Episode Categories:
Resources:
Join the discussion: Got something to say? Let us know on LinkedIn and network with other AI leaders
Upcoming Episodes: Check out the upcoming Everyday AI Livestream lineup
Connect with Jordan Wilson: LinkedIn Profile
Start Here Series in our Inner Circle Community: Join for free access
AI Language for Business Leaders: Essential Terms and Strategies for 2026
As artificial intelligence continues its rapid integration into business operations, understanding AI language is no longer optional—it's essential. Recent insights highlight the obstacles and opportunities faced by organizations when navigating AI adoption, especially the ever-changing jargon and foundational concepts that create barriers between technical teams and decision makers. This article breaks down the specific terminology, core business processes, and actionable strategies demonstrated in the "AI Without the Jargon" episode of Everyday AI to help leaders drive measurable results.
AI Adoption: Bridging the Capability Gap in the Enterprise
The expansion of generative AI has created a paradox: widespread usage, yet persistent gaps in understanding. Although tools like ChatGPT reached 100 million users in just two months, many business leaders still perceive AI as a simple chatbot. In reality, large language models (LLMs) routinely outperform human experts in blind studies, taking on complex workflows from research and personalization to document creation. This mismatch in perception and capability stems directly from three pain points: technology evolving faster than business training, lack of employee education, and a persistent disconnect between what users think these models can do versus their real capabilities.
Large Language Models: Practical Terminology Every Organization Needs
Business application of LLMs hinges on understanding a trio of core terms:
Prompt: The instruction or input provided by a human to an AI model, initiating the action.
Action: The model's internal sequence—summarizing, searching, drafting—performed within its technical architecture.
Outcome: The final result, which often surpasses human benchmarks for speed and quality.
Choosing the right model and mode for a specific business task is as critical as prompt engineering. Misunderstanding this selection process often leads to stalled pilots and wasted investments due to jargon overload and misaligned expectations.
Tokens, Context Window, and Parameters: Core Mechanics Explained
Every AI interaction translates human language into tokens—text chunks averaging four characters each. These tokens are processed within a finite "context window" (like a self-purging hard drive): when the model’s memory limit is reached, early information is silently dropped, sometimes sacrificing essential instructions. Business leaders must recognize that the structure and sequencing of their prompts directly impact outcomes.
Parameters reflect the "horsepower" of the model; more parameters enable more complex decision-making at higher computational cost. Notably, industry leaders are shifting toward smaller, specialized models for finance, translation, and healthcare, promising lower resource usage and faster deployment cycles.
Retrieval Augmented Generation (RAG), Embeddings, and Vector Databases: Business Data Integration
For businesses leveraging proprietary data, retrieval augmented generation (RAG) allows organizations to "ground" AI outputs in relevant documents, minimizing errors and hallucinations. This involves transforming text into embeddings—vector representations that facilitate efficient similarity searches inside vector databases. Proper chunking and overlap settings are crucial to ensure the AI understands and accesses logical, context-rich business information, which directly impacts accuracy.
Front-end users can benefit from RAG-driven enhancements even without deep technical involvement, while back-end teams should master the implementation of RAG, embeddings, and vector databases for custom solutions.
Agentic Models and Connectors: AI as Autonomous Workers
Modern LLMs have evolved beyond static assistants into agentic models: autonomous systems capable of decision-making, tool execution, data retrieval, and interaction with external services. These agentic capabilities, accessed via connectors (integrations with services like email, calendars, Salesforce, and cloud drives), simplify access to previously siloed business data and automate workflows.
Additionally, the emergence of Model Context Protocol (MCP) enables interoperability across AI platforms, streamlining multi-model workflows and reducing technical bottlenecks for business software integration.
Business Risk Management: Hallucinations, Prompt Injections, and Guardrails
AI risks are specific and persistent. Hallucinations—confidently delivered but false outputs—can derail decision-making. However, context engineering and the application of best practices in prompt construction and model selection can nearly eliminate hallucinations.
Prompt injections pose risks both at the model training level and during autonomous web browsing, potentially leading to unintended actions or security breaches. Leaders must implement robust guardrails: well-defined policies and filters combining human oversight and technical safeguards to limit misuse and enhance compliance.
