Top 10 AI Prompts and Use Cases and in the Retail Industry in Colombia

AI is already reshaping Colombian retail - from smarter inventory and predictive analytics to personalized recommendations and chatbots - and the country's mix of talent and momentum makes it a practical place to act now: Colombia graduates over 13,000 STEM professionals a year and offers compelling nearshore advantages highlighted by nearshore AI development services in Colombia, while market analysts expect AI in Latin American retail to surge (the region's retail AI market is projected to grow sharply through 2032) according to the Latin America AI in retail market forecast.

For Colombian retailers and partners this means faster pilots, better demand forecasting, and richer omnichannel experiences; staff who learn to write effective prompts and operationalize models - through practical programs like the AI Essentials for Work bootcamp - can turn these tools into measurable cost savings and smoother shopping moments for customers.

Bootcamp	Length	Early bird cost
AI Essentials for Work	15 Weeks	$3,582

“We are in the age of artificial intelligence and data. Beyond companies that develop technology directly, we believe that all companies, even those that do not have AI at the core of their business today, will be able to use it to substantially improve their productivity.”

Selection of the top 10 prompts began with Colombia's strategic context - priority was given to use cases that align with the national AI roadmap, local talent pipelines, and the CONPES‑style regulatory focus noted in recent analysis, so candidates had to promise measurable retail impact while fitting Colombian governance and workforce realities; sources and constraints also shaped prompt design, drawing on common retail needs (personalized shopping, inventory management, demand forecasting, chatbots and checkout optimization) highlighted by industry research and on infrastructure realities such as the power, cooling and GPU capacity required for production AI. Prompts were therefore chosen for pilotability (clear KPIs like improved inventory availability or reduced checkout downtime), ethical and privacy readiness (see guidance on data privacy and regulation for Colombian retailers), and practical deployment paths that account for local infrastructure limits described in AI infrastructure briefs - so the result is a set of prompts built to be actionable in Colombian stores and distribution centers, not abstract experiments; think of a pilot that's as tangible as the hum of GPUs and direct‑to‑chip liquid cooling in a freshly upgraded data center.

“Artificial intelligence is presented as a fundamental tool that can positively shape the future of our nation. But its development must be guided by solid ethical principles and a strategic vision that guarantees the well-being of all Colombians.” - Yesenia Olaya

Farmatodo's Colombia app blends fast convenience with smarter product discovery: the Store's mobile experience (4.7 stars, 1M+ downloads) lets shoppers search by brand, presentation, size or supplier and reuse past orders for one‑tap repurchases, and its promise of deliveries in less than 45 minutes makes quick recommendations especially valuable; Google Cloud even lists Farmatodo among real‑world examples using generative AI to personalize product recommendations and improve semantic search, showing how embeddings and natural‑language queries can surface the exact medicine, cosmetic or baby item a customer needs - right when they need it - helping Colombian retailers close the gap between discovery and delivery in a market where consumers spend a large share of their day online.

Farmatodo's strong app ranking in local shopping charts underlines that personalization plus speedy fulfillment resonates with Colombian users and creates a practical pilot lane for other chains looking to deploy recommendation engines and semantic search.

Farmatodo Colombia mobile app (Google Play) and examples at Google Cloud generative AI real-world use cases (see local app rankings via Colombia shopping app rankings on Similarweb).

Metric	Value
Rating	4.7 stars
Reviews	60.3K
Downloads	1M+
Delivery promise	Orders in less than 45 minutes
Last updated	2 Sept 2025

For Coop in Colombia, moving inventory planning from heuristics to AI means fewer empty shelves and smarter stocking decisions at the store and SKU level: analytics-driven pilots can reproduce the Prescience example that delivered weekly product‑level forecasts with over 95% accuracy and lifted availability above 85% by combining multiple models and automated reorder alerts (Prescience demand forecasting case study - 95% SKU forecast accuracy).

Alternatives like ForecastSmart show how next‑gen engines blend external signals (promotions, weather, mobility) to adapt faster than legacy systems and drive marked drops in lost sales and clearance risk (ForecastSmart NextGen retail demand planning (promotions, weather, mobility)).

For Coop's rollout, start small - SKU-store pilots, short horizons, automated alerts - and scale with part‑level tooling such as Databricks' accelerator so models train in parallel across thousands of SKUs; the payoff is tangible: imagine supplier trucks pulling up with the exact mix your AI foreshadowed, turning reactive restocking into predictable fulfillment.

