A small Bangkok accounting firm goes digital, Part 1: it lacks intake, tracking, tiers and extraction — why still not buy it all at once

This whole series follows a small Bangkok accounting firm, step by step, into digital. The scenario comes from our real implementation experience in accounting services — if you run a firm that collects vouchers, chases clients and books by hand, every chapter should show you a corner of your own office. Each chapter says what was built, what was deliberately not built, and what evidence made it safe to move on.

This firm genuinely lacks everything

It is a small accounting firm in Bangkok: one licensed CPA who also runs the firm, two bookkeepers, serving thirty to forty SME clients. Bookkeeping, VAT filings, withholding tax, payroll and social security, annual statements — all done by this small team. The owner knows exactly how things stand today:

• Intake runs on mail, drop-off and LINE. Every month clients send last month’s vouchers by paper mail, walk them in, or photograph them into LINE/Email. The formats vary wildly; missing items, duplicates and blurry photos are routine.
• Tracking lives in Excel and memory. Who has delivered what for which month sits in a spreadsheet and, more often, in a bookkeeper’s head. When someone leaves, the knowledge goes with them.
• Bookkeeping is done by hand. Vouchers arrive, are checked, typed in, posted and filed. In the busy season — VAT deadlines and the annual-report window — everyone works overtime and still falls behind.

But off-the-shelf cloud accounting systems are also hard to buy

The owner has looked at the cloud accounting software on the market. He came away with three conclusions, each of which kills the idea that one purchase settles it:

• They lock in their own accounting logic. Cloud systems bring their own charts of accounts, reports and posting conventions. The firm handles dozens of clients whose historical data follows its existing habits — migrating them means redoing, reconciling, and risking mistakes.
• Per-seat subscriptions don’t add up. Software is priced per user, and the firm has a handful of people paying for collaboration features it never uses; multiply by three to five years and it is real money, without solving how clients actually get their vouchers in.
• They sell software, not a workflow. The vendor hands over the license and stops. Nobody tells him what to build first, how to get clients submitting every month, how to track missing documents, or how to turn the service into transparent tiers.

The owner’s conclusion: digital is the end goal, but it must be built in stages. Not because the firm does not need these capabilities today — it desperately does — but because the buy-it-all-at-once path is expensive, does not match the existing workflow, and nobody has designed how it lands.

Building in stages is not downgrading the goal

Building in stages does not mean:

• treating the firm as if it does not need the full capabilities;
• splitting a system apart just to sell pieces;
• keeping inefficient manual work forever;
• replacing bookkeeping and filing with a website.

It does mean:

• keeping the full direction, but building one step that can land and be verified first;
• changing a limited slice at a time, so bookkeepers and clients can absorb it;
• using real operations to test which steps actually pay off and which are only “feels like it should exist”;
• making the data and experience of each phase the input to the next.

The route this series follows comes from one simple judgment: first make submissions stable (the foundation of everything), then build an internal intake list (make tracking light), then define tiers from real data (make pricing transparent), then build the PWA client portal (make it visible and upgradable), and finally bring AI into extraction and review (save typing hours). Every step must hold up to market testing.

First step: start with the website and document uploads

Before approving any next step, the owner agrees to watch one thing first: make the monthly voucher submission stable.

Submission is the foundation of all digitalization — without a steady, trackable stream of vouchers, the intake list, tier statistics and AI extraction that come later are empty talk. Part 2 explains why the website and the upload entry come first, and how they fix the mess of mail, drop-off and LINE.

---

Boundaries matter: the website, online presentation, basic forms and the voucher upload entry are a normal website project. The intake list, tier management, PWA client portal and OCR/AI extraction in this series are independent system projects with their own scope — they are not part of the standard website package.