Workflow guide
The daily business cockpit: 13 detectors that tell Indian shop owners what to fix first
Quick answer
Most Indian SMB dashboards count. Few prioritise. Tiram's What's Up engine reads 13 deterministic detectors across cash, inventory, margin, expenses, and GST — then surfaces the top four ranked on a five-tier severity scale. Per-user snooze prevents stale cards from clogging the screen. No ML, no noise, no A/B toggles — same inputs always produce the same output, and you can reason about why each card appeared.
Try the connected workflow
Open the cockpit in Tiram
30-day free trial. No credit card needed. The 13 detectors run from day one — no setup, no configuration, no opaque scoring.
30-day free trial · No card needed
Scan first
The 13 detectors: what each one catches
| Detector | Severity tier(s) | What it catches | Threshold |
|---|---|---|---|
Overdue receivables | Critical / High | Invoices past due with no payment recorded | Critical: >₹1L overdue >60 days · High: >₹25K overdue >30 days |
Out-of-stock, recently-sold | Critical | SKUs that sold in the last 30 days but currently show zero stock | Any qualifying SKU |
Low-stock reorder | High / Medium | SKUs below the configured reorder point with positive sell-through | Below reorder point, recent sales > 0 |
Dead-stock capital | Critical / Medium / Info | Products with no stock movement for >90 days, weighted by capital locked | Critical: >20% of inventory value · Medium: >5% · Info: >1% |
Revenue decline (MoM) | Critical / High | Month-on-month revenue trending down | Critical: <85% of prior month · High: <90% |
Margin decline | Critical / High | Gross margin dropping vs the 3-month average | Critical: 3pp drop · High: 1.5pp drop |
Negative cash flow | Critical | Cash position trending negative over the current window | Net cash flow < 0 over the rolling 30-day window |
Expense spike | High / Medium | Expense category growing significantly MoM | ≥30% MoM (vs floor of ₹1,000) |
Credit-limit near-breach | High | A customer approaching their configured credit limit | Outstanding ≥ 90% of limit |
GST filing due | High / Medium | GSTR-1 or GSTR-3B filing window opening | 3–5 days before each deadline (monthly or QRMP) |
Promote top product | Opportunity | A high-margin, high-velocity SKU underrepresented in current promotion | Top-decile margin × velocity, low recent push |
Re-engage inactive customer | Opportunity | A previously-active customer who hasn't ordered recently | No order >60 days, prior 90-day frequency >1/mo |
High-margin under-exposed | Opportunity | A high-margin SKU with low sales relative to peers in its category | Above-median margin, below-median volume in category |
Workflow step
The 'I open seven tools every morning' problem
An Indian shop owner's morning currently looks like this: open the billing software, glance at yesterday's sales. Open the bank app, check overnight UPI receipts. Open WhatsApp, scan three customer complaints. Open the inventory tab in a spreadsheet, count what's running low. Open the GSTN portal, check filing status. Open the staff WhatsApp group, see who's coming in today. Open a notes app, remember what's been left undone.
Seven surfaces, twenty minutes, and the question 'what's the most important thing to fix today?' is still unanswered. The bottleneck isn't data — it's prioritisation.
Workflow step
Why a cockpit beats a wall of reports
A typical SMB dashboard shows 24 KPIs at once: sales MTD, sales YTD, customers, invoices, dues, top product, top customer, GST collected, inventory value, low-stock count, expenses, margin... You can read all of them and still not know what to act on, because every number lives at the same visual weight.
A cockpit makes one specific assertion: 'these four things matter right now, in this order'. It deliberately drops the other twenty numbers from the top of the screen — they're still available as drill-downs, but they don't compete for your attention at 8 AM. The four-card cap is the most important design constraint of the engine. Anything more becomes noise; anything fewer doesn't cover the realistic distribution of issues a shop has on any given day.