Actionable AI Strategy: Sprint Planning, Measurement, and Continuous Evaluation
Effective AI adoption is rooted in precise planning. Leaders must continuously ask:
What specific business problem is being solved? What are the current costs?
What models, data, and tools are involved, and who authorizes these solutions?
How can progress be measured rapidly and responsibly, with transparent QA in place?
What are the hallucination, security, and operational risks?
Are observability and traceability practices ("receipts") in place to support expert-driven continuous improvement?
The AI landscape demands a modular, fast-paced approach: pilots must move quickly, with evaluation cycles aligned to model improvements—long implementations risk obsolescence before realizing ROI.
Staying Current: Terminology and Knowledge Management
AI vocabulary and definitions shift monthly—especially for concepts like artificial general intelligence (AGI). Maintaining a living "AI translation playbook" in organizational knowledge systems enables teams to avoid miscommunication and maximize technology value, ensuring everyone speaks the same language for decision-making.
Bottom Line: AI Language Mastery Drives Business Outcomes
The ability to translate AI terminology into actionable business outcomes is now the deciding factor for ROI, competitiveness, and risk management. Leaders who build foundational knowledge—prompt engineering, context awareness, data integration, and risk mitigation—stand to optimize workflows, cut costs, and leverage AI-powered tools far beyond basic chatbot use.
To future-proof strategy, organizations should invest in ongoing education, up-to-date resources, and community access to maximize AI literacy and connect technical advances to real business impact.
For further resources, educational materials, and top industry frameworks, refer to specialized AI inner circle communities and continuous-access learning platforms tailored for business application.
Topics Covered in This Episode:
- AI Jargon Barrier for Business Leaders
- Generative AI Basics and Lingo Breakdown
- ChatGPT, Claude, and Gemini Model Comparison
- Understanding Tokens and Context Windows
- Large Language Models: Prompt to Outcome
- Parameters, Model Power, and Cost Implications
- Retrieval Augmented Generation (RAG) Explained
- Embeddings, Vector Databases, and Chunking
- Agentic Models vs. Transformer Models
- AI Risks: Hallucinations, Prompt Injection, Guardrails
- Model Context Protocol (MCP) and Connectors
- Scaffolding for Complex AI Workflows
- AI Success: ROI, Risk, and Implementation Strategies
Episode Transcript
Jordan Wilson [00:00:17]:
One of the most challenging things about AI that no one talks about is not being able to talk about AI effectively. And I think that there's three main reasons for this. Number one, the tech changes faster than literally anyone can keep up with, even someone that talks about AI every day like myself. Two, companies don't train their employees. And three, it's kind of the summation of those parts. There's a huge gap between what people know about AI and what the models are capable of and can actually accomplish. And so much of this can be solved by talking about it, but therein lies the problem. AI is a mystery wrapped up in ever changing jargon.
Jordan Wilson [00:01:04]:
So we're kicking off volume two of our start here series where we're gonna karate chop the AI jargon in the jugular and break down the basics of AI language and lingo that every business leader needs to know in 2026. Alright. I'm excited. Let's get into it. Welcome to our start here series and everyday AI. So this brand new series is our essential beginner in advanced AI overviews to help you launch your year strong. So if one of your big goals is to double down or to just better learn AI, this new start here series is for you. Alright.
Jordan Wilson [00:01:47]:
So if you haven't already, please go to starthereseries.com. Yeah. New URL. So, there, you're gonna be forwarded straight to, sign up page to get free access to our inner circle community. And not only will you get, instant access to our prime prompt polished chat GPT prompt engineering course, but you will also get access to the start here series. So whether you're listening to this in mid January or sometime in March, we're gonna be putting the recap of every single start here series, episode inside the community there, inside the community there so you can listen to them all in order. Alright. And if you are brand new brand new, make sure you go listen to the first, series of this, which was episode six ninety one or volume one in the start here series, which was generative AI, how it works, and why it matters in 2026 more than ever.
