-->

Keep up with all of the essential KM news with a FREE subscription to KMWorld magazine. Find out more and subscribe today!

  • November 4, 2024
  • By Marydee Ojala Editor in Chief, KMWorld, Conference Program Director, Information Today, Inc.
  • Features

Lucidworks shares results from a global survey about GenAI adoption, fears, and strategies in 2024

Article Featured Image

The second part, after the business case, is building out a work plan—what are the use cases and how is success measured against those use cases? Have a specific work plan for the next quarter and a pretty general work plan for the next couple of quarters. It’s pointless to plan anything more than a year out right now in GenAI because the technology is changing so fast. But absolutely have a road map. Remember, there are multiple stakeholders—it can’t just be the IT group saying, “All right, here’s our plan to deploy this AI tool.”

Some of the most successful applications of GenAI, said Sinoway, have been in the accounting and cost reduction area. Coding for new computer systems and testing that code are oth- ers. The human resource area has many use cases, whether it’s developing training materials or customizing training programs by positions, helping craft job descriptions, or helping screen candidate resumes.

Tools are being deployed within finance to do procurement activities and manage accounts such as cash receivables and accounts payable. Sinoway expects to see these tools realize some of the vision that companies have for sales and marketing. Another area for GenAI is content creation for marketing sites, promotion, and those kind of things. KM use cases involve pro- cessing signals from different parts of the organization to help make better decisions.

Pilot testing

The next logical step is to pilot test these tools and put them through a rigorous evaluation before being fully deployed. Before deployment, ensure the right data is in the tool, make sure the information is accurate, that the right people have access to it, and that it doesn’t break the bank. A small-scale pilot test to learn how to use these tools is valuable.

At Lucidworks, a year of pilot testing has given us insights into how to chunk a document. Chunking is when you take these big documents and break them up into small enough bites that an AI tool can process them accurately and quickly, but not breaking them up so much that you’re disconnecting the key insights.

It’s the same thing with using different models. A model that does the indexing may not be the same model you want to do the query, the query response, indexing, and summarizing. Indexing can happen in batches and should be very cost effective. But when someone asks a question, you want it to be in very natural, conversational language. Those could be two different model use cases, and figuring out which two models to use entails a certain amount of trial and error, pilot testing, and proof of concept.

To summarize, figure out what to deploy and how to deploy it. Start with governance, followed by business case, followed by proof of concept and testing, and then take baby steps to see what works.

AI tools as workplace differentiators

Sinoway is convinced that jobs won’t be replaced by AI. Instead, AI tools will be a differentiator in the workforce. KM professionals don’t need to know about weightings, transformers, or back propagation. They absolutely do need to know how to use these tools, how to put an effective prompt into the system, and, most importantly, how to use the output to be more effective.

KM professionals need to stay up-to-date because AI technologies are changing so fast. The tools you get familiar with today won’t be the leaders next week and certainly not next month. So staying on top of the developments and understanding how those developments apply to your specific function or area is going to be critical.

KMWorld Covers
Free
for qualified subscribers
Subscribe Now Current Issue Past Issues