For monthly board reporting in mid-market LATAM companies, there are 28 standard KPIs across 5 dimensions: financial, sales, operations, service, and system. Each one must be calculated in auditable TypeScript code—not in an Excel formula or an AI response (hallucination risk). Real case from Catalizadora: reporting with 5 sections, 28 hardcoded KPIs, AI engine for narrative only, SHA-256 hash chain audit trail.
If you run a mid-market company in LATAM (20 to 300 employees) and build monthly board reports, this is your actionable list.
The 28 Standard Board KPIs for Mid-Market Companies
Distributed across 5 dimensions:
Financial (6 KPIs):
- Monthly revenue vs. budget
- Actual gross margin vs. projected
- Cash position plus burn rate
- Average accounts receivable days
- Overdue receivables ratio (60+ days)
- Invoiced but uncollected revenue
Sales (6 KPIs):
- Active deals by pipeline stage
- Conversion rate by funnel stage
- Average lead time from lead to close
- Average ticket size for the period
- Conversion by channel (organic, paid, bot, referral)
- Dormant leads with no contact in 14+ days
Operations (5 KPIs):
- Pending orders vs. operational capacity
- First-message response time to customers
- OTIF (on time in full) for deliveries
- Stockouts on top SKUs
- Productivity per team member
Service (5 KPIs):
- Open tickets by category
- NPS over the last 30 days
- Average resolution time
- Complaints by type and trend
- Escalations to management
System (6 KPIs):
- Daily and monthly active users
- Slow queries or timeouts in the database
- Production errors in the last 24 hours
- Failed syncs between systems
- Storage usage vs. quota
- Third-party API calls hitting rate limits
Why KPIs Must Live in Code—Not in AI or Excel
Three critical reasons to calculate KPIs in TypeScript code:
- Auditability: every figure is traceable to an auditable function—not to an Excel cell nobody can explain
- Zero hallucinations: if you let AI calculate KPIs, it produces plausible but sometimes incorrect answers. AI is for narrative only, on top of verified data
- Immutable audit trail: with a SHA-256 hash chain, every figure is defensible to your board or an external auditor
This is the difference between guardrails and a ChatGPT wrapper. KPIs in code, not hallucinations. When data is unified, problems announce themselves.
The Real Case: 28 KPIs in Code, 5 Sections, SHA-256 Audit Trail
A multi-country distributor with 100 franchises engaged Catalizadora for advanced reporting. Result:
- Reporting with 5 sections (Financials, Sales, Services, Complaints, System Usage)
- 28 KPIs hardcoded in JavaScript with traceable calculation
- AI engine for narrative only—not for calculation (zero hallucinations)
- Browser-side compute (zero server CPU)
- Two-level pattern: numeric KPI headline plus AI explanatory paragraph
- Immutable audit trail with SHA-256 hash chain
- Every decision by every user traceable and verifiable
Cost: included in MAGIA / Core at $26,000 USD. Implemented in sprint S3 (May 18–25, 2026).
The Two-Level Pattern Format for Board Reporting
The most effective format for board reporting is the two-level pattern:
| Level | Content | Generated By |
|---|---|---|
| Headline | Numeric KPI with MoM and YoY delta | TypeScript code |
| Narrative | 2–4 sentence paragraph explaining what changed and why | AI engine on verified data |
Example:
Revenue MTD: $1.84M USD (+12% vs. prior month, +28% YoY)
"Growth is concentrated in the Bogotá franchise, which contributed $320,000 USD via the Cross Sell module launched in the previous sprint. Three franchises in the southern region show a 4–7% decline due to typical seasonality. Multi-channel attribution indicates that 38% of the increase came from the WhatsApp bot implemented in April."
Auditable, defensible, actionable.
What Should NOT Go in a Monthly Board Report
Three common mistakes:
- Vanity metrics without context: web visits, followers, impressions. They don't translate into action
- KPIs without a baseline or target: a single number says nothing. Always include delta and target
- Too many KPIs (40+): the board gets lost in noise. 28 to 35 is the optimal range
For context, see Wikipedia · Balanced Scorecard.
How Long It Takes to Build Reporting the Right Way
Typical implementation with MAGIA / Core:
| Phase | Weeks | Deliverable |
|---|---|---|
| Mapping | 1–2 | Unified Data Lake, prioritized KPI list |
| Architecture | 3–4 | KPI calculation in code, API contracts |
| Generation | 5–8 | Role-based dashboards, AI narrative engine |
| Deployment | 9–10 | Audit trail, parallel rollout |
| Autonomy | 11–12 | Operations manual, baseline, handoff |
12 weeks to have complete reporting with 28 KPIs in code and a full audit trail.
Three Common Mistakes When Implementing Board Reporting
- Trusting AI to calculate KPIs: it produces hallucinations. KPIs in code, AI for narrative only
- Manually generated Excel reports: every month someone copies and pastes for 4–8 hours—error-prone by design
- No immutable audit trail: when someone asks "who approved this," there's no answer
How Is It Priced and Delivered?
MAGIA / Core for executive reporting:
- $15,000 USD fixed (12 weeks)
- Unified Data Lake plus executive blueprint
- 28 KPIs in TypeScript code
- Role-based dashboards with RBAC
- AI engine with guardrails for narrative
- SHA-256 hash chain audit trail
- Operations manual plus training
- 4 weeks of support included
- Code and data 100% owned by the client
No retainers. No per-user licenses.
Next Steps
If your company builds monthly board reports out of Excel and nobody can defend the numbers, you don't need another analyst. You need KPIs in code. MAGIA / Core delivers a unified Data Lake plus 28 KPIs in code in 12 weeks for $15,000 USD. 30-minute call, no pitch deck.
- MAGIA / Core enterprise automation with KPIs in code
- MAGIA / Forge custom software with AI engine and guardrails