AI: Start With the Problem, Not the Solution

Technology

Sep 29

Written By Berkshire Hospitality & Healthcare

To use or not to use AI? That may NOT be the question.

New AI tools and features are constantly being introduced.

At the same time, concerns about AI’s environmental and societal impacts are growing.

So how do we use it responsibly?

One way is to start with the problem, not the solution.

What are we trying to improve?

Efficiency?
Cost?
Speed?
Quality?

Here are a few examples:

  
      Area
      Problem
      Possible AI Application
    
      Fiber Production
      Cotton yield prediction
      Optimizing and managing crop production relies on accurate yield predictions. A study used machine learning and satellite imagery to predict cotton lint yield at different spatial and temporal scales.
    
      Textile Manufacturing
      Dyeing errors
      Incorrect dyeing can result in significant fabric waste. Because fabric can change color as it dries, researchers developed machine-learning models that can predict the final dry color while the fabric is still wet.
    
      Inventory Management
      Fluctuating demand
      A study focused on textile and PPE supply chains combined machine-learning demand forecasting with an optimization system to recommend replenishment and allocation decisions. In the study's experiments, the system reduced inventory costs and stockouts while improving service level.
    
      Hotel Housekeeping
      Efficiency
      One study used an AI system to dynamically coordinate housekeeping schedules and balance staff workload. In a 30-day hotel trial, the researchers reported a more than 50% reduction in room turnaround time, along with improvements in task completion and workload distribution.

Area	Problem	Possible AI Application
Fiber Production	Cotton yield prediction	Optimizing and managing crop production relies on accurate yield predictions. A study used machine learning and satellite imagery to predict cotton lint yield at different spatial and temporal scales.
Textile Manufacturing	Dyeing errors	Incorrect dyeing can result in significant fabric waste. Because fabric can change color as it dries, researchers developed machine-learning models that can predict the final dry color while the fabric is still wet.
Inventory Management	Fluctuating demand	A study focused on textile and PPE supply chains combined machine-learning demand forecasting with an optimization system to recommend replenishment and allocation decisions. In the study's experiments, the system reduced inventory costs and stockouts while improving service level.
Hotel Housekeeping	Efficiency	One study used an AI system to dynamically coordinate housekeeping schedules and balance staff workload. In a 30-day hotel trial, the researchers reported a more than 50% reduction in room turnaround time, along with improvements in task completion and workload distribution.