AI CFO for DTC brands: GST reconciliation, multi-channel revenue rec, and the month-end close that no longer takes a week

If you run a DTC brand doing meaningful volume across multiple marketplaces, you already know the month-end accounting problem. Every platform settles at different times, in different formats, with different fee structures, deductions, and return handling. Reconciling these against your GST returns is a multi-day manual exercise that your controller hates.

Here’s the honest picture: a ₹50 crore DTC brand selling across 4–5 channels typically has 3–5 accounting staff spending 3–5 days every month just on the reconciliation and GST compliance portion of the close. That’s 150–300 person-hours per month on work that is fundamentally pattern-matching and data transformation.

AI is very good at pattern-matching and data transformation.

The DTC reconciliation problem — channel by channel

The AI CFO stack for DTC — five workflows

Workflow 1: Automated settlement file parser

Each platform exports settlements in different formats: Amazon uses TSV files, Myntra uses Excel with pivot-style summaries, Flipkart uses CSVs with subtly different column headers across versions. A human accountant has to map these manually every month.

The AI workflow:

1. Receives all settlement files via email or a shared folder
2. Parses each format (the AI learns the format pattern from a few examples)
3. Normalises to a standard ledger format: date, amount, type (revenue/fee/return), platform, order ID, GSTIN of counterparty
4. Loads directly into Zoho Books / Tally / SAP B1 via API

Time saving: the parsing and import step goes from 2–3 days to 2–3 hours. The AI catches format changes (when Amazon updates their settlement file structure) and flags them for human review rather than silently mismapping.

Workflow 2: GSTR-1 vs. platform settlement reconciliation

Your GSTR-1 shows the invoices you raised. Your platform settlement shows what they paid. These should match, adjusted for timing, returns, and fee deductions — but in practice, discrepancies accumulate every month.

The AI workflow cross-references:

• GSTR-1 invoice data (extracted from your GST portal or GSTN API)
• Platform settlement data (normalised from Workflow 1)
• Returns and credit note register

It surfaces three categories of discrepancies: timing differences (invoice in period A, settled in period B — expected), amount differences (invoice value vs. net settlement after fees — needs reconciliation item), and missing matches (invoice with no corresponding settlement — needs investigation).

Workflow 3: GST filing preparation

Multi-channel DTC brands have a complex GST position:

• Marketplace sales where TCS (Tax Collected at Source) is deducted by the operator (Amazon, Flipkart)
• Direct sales where you collect and remit GST yourself
• Inter-state vs. intra-state sales with different IGST/CGST/SGST treatment
• Returns that generate credit notes which must appear in GSTR-1
• Input tax credit on purchases and operational expenses

The AI workflow prepares a consolidated GST position report every month: GSTR-1 entries across all channels, TCS summary by operator, ITC claimed vs. available, net liability by state, and the GSTR-3B pre-fill. Your CA reviews and files. The AI handles the compilation; the CA handles the final judgment.

Workflow 4: Channel P&L by SKU

Most DTC finance teams can tell you their total revenue. Very few can tell you the contribution margin by SKU by channel in real-time — because assembling that view requires normalising ad spend allocation (which most brands track in yet another format), platform commission rates (which vary by category and change without notice), return rates (which differ by channel), and COGS per variant.

The AI workflow assembles this view monthly. Output: a channel-SKU-level P&L that shows which products actually make money on which platform after all deductions. For most DTC brands, this view changes their marketing allocation decisions within the first month of seeing it.

Workflow 5: Cash flow forecasting

DTC brands often have predictable settlement timing (Amazon settles T+7, Myntra settles T+14) but unpredictable order volume. An AI cash flow model uses:

• Historical settlement patterns by platform
• Current order velocity (from your OMS or Shopify)
• Known outflows (salaries, COGS purchases, ad spend commitments)

To produce a rolling 13-week cash flow forecast updated daily. For DTC brands managing working capital against marketplace payables, this is the visibility that prevents cash crunches.

The month-end close — before and after

What you need to make this work

Three prerequisites for implementing the AI CFO stack in a DTC business:

1. Consistent data exports. You need to establish a standard process for pulling settlement files from each platform at month-end. This sounds obvious — surprisingly many teams pull them ad-hoc, in different formats, with different date ranges.
2. API access to your accounting software. Zoho Books has an excellent API. Tally Prime has TallyPrime Data Access. SAP B1 has full API access. If you’re on a legacy system, the import step will require an intermediate step. We’ve built this for all three.
3. A data room or shared folder. The AI workflow needs somewhere to receive files. We typically use Google Drive or a Zoho WorkDrive folder that the client team dumps files into at month-end. The AI agent monitors the folder and processes automatically.