upload accepts photos from the phone; the feedback is immediate. It’s easier than guessing what to send you and waiting for your answer. The progress bar pushes them to complete.

Does it connect with the bank or the lending body? Where their systems allow it, yes, and it closes the circle (file sending, status return). Where it’s not possible — often, with closed systems — the portal still gets a complete, orderly and verified dossier to the lender, which is resolved faster and comes back less for integrations.

How much time is really saved? The big item is the collection days per file: if every dossier remains stalled five days less hunting for attachments, you close more files a year with the same staff. The preliminary investigation time is margin. Add the saved update phone calls and the calmer customers.

Do documents remain safe and under control? Yes: permissions to decide who sees what, audit to know who did what, and the possibility to manage the conservation and deletion of documents according to rules, instead of leaving them forever in a folder. It’s the level of control such sensitive data require.

What if the required documents or rules change? Checklists and file types are updated: when the lender changes the required documents or a new file type is born, the list is modified without redoing the portal. It’s part of the maintenance, and must be foreseen, because the world of files changes continuously.

How much does such a portal cost? It depends on how many file types you manage and how articulated permissions and integrations are. It’s not an archive for a few hundred euros, but neither is it a banking management system. The right way to evaluate it is to compare it with the hidden costs you remove — the 600-800 hours a year chasing attachments and, above all, the extra files you close freeing the dead days of collection. On that comparison, a well-made portal pays for itself in a horizon of months, and remains your asset: code, data and rules remain your property, without vendor lock-in.

Do data and portal remain mine? Yes, and on such sensitive documents it’s essential. Code, customer data, file history and access log are yours and remain exportable: if tomorrow you change who maintains the system, you take everything with you, without being held hostage by a supplier precisely on the most delicate data you have.

In one line

If you make a living from documentary investigations and spend half your time chasing attachments via email — with sensitive documents parked on a shared Drive that doesn’t know who sees what — you are losing margin (collection time is margin) and taking a risk you haven’t evaluated. A credit dossier portal transforms the hunt into a process: a checklist guiding the customer, upload with check, status visible to everyone, permissions and audit on delicate data, and AI checking completeness while the decision remains human. Done this way, files close sooner, customers are calmer, and documents stay where they should stay.

If you want to understand what a portal on your files and your documents would look like, look at the projects I’ve built or drop me a couple of lines: we start from your true file types and where collection gets stuck today, not from a demo.