Workflow step
The 13 detectors, grouped
See the comparison table above for the full list with severities and thresholds. Each detector follows the same shape: a deterministic query against the org's data (no ML, no opaque scoring), a tuned threshold expressed as a number or percentage, and a severity tier that determines whether the resulting card surfaces at all (the engine caps at four). The detectors fall into five families:
- Cash — overdue receivables, negative cash flow, credit-limit near-breach.
- Inventory — out-of-stock-but-recently-sold, low-stock reorder, dead-stock capital.
- Margin and revenue — revenue decline (MoM), margin decline, expense spike.
- GST and compliance — GSTR-1 due, GSTR-3B due.
- Opportunity — promote top product, re-engage inactive customer, high-margin under-exposed.
The families exist for diagnosis, not for filtering — there's no 'show me only cash cards' toggle. The point of the cockpit is 'what's the most important thing today, regardless of family'.
Workflow step
Severity ranking and per-user snooze
Five tiers in priority order: critical > high > medium > info > opportunity. A critical card always beats a high card; opportunity cards only surface when there are no actionable issues higher up the stack. This is what stops the cockpit from being a vanity dashboard — the screen genuinely changes shape as the business state changes.
Per-user snooze prevents the same card from appearing every morning for a week if it's been seen and queued. Snoozes carry a fingerprint of the underlying issue (e.g. 'overdue ₹1.2L from Customer ABC') so when the issue genuinely resolves and reoccurs later, it reappears as a fresh card. Dismissals are also per-user — two people sharing a dashboard don't lose visibility because one of them snoozed.
Workflow step
Three worked examples
**Overdue receivables (critical)**. Your B2B customer Anand Hardware has ₹1.4 lakh outstanding, the oldest invoice is 67 days past due, and they've gone quiet on WhatsApp for two weeks. The detector raises a critical card with the customer name, the aged amount, and a one-click action to send a payment reminder.
**Dead-stock capital (medium)**. You stock 800 SKUs. 47 of them haven't moved in 90+ days. Together they account for ₹1.8 lakh of inventory value — 6% of your total. The detector raises a medium card, links to the slow-stock report, and suggests a clearance discount as the most common resolution.
**GST filing due (high)**. It's the 8th of the month; GSTR-1 for last month is due on the 11th. The detector raises a high card three days out, deep-links to Reports → GSTR-1, and shows a count of invoices in the period for which the export would succeed — so you know whether you're done drafting or have validation errors to fix first.
Workflow step
How it composes with the MTD tiles
The cockpit is layered. Above the four What's Up cards sit the dashboard tiles — MTD GST collected, MTD vendor spend, dead-stock ratio, credit default rate — which give a steady-state view of the business. Below the cards sit the regular reports (sales, inventory, GST, P&L, receivables aging). The cards live in the middle because they bridge the two: 'here's what matters today; the tiles tell you the long-term trend; the reports let you drill in'.
The tile order is per-user (saved as a `DashboardLayout`) so two cashiers on the same org can each prioritise the tiles that matter to their role.
Workflow step
What this engine is NOT
It's important to be explicit about what the priority engine is not, because every adjacent product over-claims this surface:
- It's not AI. It's deterministic rules with tuned thresholds. Every card has a comprehensible 'why'.
- It's not predictive. It diagnoses what's already true (overdue, dead, declining), not what might happen. Forecasting needs different math and Tiram doesn't ship it.
- It's not customisable in the UI. Thresholds are tuned by the product team based on observed Indian SMB data. Surfacing every threshold as a setting would turn the cockpit into a configuration project; it stays opinionated on purpose.
- It doesn't replace your CA. GSTR-due cards remind you to file; they don't file for you, and they don't draft GSTR-3B. The boundary is covered in the companion article: GSTR-1 vs GSTR-3B: difference, reconciliation, and how to draft a clean GSTR-1.
