Now, the conversations have evolved and become more specific. What are the latest models and companies? What AI technologies should I use for this specific initiative?
Pluralsight Author and AI expert Amber Israelsen answers these questions. She also shows how to use different AI tools to improve efficiency and decision-making in your organization.
Want to see these AI solutions in action? Watch the on-demand webinar to see Amber’s demos.
Table of contents
Practical AI use cases and tools for organizations
- Generate videos with tools like OpenAI Sora
- Create different types of music
- Generate code from images
- Adopt AI agents for better productivity
- Build custom GPTs for specific tasks
- Use RAG for content management
Keep AI security in mind
AI is a partner for people, not their replacement
Practical AI use cases and tools for organizations
When it comes to exploring AI tools (or any new technology), it’s often best to get your hands dirty and just start experimenting with them. But if you don’t know where to start or don’t have time to jump into them yourself, Amber recommends several solutions you can use to power up your organization.
“Just because I'm showing one tool doesn't mean it's the only tool or the best tool for the job. I really just want to give you some examples of what's out there,” she says.
1. Generate videos with tools like OpenAI Sora
Let’s say you want to create a marketing video for your product, but you’re short on producers and editors. You can use tools like OpenAI’s video generator Sora, Adobe Firefly, or Synthesia to create videos. You can also change different settings like:
- Video style
- Aspect ratio
- Resolution
- Length
2. Create different types of music
If you need a certain style of music for a video, an ad, or anything else, AI can produce exactly what you need.
Tools like ecrett music allow you to set a scene, mood, and genre before creating a tune that matches your criteria. Other, more advanced tools, give you the option to get even more specific.
3. Generate code from images
AI makes it easy to generate new images. You can also upload images to tools like Microsoft Copilot and generate code based on your photo.
“Maybe you want to build a simple website or a mobile app or something like that. You don't really know how to code it. But I've got a picture of a coffee cup and a title and a description, and a buy now button. I've got navigation,” says Amber.
“So you can just upload this to Copilot, or the other tools out there as well, and it will generate some code for you. So we've got HTML. We've got CSS. It works on all the popular languages, as you would expect. You could take this [code] right here, and you could theoretically run this website.”
4. Adopt AI agents for better productivity
AI agents can accomplish complex tasks on their own, which makes them a useful tool for boosting productivity and aiding decision-making.
Amber gives the example of an IT Ops manager who wants an AI assistant that can research pricing for big cloud providers, calculate the monthly cost difference between them, and recommend one for their organization.
Normally, that process would require the IT Ops manager to:
- Use a search engine to find more information on each cloud provider
- Perform calculations in Excel or some other calculator tool
- Summarize findings in a document or text editor
“It is in early days, and it consistently fails at the moment, at least for me,” says Amber. “But I would imagine this is only going to get better. . . . You can also build your own agents with various agentic frameworks that are popping up out there as well.”
Learn how to use the AI agent ChatGPT Pro: OpenAI Operator.
5. Build custom GPTs for specific tasks
Custom GPTs are designed for a particular task or skill. ChatGPT provides pre-made GPTs with specific instructions, knowledge, and specializations. The Data Analyst GPT can be particularly helpful for analyzing and visualizing your data.
You can also create your own custom GPT specific to your organization or processes.
“You don't have to know how to code. There's no crazy configuration here. You basically just have a conversation, and it will configure things for you on the back end,” says Amber.
For example, maybe you want to create a tech hiring assistant to help recruiters with job descriptions, resume evaluations, and interview questions. Give the GPT some context and any additional documents or domain-specific information it should know.
Once the assistant is built, it can do things like generate a job description for a machine learning engineer, go out to external sites, call APIs, and more.
6. Use RAG for content management
Retrieval augmented generation (RAG) is an AI technique that retrieves relevant information from sources to generate responses.
Feed it your documents and data, and you can ask questions, like What were my best-selling products last quarter? Or Which customers are at risk of churning next month? The tool will provide answers based on your data.
Some solutions, like Amazon Q Business, will even show you its sources so you can double check its responses or dive deeper into particular documents. These tools can make mistakes and hallucinate, so it’s important to verify their responses for accuracy.
“A great use case for this is just content management,” says Amber. “If you think of your company intranet—at least every one I've ever used—it's a little difficult to find things. So if you can just point a chat bot at all of your data that you have out there and use RAG, then people can just ask questions about it instead of having to know exactly where to navigate.”
Keep AI security in mind
When using these AI tools, don’t upload sensitive or company-specific data and documents unless you’re using a private enterprise version of the tool. Amber explains why.
“[ChatGPT, Copilot, and others] can and do use your data. They can use your data to train the models. They can review your prompts.
“Part of why this AI is so good is because there's a human in the loop, right? They're reviewing the answers to make sure they actually sound like a human. But somebody at Microsoft or OpenAI can look at your stuff, which obviously is a huge concern,” she says.
“There are offerings from the big players, [like] Microsoft, OpenAI, and Amazon, where you can kind of sandbox things. . . put it into your own environment and wall it off from the rest of the world so that information is not going back to OpenAI or whatever. And then you have control over what's happening with your data and with your model.”
The bottom line? “Don't upload company sensitive data,” says Amber. “Don't upload PII, all of that kind of stuff. You would want to be on the enterprise version [of tools] for that kind of stuff.”
AI is a partner for people, not their replacement
AI is a powerful tool for specific tasks. But humans are still part of the equation.
“If you think about driving and using a GPS, the GPS is telling you to do stuff: Take a turn, get off on this exit, that kind of stuff. But you as the human are still driving. And if you know that it's taking a wrong turn, you'll just not do that, right? [With AI], you need to still be the pilot,” says Amber. “Be a human. Be yourself.”
See Amber demo these tools and explore more use cases for AI in your organization. Watch the on-demand webinar now.
