Skip to content
Back to Blog
Case study Integrations Metrology Colombia Latin America

How Global Sensors Moved Its ISO 17025 Calibration Lab from Excel to One Platform (Cold Chain, Colombia)

Calibration lab management software for an ISO 17025 cold-chain lab: 15 systems in production, live with real data about 7 weeks after the first commit, from CRM and signed offers to ISO sign-off and audit.

October 21, 2026 · 8 min read
Share:
How Global Sensors Moved Its ISO 17025 Calibration Lab from Excel to One Platform (Cold Chain, Colombia)

Global Sensors runs ISO 17025-accredited calibration for the cold-chain industry, and it was running sales, orders, invoicing and compliance paperwork on spreadsheets and email. WebMinds designed and built its calibration lab management software: one platform with a commercial CRM, a quoting engine with signed PDFs, a service-order and invoicing backbone, a client-signature compliance form, a structured WhatsApp intake and a reporting layer, all on a hardened Postgres backend with row-level security on every table. It went live with real data about seven weeks after the first commit.

Key facts

  • Client: Global Sensors, ISO 17025-accredited calibration for the cold-chain industry
  • Scope: 15 systems in production: CRM, quoting, orders, invoicing, ISO forms, WhatsApp intake, reports, client portal, automation and more
  • Speed: about 7 weeks from first commit to live data (June 19 to August 5, 2026), in continuous evolution since
  • Security: 46 of 46 tables under row-level security, with 103 policies and a role-permission matrix
  • Migration: 1,350+ clients migrated from legacy data, with NIT check-digit validation and duplicate detection
  • Audit: 13,000+ audited events since launch

The context: a compliance business running on Excel

Global Sensors ran its whole commercial and operational cycle in Excel and email: no traceability, no offer versioning, no role permissions and manual numbering prone to error.

For an accredited calibration lab, that is more than an inconvenience. ISO 17025 is about demonstrable competence and traceable records, and the commercial side of the business feeds directly into that record: what the client asked for, what was quoted, what the client signed, which order was executed and what was invoiced.

The problems we found were concrete:

  • Critical processes in spreadsheets, with no single source of truth and no record of who changed what.
  • Manual ISO 17025 compliance. The client form required before calibrating was filled in and signed outside any system.
  • Scattered inbound. Leads arrived through WhatsApp, the web form and phone calls, with no structured capture or advisor assignment.
  • No management visibility. Sales, conversion and billing metrics were assembled by hand, mixing quoted amounts with what was actually invoiced.

Why generic CRMs and lab tools did not fit

A generic CRM can track leads, and a generic invoicing tool can issue invoices. Neither knows that, in this business, no calibration order should exist without a client-signed ISO 17025 form. Neither knows that an offer needs annexes from a library, per-item VAT and a versioned history. And stitching several SaaS tools together would have recreated the original problem: several sources of truth and no single audit trail.

What Global Sensors needed was its full lead-to-offer-to-order-to-invoice cycle in one data model, with compliance rules enforced by the system itself rather than by memory. That called for custom web application development with most of the effort below the interface. By our estimate, three quarters of the work lives in the backend: a relational domain model, row-level security on every table, business-rule triggers, SLA engines and external integrations.

What we built: calibration lab management software for the full cycle

Fifteen systems on a single Supabase/Postgres backend. You can see the screens on the Global Sensors case study page.

Quoting engine with signed PDFs

Offers support per-item VAT, annexes from a library, server-side PDFs with merged annexes cached by content signature, a public approval link with client signature, versioning and automatic expiry.

PDFs only render when they change: they are generated on the server with annexes merged in, cached by a content signature, and hardened against page breaks splitting rows, totals or signatures.

Lead-to-cash pipeline

A CRM in Kanban and list views with bulk CSV import, service orders with purchase-order evidence, and invoicing reconciled automatically against orders. Management reports count billed revenue only, with or without VAT, so quoted amounts no longer inflate the numbers.

ISO 17025 client signature

A multi-equipment, multi-channel form the client signs before calibration, covering equipment, magnitudes and channels. No calibration order can be created without it.

The key decision was a signature that freezes the facts. When the client signs, their data is snapshotted, so certificates already issued never change if the client record is edited later. That is what makes this calibration lab management software defensible in an audit.

Structured WhatsApp intake

WhatsApp is the top lead channel for Global Sensors, about half of all leads. We built a five-question flow on the WhatsApp Cloud API that links each message to the CRM by phone number or creates the lead, stores media and tracks sent, delivered and read status.

It is worth being precise here: the intake is a deterministic flow, by design. The product itself runs no AI. For this use case, a predictable sequence of questions was the right tool, and it is a good example of integrations doing the heavy lifting without any model in the loop.