Workflow step
If your dashboard counts but doesn't prioritise, you have a counter, not a cockpit
Most SMB dashboards in the Indian market today are counter UIs: 200 numbers, no opinion, no act-on-this-first. They were designed by people who weren't sure what to surface, so they surfaced everything. The cockpit pattern works because it's willing to be wrong about the fourth-most-important issue today — and right about the first one. The cost of being wrong on #4 is a missed opportunity. The cost of having no #1 is a missed bill, a missed reorder, a missed filing, a customer that walked away.
Quick answers
- What dashboard tells a shop owner what to fix first each morning?
- A prioritised cockpit, not a wall of counters. Tiram surfaces the top four issues across receivables, stock, margin, expenses, and GST due dates, ranked on a five-tier severity scale with per-user snooze. Maximum four cards on screen at once.
- How do I detect dead stock automatically in my inventory software?
- A dead-stock detector flags products with no movement for 90+ days, weighted by the capital they lock. In Tiram, dead stock at >20% of inventory value is critical, >5% is medium, >1% is info — surfaced as a cockpit card with a deep-link to the slow-stock report.
- What's the best way to track overdue invoices for an Indian small business?
- Aged-receivables logic with severity tiers, surfaced as a daily card rather than a buried report. Tiram raises a critical card for >₹1 lakh overdue >60 days, a high card for >₹25K overdue >30 days, with a one-click action to send a payment reminder.
- Can billing software remind me when GSTR-1 is due?
- Yes. The cockpit surfaces a GSTR-1 due-date card three to five days before each filing window. The detector knows whether you're a monthly or QRMP filer and shows the right deadline. A separate detector covers GSTR-3B.
- How do I know when my expenses spike in a small retail business?
- An expense-spike detector compares each category's spend to its prior month. Tiram raises a card when a category grows ≥30% MoM (with a ₹1,000 floor to suppress noise on tiny categories). The card shows the category, the prior and current totals, and the % change.
FAQs
- Is the priority engine AI-powered?
- No. The detectors are deterministic rules with tuned thresholds. The same inputs always produce the same output, and you can reason about why a card appeared — which matters when money is on the line. Tiram does not ship ML in the cockpit.
- Can I customise the thresholds?
- Not in the UI today. Thresholds are tuned by the product team based on observed Indian SMB data and adjusted across releases. We deliberately don't surface every threshold as a setting because turning the cockpit into a configuration project would defeat the point. If a threshold consistently misfires for your business, let us know.
- What if I disagree with which card is shown first?
- The severity ordering (critical > high > medium > info > opportunity) is fixed. Within a tier, cards are ordered by impact magnitude (largest overdue rupee value first, biggest dead-stock pile first, etc.). You can snooze any card that's not currently actionable; it returns when the underlying issue genuinely changes.
- Does the cockpit work with multi-warehouse setups?
- Yes. Inventory detectors (out-of-stock, low-stock, dead-stock) read from the per-warehouse stock cache and surface issues per-warehouse, not just at the org level. A single SKU dead in one warehouse and fast-moving in another shows up correctly.
- Can two users see different cockpit states?
- Yes. The priority engine is org-scoped (same data), but snooze and dismissal are per-user (different surfaces). If the cashier dismisses 'low stock on Item X', the owner still sees it. Dashboard tile order is also per-user.
- Does this work for distributors, not just shops?
- Yes. The detectors are calibrated for SMB shape (turnover roughly ₹10 lakh to ₹50 crore) and work across retail, distribution, and multi-branch chains. Distributor-specific signals like credit-limit near-breach leverage the same engine.
- How is this different from Tally's reports or Vyapar's dashboard?
- Tally ships exhaustive reports — you have to know what to look at. Vyapar shows MTD counters. Neither prioritises a small set of acts-today issues. The What's Up engine's value isn't a single feature; it's the editorial discipline of capping at four cards, ranking by severity, and refusing to ship a configurable dashboard.
Related reading
Open the cockpit in Tiram
30-day free trial. No credit card needed. The 13 detectors run from day one — no setup, no configuration, no opaque scoring.
Start free