Artificial intelligence, especially generative AI, is shaking up the business world. There’s a buzz out there, almost like we’re at the height of the Gartner hype cycle. But here’s the kicker: many companies are about to discover that implementing AI isn’t as straightforward as it seems.
One major challenge is finding practical use cases that provide a quick return on investment and can scale without breaking the bank. This is especially true when using large language models (LLMs), where expenses can pile up fast. Computer Weekly has noticed a trend where businesses are leaning away from public LLMs, opting instead for smaller, on-premise AI models that can be fine-tuned.
Experts stress that to train AI effectively with internal data, that data has to be high quality. But enterprise data often lives in different systems, creating a messy situation where there’s no single source of truth. Consolidating this data can be complicated, as various departments own different applications, and a unified data structure may not be a top priority.
Moreover, AI often ties into how enterprise resource planning (ERP) providers are pushing organizations towards cloud solutions. But many companies use heavily customized ERP systems. Transitioning these customizations can be a nightmare. Take Birmingham City Council’s switch from SAP to Oracle. The true costs ballooned far beyond initial estimates, and manual processes stalled financial reporting, leaving a major budget gap.
IT leaders wrestling with long-term ERP upgrades face another hurdle: the software they chose at the start may no longer fit their needs just a year later. The temptation to snatch up discounted software bundles can lead to shelfware. While the upfront price might look good, ongoing maintenance costs can quickly chip away at those savings.
Some companies turn to third-party support to manage these costs, which can keep older software running longer. These providers can access any patches and support documents the customer has downloaded before making the switch. Once they do, though, they lose direct access to updates from the original provider. That said, since many products moving to third-party support have been in use for a while, the likelihood of new, drastic issues arising is low.
Here are the top 10 business application articles from Computer Weekly this year:
-
Corporate credit card pitfalls: An easy purchase process for SaaS can lead to unexpected costs and data leaks.
-
Rising VMware costs for schools: Academic organizations, like London Grid for Learning and KU Leuven, face drastic licensing increases as Broadcom ends its discount program.
-
Birmingham’s Oracle blunder: After launching in April 2022, Birmingham City Council’s Oracle implementation suffered a major IT failure.
-
Generative AI in real life: We look at five organizations testing generative AI and the lessons they’ve learned.
-
SAP upgrades: The chair of the Independent UK and Ireland SAP User Group describes the upgrade journey as a marathon, highlighting that SAP Rise is on the horizon.
-
Are you AI-ready? This article breaks down what it means to prepare your business for AI adoption.
-
Building trust in AI: New platforms aim to boost confidence in AI by helping businesses identify and manage associated risks.
-
ServiceNow’s AI ambitions: We report how a partnership with Rimini Street positions ServiceNow as a go-to platform for enterprise AI.
-
Philippine Airlines’ transformation: The airline shares how its digital initiatives have enhanced customer experience and optimized workflows.
- Open source’s value: Companies are discovering that open source software offers significant benefits, like increased productivity and lower operational expenses.