Automation, audit and reporting

A daily SLA engine, multichannel notifications (in-app, email and web push), audit triggers on every key entity, a client portal and management reports with or without VAT.

Leads close automatically after 15 days without follow-up, measured from the last real activity rather than from a timestamp that background processes kept refreshing, and long-cycle projects are excluded.

Row security without breaking the reports

Advisors see only their own orders, while coordinators and the technical lead see everything, and every dashboard still adds up. Getting row-level security and accurate aggregate reporting to coexist is one of those details that only shows up when it is done wrong.

Handover without downtime

The codebase moved to the client's own organization and the hosting was reconnected with no production downtime. WebMinds stays on as technical partner for support and new features.

Hybrid AI Coding

Senior WebMinds engineering steered AI agents (Claude Code) for implementation, with a QA agent reviewing every change. AI accelerated how we built the platform; the platform itself does not use AI.

Is your lab's compliance paperwork still living outside your systems? Get your free assessment and we will map how to bring sales, orders and compliance into one platform.

Under the hood

This was never a website. The numbers below the interface:

  • 46 of 46 tables with row-level security
  • 103 security policies
  • 66 SQL functions
  • 57 database triggers
  • 117 versioned migrations
  • 13,000+ audited events since launch

The stack: Next.js (as a PWA), TypeScript end to end, Supabase (Postgres, Auth, pg_cron), PostgreSQL row-level security and triggers, React-PDF for signed offers, the WhatsApp Cloud API, Resend for email, web push, Playwright for end-to-end QA and Vercel for hosting.

Timeline

  1. June 2026: Design, mockups and use cases.
  2. June 19–26, 2026: Platform bootstrap, then SSO, CRM pipeline, offers with PDF, orders, invoicing, client portal, audit and notifications.
  3. July 2026: Row-level security hardening after QA and authentication on every endpoint.
  4. August 4–5, 2026: Final delivery approved and live with real data.
  5. September 2026: WhatsApp automation switch, delivery timestamps, per-item VAT in PDF, owner-based row security.
  6. October 6, 2026: Codebase handed over to the client's organization; support and new features ongoing.

Results

We report only what the platform shows:

  • From Excel to a single source of truth. The full lead-to-offer-to-order-to-invoice cycle lives in one platform with traceability, versioning and audit: more than 1,350 clients migrated and over 13,000 audited events since launch.
  • ISO compliance that enforces itself. The client-signed ISO 17025 form is now a requirement the system enforces before any calibration order, with an immutable client snapshot and full traceability.
  • A commercial operation that runs itself. WhatsApp, the top lead channel at about half of all leads, flows into the CRM structured. The pipeline manages itself by SLA, and invoices and offers reconcile and expire on their own.

See how we approach similar back-office work on our internal automation and data and analytics pages. For a sibling project, read how DYMM Ingeniería replaced spreadsheets with a nine-role operations platform, or how we built scheduling for occupational health at IPS Aleso.

FAQ

What is calibration lab management software?

It is the system an accredited calibration lab uses to run its business and its records together: leads, quotes, client sign-off, service orders, invoicing and audit. For Global Sensors, it also enforces the ISO 17025 client form before any calibration order can exist.

How does the platform support ISO 17025 compliance?

The client signs a multi-equipment form before calibration, and the system blocks calibration orders without it. At signing, client data is frozen in a snapshot, and audit triggers record every change to clients, leads, offers, orders and invoices.

Is the WhatsApp intake an AI chatbot?

No. It is a deterministic five-question flow on the WhatsApp Cloud API, by design. It links each conversation to the CRM by phone or creates the lead, stores media and tracks delivery status.

How long did it take to go live?

About seven weeks: the first commit was on June 19, 2026, and final delivery was approved and live with real data on August 4–5, 2026. The platform has kept evolving since.

Who owns the code?

Global Sensors does. On October 6, 2026, the codebase was handed over to the client's own organization with no production downtime, and WebMinds continues as technical partner.

Does WebMinds build this kind of platform outside the US?

Yes. WebMinds is an AI and software development company based in Atlanta, Georgia, that works with companies in the United States and Latin America. Global Sensors works with NIT tax IDs, which is why client migration included NIT check-digit validation.

Bring your lab onto one platform

Read the full Global Sensors case study or tell us how your lab runs today.

Get your free assessment or book a call with our team.

Ready to explore these ideas?

Let us help you turn insights into action. Get a free, no-obligation assessment of your digital strategy.

Book a Free Assessment→
Stay in the loop

Want more insights?

Subscribe to get weekly insights on strategy, technology, and growth delivered straight to your inbox.

Hi! I'm Atlas