Jordan Wilson [00:02:44]:
Alright. If you're new, everyday AI, it's a daily livestream podcast and free daily newsletter. But I've done, like, 700 episodes, and the most common question I get is where do I start? That's the whole reason that we're doing this series. So let's cut the jargon down, shall we? We're gonna get straight into it. So don't worry. This isn't gonna be a a super boring run through of random vocabulary words. One thing that I've realized that I kind of have a knack for, over the years is almost being a a a translator. Right? And, like, literally, there is there is an instance.
Jordan Wilson [00:03:21]:
I was working with a pretty big company, right, doing, like, 20,000,000,000 in in revenue. And I was sitting with their c suite, and and I was sitting with their their head of AI. And it seemed like sometimes these two sides weren't speaking the same language. And that's because there's so much jargon. And because, like I said, kind of that capabilities and understanding gap. And that's really what I've learned to do by doing this show over the last three years is connecting those two. So whether you are that highly technical person, you know, who understands all these terms, hopefully, this will help you understand where the nontechnical business leaders are coming from. And for our nontechnical business leaders, hopefully, this serves as a nice foundational look at the language and lingo you need to know to, you know, talk about it.
Jordan Wilson [00:04:07]:
Let's start. Start where we do. AI is not going anywhere. It's here to stay. Alright. We went over that in the first episode, but it's worth repeating. Chatt GPT, when it launched in November 2022, it was literally world breaking technology. No one had ever seen anything like it before, not just its capabilities, but also the sheer numbers.
Jordan Wilson [00:04:27]:
Right? They had a 100,000,000 users within two months after launch, and Chatt GPT now has nearly, 900,000,000 weekly active users. I think officially, still at the 800,000,000 weekly active users, and 40% of Americans use generative AI in just two years. For comparison, it took the Internet, like, five times as long to get that number of people actually using it. And I think, honestly, Chad GPT's overwhelming popularity is one of the reasons that we have this huge capability divide. Right? Do you know, as an example, you can give a single prompt to a large language model, like Claude or, Chad GVT, I think are probably the two best at it right now. And they can go research the web. They can, you know, personalize, things for you. They can grab, images.
Jordan Wilson [00:05:25]:
They can put together spreadsheets that you can actually download, PowerPoints that look really good. Right? All in one prompt. Right? Most people don't know that. Right? Unless you listen to the show, then you probably do. Most people think AI is just a chatbot. It's not. It is something that is completing, economically valuable work at a rate higher than human experts, and recent benchmarks show that according to blind studies. So essentially, yesterday's expert is today's beginner, and that is how quickly things change.
Jordan Wilson [00:05:58]:
Because your technical teams now are probably just spitting out things like tokens and, you know, MCP and rag and vector databases and Ralph Wiggins. Right? Like, it's this misunderstood jargon that just drives waste. It stalls pilots, and the vocabulary shifts faster than anyone can track. To get it right, you require the right words. Right? There is a reason why this is volume two in our, start here series. Because number one, we set the foundation. We said here is generative AI. Here's why it's not going away.
Jordan Wilson [00:06:37]:
Here's why you need to invest in it. And the next thing is you have to be able to communicate effectively about it. Alright? Because they require the right words. You can't evaluate vendors or manage risk alone. Right. You need to bring your technical teams together with your people leading change management, you know, with your HR, with your marketing, with your sales. You all need to be able to speak the same language. And, you you know, this is the new business terms.
Jordan Wilson [00:07:04]:
Right? You you know, MCPs and and, you know, a two a's are the new KPIs. They are. So you have to be able to translate AI phrases into things that move the needle for your company. This is ultimately, at least when we look at most large language models, and, assuming if you're listening to the show, you know, you're either using them on the front end or the back end or both. But this is all it comes down to. There's a prompt, there's an action, and there's an outcome. Alright. So humans prompt a large language model.
Jordan Wilson [00:07:38]:
Hopefully you're using our prime prompt polished context engineering way. Right? So again, go to starthereseries.com. Get free access to, our new and updated PPP course that teaches you how to get the most out of a large language model. But it starts with human prompting a large language model. Then the model will kind of go on with the task. You know, it'll summarize, search, or draft, and the model will read the prompt inside of its context window, and it works within the confines of its architecture. Right? So much of the outcome is not even decided by the prompt. It's decided by you using the right model, the right mode, for the right reason, for the right task.