Databricks part-level demand forecasting accelerator - scalable SKU forecasting

Source	Key result
Prescience case study	Weekly SKU forecasts >95% accuracy; availability >85%
ForecastSmart / Impact Analytics	90%+ agile location forecasts; 99%+ on‑shelf availability (product claims)
Databricks accelerator	Part‑level, scalable forecasting across thousands of SKUs

Revionics-style dynamic pricing and promotion optimization gives Colombian retailers a practical lever to protect margin, move slow inventory and react to local competition - think of a Bogotá convenience aisle where prices and promotions subtly shift between the morning commuter surge and a quiet afternoon, nudging purchases without breaking trust; start by learning the core mechanics in a clear primer like Tredence's dynamic pricing guide and then test vendor playbooks that show measurable uplifts (see Invent.ai's dynamic pricing use cases) while keeping local data‑privacy and regulatory guardrails in view for compliant personalization.

A sensible pilot in Colombia pairs elasticity modeling with simple guardrails (min/max price bands, promo caps and clear customer communication), runs A/B tests on background SKUs, and measures revenue, margin and conversion impacts before broader rollout - this staged approach balances the upside of real‑time pricing with the reputational risks called out by pricing specialists.

Metric	Reported range
Revenue uplift (Invent.ai)	2–7%
Gross margin improvement (Invent.ai)	2–4%
Conversion lift (Invent.ai)	5–20%
Operating profit sensitivity (McKinsey cited)	1% price ↑ → ~8.7% op profit ↑

“Teknosa doubled its net-profitability within one year of starting to implement price optimization results systematically. This transformation revealed the significance of using big data and analytics accurately.” - Baris Oran, CEO, Teknosa

For Colombian retailers building smarter contact centers, Google Dialogflow CX (and Contact Center AI) is a practical way to deliver omnichannel, multi‑step support - chat, voice and phone - without forcing customers to repeat themselves across channels: the platform's flow-and-page model in the Dialogflow retail codelab shows a full order/status/returns agent you can adapt for Bogotá stores, while the Dialogflow CX best practices guide explains concrete production steps - use agent versions and backups, reuse SessionsClient instances, implement webhook retries and load testing, and design clear no‑match/no‑input fallbacks to avoid dead‑end conversations.

That combination matters in Colombia where seamless handoffs (chat → phone → agent) and unified CRM data are expectations, not nice‑to‑haves: follow omnichannel customer service best practices to integrate channels so a shopper never has to repeat an order number or delivery address.

Start small with the retail codelab patterns (flows, parameters, conditional routes) and validate latency and retry behaviors noted in the docs - remember that LLM and speech components add variable latency - then pilot on a high‑value use case (order status or refunds) so the first customer interaction feels as effortless as a sales clerk handing over a receipt.

Learn more in the Dialogflow CX production best practices guide, the Dialogflow CX retail chatbot codelab, and the omnichannel customer service guide for retail to fast‑track a compliant, testable Colombian rollout: Dialogflow CX production best practices guide, Dialogflow CX retail chatbot codelab, and the omnichannel customer service guide for retail.

Vertex AI Vision brings two tightly connected capabilities Colombian retailers can pilot today: image-based product discovery for online shoppers and live shelf/compliance recognition in stores.

For marketplaces and apps, Vision's product-search embeddings accelerate the “find it from a photo” flow that reduces search abandonment - pairing neatly with Vertex AI Search for commerce to return visually and semantically similar SKUs with personalized ranking (Vertex AI Search for Commerce product-search solution).

In brick-and-mortar pilots, the platform's serverless video ingest and pretrained detectors make it realistic to run continuous shelf checks and planogram audits at scale, streamlining alerts into BigQuery for analytics so teams can act in near real time - imagine an aisle “eye” that notices a missing SKU and surfaces it to a store manager before the lunch rush.

The managed, low-code app builder and import options for custom models reduce build time and cost, while proven retail deployments (see a practical example of image-based product matching that cut processing time and lifted conversions) show the business case for Colombian chains (Vertex AI Vision documentation and features, Miinto case study: image-based product matching with Vertex AI Vision).

Metric / feature	Value / note
Real-time video ingest	Serverless streaming at global scale
Cost reduction claim	Monthly pricing ~ one tenth prior offerings
Miinto: arrival processing	Efficiency ↑ >40%
Miinto: duplicate matching	Duplicate in top‑5 results 100%; first result 98%; matching <1s
Miinto: conversion impact	Conversion ↑ up to 20%

"Vertex AI Vision is changing the game for use cases that were previously economically unviable at scale. The ability to run computer vision models on streaming video with up to a 100X cost reduction for Plainsight is creating entirely new business opportunities for our customers." - Elizabeth Spears, Co-Founder & CPO, Plainsight

Amazon Go's “Just Walk Out” model is the clearest example of cashier‑less checkout Colombian retailers should study: ceiling cameras, sensor fusion (including shelf weight/load sensors) and deep‑learning object recognition keep a virtual cart updated as shoppers move through the store, and payment is finalized automatically when customers leave - so curious crowds first lined up not to pay but to watch the tech in action.

