Tech Stack & AI Strategy Advice: Logistics No-Code Project for freezers follow-ups in our juice factory

Hi guys,

I am currently building a field operations and asset management system for our commercial freezer fleet at "Juice Company" (moving from manual WhatsApp reporting to a digital cloud ecosystem). I have zero coding experience, but my English is good, and I am relying heavily on AI (LLMs) to architect, code, and guide me through the development.

I would highly appreciate your expert advice on the optimal platform choice and AI-coaching strategy.

1. Core Concept & 10 Standard Operations

We need a dynamic system to track our distributed freezers and field representatives' movements via a dual-phase flow: "Service Request (Remote)" & "Execution (On-site)".

Field agents must select from 10 standard operations with conditional data entry:

1.      Refill: Inventory replenishment (Small cup, Large Cup, etc.).

2.      Fix: Maintenance logs (Compressor, Fan, Freon, etc.).

3.      Relocate: Moving a freezer between points of sale.

4.      Withdraw: Returning asset to the warehouse.

5.      Lost: Reporting missing assets immediately.

6.      Visit: Routine checks & cleanliness audits.

7.      Update File: Photographing and updating customer commercial licenses.

8.      Sales Issue: General marketing and sales obstacles.

9.      Extra Freezer: Requesting more freezer assets for high-volume sites.

10.  Change Freezer: Swapping a faulty unit with a new one.

2. Operational Requirements:

·         Offline-First: Representatives frequently work in superstore basements or remote areas with zero cellular connection. The app must store logs locally and auto-sync when back online. (or at least manual location paste if the connectivity issue forces a workaround).

·         GPS Tracking & Heatmaps: We need freezer locations mapped onto a main dashboard Heatmap to identify active vs. dead zones.

o    Workaround: If mandatory GPS locking is too complex, reps can manually copy-paste Google Maps URLs when connected, with the database coloring.

·         Future AI Integration: The stack should scale so that we can eventually integrate AI Computer Vision (e.g., photo-scanning shelves for auto-stock audits, automated barcode/serial reads).

·         Low Middleware Complexity: I want to avoid complex background piping (no heavy reliance on Make.com or Zapier if a single native platform can handle it).

3. The Dilemma: Platform Choice

·         Option A: AppSheet + Google Sheets

·         Option B: Airtable + Native Interfaces

My Questions to You:

1.      Which No-Code Platform is Best? Considering our offline-first requirement, GPS tracking, and Heatmap needs, which route is structurally more stable and scalable for a non-coder?

2.      Which AI Assistant is Best? Since I am leveraging AI to build and debug this, which LLM model (Claude 3.5 Sonnet, GPT-4o, Gemini, etc.) have you found to be the most accurate for architecting AppSheet expressions, Airtable formulas, or workflow logic?

3.      Are There Other Platforms? Is there another No-Code platform I missed that bridges the offline/GPS capabilities of AppSheet with the beautiful UI, Map Views, and future AI scalability of Airtable?

Thanks in advance for your insights!