Jordan Wilson [00:08:15]:
Right? But, essentially, the model reads the prompt the human puts inside of its context window. It you know, people, you need to understand things like context window. You need to understand tokens, and we're gonna be going over that today. Then the model retrieves whatever data that it needs to be a helpful assistant and respond to your query. Right? It's now models are agentic by nature. Right? Agency, they make decisions on their own to accomplish a goal, which includes using different tools in their arsenal, retrieving data on their own, you know, exploring multiple pathways and going back and saying, oh, this wasn't the right one. And then they output a result. And then the humans hopefully observe what the model did.
Jordan Wilson [00:09:00]:
They trace, the paths that it took. They verify the information, and then they use the outcomes. That's essentially. Right? If if you're like, well, how does AI even work? There you go. Right? From prompt to action to outcome. Are you still running in circles trying to figure out how to actually grow your business with AI? Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on GenAI. Hey. This is Jordan Wilson, host of this very podcast.
Jordan Wilson [00:09:35]:
Companies like Adobe, Microsoft, and NVIDIA have partnered with us because they trust our expertise in educating the masses around generative AI to get ahead. And some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use GenAI. So whether you're looking for chat g p t training for thousands or just need help building your front end AI strategy, you can partner with us too, just like some of the biggest companies in the world do. Go to your everydayai.com/partner to get in contact with our team, or you can just click on the partner section of our website. We'll help you stop running in those AI circles and help get your team ahead and build a straight path to ROI on GenAI. And this is the foundation for every AI term out there. Alright? And and and we did cover this last one. So this is the the thirty second, refresher from volume one in this series.
Jordan Wilson [00:10:37]:
So generative AI systems, they generate something. Right? Whether that's text, images, code, video, music. Right? Generative AI is not just chatbots. It is now everything. Large language models which fall under, you know, kind of the generative AI umbrella. They generate responses by predicting the next word, or token in a sequence. And they do that, you know, whether they're using, the GPT technology, which is what a lot of today's most of today's large language models are built on top of. Some of the, you know, image models use diffusion technology, which is essentially they kind of fill in the dots.
Jordan Wilson [00:11:13]:
Right? Where large language models use more kind of next, next word or next token prediction. And most current AI products that you use, right, they're just wrapped around these other products. Right? So essentially, you have three big AI labs that dominate. Right? You have Anthropic, you have Google, you have OpenAI and their models, Anthropic Claude, OpenAI's, Chad GPT models, so the GPT models. And then you have the Google Gemini models. And you're like, oh, wait. What about Microsoft Copilot? Well, they did just start making some of their own models, but for the most part, Copilot just uses, the OpenAI models. They use a little bit of, anthropics models as well inside of their Microsoft Microsoft Copilot three sixty, or three sixty five.
Jordan Wilson [00:11:59]:
Sorry. Yeah. Left out. That was Microsoft Xbox three sixty. So the, you know, a lot of this other software that you use, most people don't know this. Right? They're not out there creating their own models. They're not competing with OpenAI. For the most part, you know, if you're using any AI product and there's literally hundreds of thousands of AI products out there, right, pieces of software, For the most part, they're using one of those three, probably.
Jordan Wilson [00:12:26]:
There's other smaller, AI labs, but for the most part, those are the three that dominate. And, essentially, these models that we use, they don't technically understand our words. Right? The human who inputs the words, those words get translated to tokens. And the model itself, it doesn't think in words. It thinks and produces, in tokens and then converts it back. So there's almost like a, you know, it's kinda like when you go into a new country, you have to convert your money, and then when you leave, you gotta convert it back. It's the same thing. Right? When I'm giving something to a large language model, it's con, it's converting that to tokens, right, which helps it truly understand what I'm trying to say, because it doesn't understand words.
Jordan Wilson [00:13:03]:
It converts everything to tokens. Right? Which is why sometimes, you know, how you communicate with a model in the same way of how you communicate with your teammates to get the most out of AI, how you communicate with the model is very important. Right? An example that I've given before, the word just. Right? Context matters. The word just can be tokenized by a model. Just the single word just. Seven different ways the last time I checked. Right? Something can be just.