The approach scales from compact micro‑stores to stadium concessions (Lumen Field reported dramatic transaction and sales uplifts) and relies on synthetic training data and generative techniques to handle rare edge cases, which matters in Bogotá or Medellín where store layouts and lighting vary.

For pilots in Colombia, pair a small-format site with robust edge compute and clear privacy/legal guardrails (see the technical overview of Amazon's Just Walk Out and the sensor‑fusion primer on checkout‑free shopping) so the early KPI is reliability - accurate receipts, near‑zero false charges - and the memorable outcome is simple: shoppers walk out with their receipt in an app and no line ever breaks their morning commute.

Practical rollouts will balance infrastructure (edge nodes, cameras, latency) with consumer trust and transparent data practices before scaling beyond pilot stores.

Item	Note / metric (source)
Core technologies	Computer vision, sensor fusion, weight/load sensors, mobile app authentication (About Amazon)
Training approach	Use of synthetic/generative data to improve rare‑case recognition (About Amazon)
Selected pilot results	Market Express: +300% peak customer service, +56% annual revenue; Lumen Field: large transaction and sales uplifts (Aimultiple / About Amazon)

Grupo Nutresa's reach - dozens of well‑known food brands and direct presence across Latin America - combined with a digitally modernized procurement backbone makes it a vivid example of where Gemini and Imagen can accelerate automated creative and localized marketing in Colombia: rather than manual agency cycles, generative models can stitch supplier catalog metadata, contract terms and regional sales signals into on‑brand ad copy, label variants or short social videos tailored by city and channel, shaving days off campaign launches and keeping messaging consistent across Noel, Zenú or Colcafé shelves; see Nutresa's profile and international footprint for context and its procurement results with SAP Business Network that freed buyers to focus on strategy, not paperwork (Grupo Nutresa company profile and international footprint, Grupo Nutresa procurement transformation with SAP Ariba and SAP Business Network case study), and note how enterprises are already using Gemini in Google Workspace to speed creative workflows at scale (Google Workspace Gemini customer success stories).

Imagine a morning coffee promo that auto‑localizes copy and art for Medellín, Cali and Barranquilla before breakfast - small creative moves that translate to real campaign velocity and measurable savings.

Metric	Value
Representative brands	Noel, Zenú, Colcafé, Doria, Sello Rojo
International presence	Multiple countries across Latin America and beyond
Q3 2023 sales growth (consolidated)	17.5% increase
EBITDA margin	11.9%
SAP Business Network: suppliers onboarded	300 strategic suppliers
Procurement savings reported	50% (using digital tools)
Contracts consolidated	100% on a single platform

“With SAP Business Network and SAP Ariba solutions, we are confident that our buyers are purchasing the right items, at the right time, at the right price, from the right suppliers.” - Carlos Peña, Digital Procurement Leader, Grupo Nutresa SA

Integrating Nequi and Daviplata with robust fraud detection and payments risk‑scoring turns mobile wallets into a frontline compliance tool for Colombia's fast‑moving digital economy: start with strong KYC and ongoing monitoring that map to national rules and UIAF reporting expectations (see a practical KYC primer for Colombia), combine quick, high‑accuracy identity checks and device/behavior signals from identity verification providers like FACEKI, and layer dynamic AML risk scoring and transaction‑monitoring to flag anomalies in real time; this stack reduces false positives, speeds onboarding, and lets fintechs apply differentiated controls (simple CDD for low‑risk users, EDD for flagged profiles) while keeping sanctions and country‑risk inputs in the loop.

A pragmatic rollout sequence - instant eKYC at signup, continuous scoring on location/behavior/amount, and automated alerts to a compliance queue - shifts fraud work from manual review to targeted investigation, so teams see risky patterns instead of drowning in noise.

Metric	Value (source)
Digital fraud growth	43.5% increase in attempts (Didit)
eKYC onboarding time	~30 seconds average (FACEKI)
eKYC accuracy	99.8% (FACEKI)

Onfleet‑style last‑mile routing and pickup optimization turns Colombia's fast‑growing delivery challenge into a competitive advantage by combining AI‑powered route planning, driver enablement and customer‑facing notifications so planners stop drawing loops on maps and start running predictable, measurable operations; the Colombian market is expanding rapidly (expected >10.06% CAGR to 2030), so investing in smarter routing pays off in urban congestion and rural‑last‑mile complexity alike - follow practical playbooks like Descartes' last‑mile best practices to standardize planning, automate appointment scheduling and grade drivers by on‑route data, and use modern route engines to produce accurate ETAs and proactive reroutes as described in the DispatchTrack route optimization guide (teams that adopt these tools can cut route‑planning time by half, boost sales 10%+, and approach 98% ETA accuracy).

