Posts

Showing posts with the label AI implementation

Why Retail Returns Become a Technology Problem as Businesses Grow

Image
  The return looks simple from the customer's side. The customer sends the product back. The retailer receives it, processes the return, issues a refund, and puts the item back into inventory if it can be sold again. Behind that simple transaction, however, several systems may need to agree. The ecommerce platform has an original order. The payment system has a transaction. The warehouse has the returned product. The inventory system needs to know whether it can sell that product again. Finance needs to record the refund. Customer service needs to know what happened. When a retailer has few returns, employees can often fill the gaps manually. As the business grows, that approach becomes harder to maintain . The return itself isn't necessarily difficult . Keeping every system in sync is. Returns Touch More Systems Than Most Retailers Expect Consider a customer who buys a jacket online and returns it to a physical store. The store needs to identify the original order, confi...

Why Better AI Models Don't Always Lead to Better Business Outcomes

Image
  Introduction Every few months, a new AI model emerges that can do what the previous generation could not. It can handle longer documents, answer more complex questions, write better code, or process information faster. Technology executives operate under the assumption that better models lead to better business outcomes. It is not always the case.   An organization might be equipped with such a highly skilled model but still not manage to extract value from it. And it may not even be about the model itself. It may be related to the fact that data is  disperse d among different systems. Th e employees may find it difficult to  use  the model's output. Some critical business processes may still be carried out using Excel and manual approvals. That is how many enterprise AI implementations fail to deliver results. A Better Model Doesn't Fix a Broken Process Consider a customer service team handling a growing number of requests. The team uses a CRM, a billing ...