Jordan Wilson [00:13:31]:
You can use just as a qualifier. I just wanna talk. Right? There's, like, so many different meanings for words. In the same way that when you're young and you're learning whatever language that you learn, you learn, oh, different words can mean different things, and it's the words surrounding, those words that give context and help us understand. So large language models kind of the same thing, but, instead of the words around, it's the tokens around. So let's talk about those tokens and context windows. This is important. Tokens are essentially text chunks.
Jordan Wilson [00:14:03]:
Right? So they're about four characters each. So a a word could be anywhere from one to, you know, up to a handful of tokens, and they're used to measure input and output length. So when you have a conversation with a large language model, like I said, everything gets put into tokens. And I'd like to think of, you know, an an easy way to think about it is a hard drive. Right? When you except let's think of a hard drive worked a different way. Right? Right now, if you're trying to download a file and your hard drive's full, it just says, oh, it's full. A context window is like a hard drive, but the difference is it's automatically gonna keep working. And instead of saying, hey.
Jordan Wilson [00:14:42]:
The context window is full. It's just gonna kick out the first thing that you put in. Right? And a lot of times, the first thing that you put into a large language model is the most important. Right? It's saying, hey. Here's the most important information. Here's what I want you to do for me. Then if you keep working with that, you know, keep spinning tokens back and forth with that large language model, eventually, the hard drive is gonna get full, except it's not gonna say hard drive is full. It's just gonna forget the first things you said.
Jordan Wilson [00:15:07]:
That's a context window. And then parameters. Parameters are important to understand about large language models as well. Essentially, these are the neural network connections of a model. The more parameters it has, the more power. Right? You can think of parameters, kind of like horsepower of a car. Right? There's more capability of these models that have more parameters, but there's higher computational costs in resource usage. Right? So, large large language models, I think, are actually getting smaller.
Jordan Wilson [00:15:41]:
A lot of times, companies don't say how many parameters there are, but, you know, kind of the last widely reported, one we knew about was the GPT four, architecture from OpenAI that had about 2,000,000,000,000, reportedly 2,000,000,000,000 parameters. So that's a lot. I've always said for a long time, large language models are gonna get smaller. And in the long run, I think most people are gonna be using more small language models. We've even seen that a ton in the last week or two. A lot of, specialized models from the big companies, right, for things like translation, medical, finance. Right? The big AI labs are starting to put out very, very small models. I do think the future, you know, there's gonna be, you know, OpenAI will have thousands or tens of thousands of models, and they'll be spinning them out faster than we can even keep up with.
Jordan Wilson [00:16:29]:
Right? But anyways, parameters is capability. Now let's talk about some dorky things. Alright? And if you are a beginner, forgive me, because for our more intermediate, people or people who just need a reminder, we really have to understand data. Right? So rag is probably a business term that you should know. And I think rag, traditional rag as we know it, is going to very much change. Right? I'm not gonna say rag is going to away, but rag is retrieval augmented generation. And that's where a model can retrieve relevant documents, and then it grounds its responses using that retrieved content. Right? So when as an example, you know, before, ChattGPT came out, the GPT technology was available, and many businesses used it.
Jordan Wilson [00:17:22]:
And then they essentially, used retrieval augmented generation, to kind of wrap OpenAI's responses in their business data. So embeddings are also something important to to know. So these embeddings are it's kinda like group of data or turns, text into vector embeddings, allowing for efficient similarity search in a vector database. Alright. This one, is a little bit more difficult to talk about. Right? I'll probably, in in the start here series, in our community there, I'll probably put a couple graphics. But it's it's a little hard to explain on a podcast without doing a deep dive. Right? But essentially, text is split into smaller segments or chunks before the embeddings happen.
Jordan Wilson [00:18:07]:
And then the chunk size and the overlap settings control how it's done. So good chunking splits information logically, so answers include the right context and details. And bad chunking breaks up context or combines unrelated info leading to wrong or incomplete answers. So this is pretty this is, much more relevant for companies that are building with a p well, building, with large language models on the back end. Right? So when I say back end and front end, front end, if you're using a large language model on the front end, you're going to chadgbt.com, and you're using that chadgbt interface. That's a front end user or claw.ai. Right? A back end means that you maybe have a team of developers or even now nontechnical people are obviously using large language models on the back end, but you're using, the company's API. And so, essentially, you're using their technology.