Start with small pilot zones - a few Bogotá or Medellín routes - measure on‑time performance, failed deliveries and customer notifications, tune AI‑based configuration for local streets and driver habits, and scale: the result is clearer windows for customers, fewer wasted kilometers and a delivery network that can actually keep up with Colombia's rising demand (Colombia last‑mile delivery market overview).

Grupo Nutresa's scale - dozens of household food brands and a modernized procurement backbone - makes it a perfect candidate for an internal “store assistant” agent that boosts employee productivity by turning complex documents and SOPs into immediate actions: unified search across contracts, a quick summary of supplier terms, step‑by‑step checklists for a store manager, or a drafted promotional brief localized by city.

Tools like Google Agentspace prompt guide show how an enterprise agent can safely index internal docs, summarize multi‑document threads with NotebookLM, and run no‑code agents that draft, iterate, and enforce procedures at the point of work.

Paired with the procurement gains Nutresa reported from its SAP rollout, an assistant that surfaces supplier risk, onboarding status, or a one‑line corrective SOP before the lunch rush becomes a practical productivity lever rather than an experiment (Grupo Nutresa & SAP Business Network case study).

The memorable win: fewer inbox searches, faster store fixes, and more time for sales‑focused customer care.

Metric	Value
Representative brands	Noel, Zenú, Colcafé, Doria, Sello Rojo
International presence	Multiple countries across Latin America and beyond
Q3 2023 sales growth (consolidated)	17.5% increase
EBITDA margin	11.9%
Suppliers onboarded (SAP Business Network)	300 strategic suppliers
Procurement savings reported	50% (using digital tools)
Contracts consolidated	100% on a single platform

“With SAP Business Network and SAP Ariba solutions, we are confident that our buyers are purchasing the right items, at the right time, at the right price, from the right suppliers.” - Carlos Peña, Digital Procurement Leader, Grupo Nutresa SA

Execution in Colombia should start small, measurable and compliant: run focused pilots that ground models in your own documents and product data, measure with a repeatable testing framework, and hard‑wire governance so AI helps - rather than surprises - store teams.

Prioritize a data preparation pipeline and a secure vector store, iterate retrieval and chunking strategies, and optimize prompt templates so outputs cite sources and follow local privacy rules; TechTarget's RAG checklist recommends exactly this phased approach to reduce hallucinations and keep proprietary knowledge usable and auditable (TechTarget RAG best practices for enterprise AI teams).

Use grounding tools like Vertex AI's RAG/grounding guides to mix retrieval, check grounding, and evaluate with golden question sets, and couple that with a Colombia‑specific compliance review (see Nucamp's primer on data privacy and regulation for local context) so pilots respect KYC, consumer protections and data minimization (Vertex AI guide: Ground responses using RAG, Nucamp Cybersecurity data privacy and regulation guide for Colombia).

The payoff is practical: a store manager who receives an AI‑backed, source‑linked action before the lunch rush instead of sifting documents - faster fixes, clearer audits, and controllable risk.

RAG Practice	Quick summary
Data preparation	Filter, chunk, clean and version authoritative sources for retrieval.
Vector database	Select scalable, secure vector DB (Pinecone/Weaviate/Milvus/Qdrant or managed service).
Retrieval strategy	Hybrid semantic+keyword search, reranking, and metadata filters.
Security & compliance	ABAC/role controls, encrypted embeddings, logging and audit trails.
Prompt engineering	Standardized templates, citation rules and iterative testing.
Governance	Monitor drift, track sources, run red‑teaming and human evaluation loops.

Getting started with AI prompts in Colombian retail means choosing a narrow, measurable pilot - think conversational order-status agents, SKU-level demand forecasts, or a visual-shelf check that flags a missing product before the lunch rush - and wiring those pilots to clear KPIs, a secure retrieval store, and governance so outputs are auditable and privacy-safe.

Local proof points and tooling make this realistic: Google Cloud's catalog of real‑world generative AI use cases highlights regional implementations (for example, GroupBy's Vertex AI Search work), and retail teams should pair prompt engineering with RAG grounding, small A/B experiments, and iterative human review to shrink hallucinations and prove value fast.

Train people to write better prompts and interpret model outputs - practical courses like the AI Essentials for Work bootcamp teach those on‑the‑job skills - and start with vendor‑agnostic blueprints (search, forecasting, chat, vision) so pilots convert into predictable operations rather than one‑off demos.

The most memorable wins are concrete: a manager alerted to a missing SKU before peak service, a routed driver meeting an ETA, or a localized promo generated in minutes - small, repeatable outcomes that build trust and a roadmap for scaling AI across Colombia's stores.

Bootcamp	Length	Early bird cost
AI Essentials for Work bootcamp	15 Weeks	$3,582

“With the right strategy, CFOs can create substantial benefits by deploying emerging technologies such as AI.”