Jordan Wilson [00:18:56]:
And when you're doing that a lot of times, then you might be working with RAG, you know, these vector databases, chunking, all of that. So if you're a front end user, you don't have to worry about that as much, but you should know what RAG is. Right? RAG is what makes, those queries, kind of as they go off to the large language model, makes them much more accurate. It reduces the the the likelihood and the prevalence of hallucinations when you are first interjecting, your company's data, your secret sauce, how your company thinks. Right? All that important stuff that helps new hires and, you know, people that have been at the company for a very long time make good decisions. It also helps large language models make good decisions. And large language models are obviously changing. You know, they started as a helpful assistant in that little chat window, and there is nothing else they can really do.
Jordan Wilson [00:19:48]:
Now models are agentic by nature. Alright? I like to say we have our old school transformer models, and then we have our newer reasoning models. Even though these reasoning models are still built for the most part on the transformer technology, the generated pre trained transformer, GPT technology. Right? I like to say we had more of these, you know, old school transformer helpful assistant chat bots. Now we have more autonomous workers. Right? These models now, specifically, you know, anything from Claude in the 4.5, family. So Opus 4.5, SONNET 4.5, Haiku 4.5. And then with Google Gemini, all the Gemini's new models, you know, are hybrid or can think.
Jordan Wilson [00:20:34]:
Right? So they can either give you that kind of next token prediction or they can think about it. Right? They can be agentic in nature or sometimes they think if they need to think about it or they just think and they're like, alright. I'm gonna spit out an answer. And then with OpenAI's models for the most part, aside from the default one. Right? But, you know, make sure that's why I always say if you're using, a model, make sure you're using a model that thinks. Because these are models that can call tools. Right? They can, decide, hey. This user is asking me about something that might be a little old in my training data or the data that the model is built on top of might be might give older outdated or wrong answers.
Jordan Wilson [00:21:13]:
So an agentic model can decide, I'm gonna call a tool. It can trigger functions like search, you know, going through your email, letting the, agentic model interact with external, services. Right? You have these things called connectors, which, you know, it's kind of like rag mini. Right? Retrievalog mini generation, but a very simple version. Right? It used to take, a lot of smart engineers and usually, six, seven, eight figures to get rag. Now you can get a very simple version in a couple of clicks. You know? Most large language models on the front end call those something like connectors or apps. Right? So you can connect things like your like your email, your calendar, your, you know, your SharePoint or your Google Drive, your OneDrive, things like that, Salesforce, Slack.
Jordan Wilson [00:22:00]:
Right? And then all of that, invaluable, right, like all of that data that your company needs and uses that you're like, man, two years ago, it was so hard to to to work with that. Now it's there almost automatically, right, with these connectors. You also have MCP. This is the model context protocol. And this is, a technology that was first, invented and popularized by Anthropic, but the rest of the industry has adopted it, which is great. You know, one thing about the AI industry, see like, it's weird. They all kinda work together. Right? So even as an example, I was just talking about connectors.
Jordan Wilson [00:22:38]:
If if you're a Google organization and you use Gmail and Google Drive, you can technically use those connectors inside OpenAI's products. You can use it inside Microsoft products, inside clause products. So in the same way that MCP is kind of this it's kinda like, you know, how the Internet talks to, you you know, websites can talk to each other with APIs. This is kind of how AI tools can talk with, to each other with MCPs. So the model context protocol is definitely the most popular one. There's a handful of other ones that are, widely used and widely adopted. But for the most part, you know, these models have essentially languages that they can talk to each other. So you don't have to, you know, send it from an AI, system to a website to use the website's API to go to another website to then send it to a model.
Jordan Wilson [00:23:27]:
Right? So it's just a direct connection. The MCP connection allows different AI models to talk to each other or different AI tools. And then scaffolding. So if you've ever heard of the word scaffolding, I say it a little. I try not to, just because I don't wanna, you know, confuse any people. But that essentially defines the structure that agents use to chain together multiple steps or tools for more complex workflows. So, you know, sometimes, honestly, like, the models themselves can all be fairly similar. Right? If you look at the benchmarks, they're, you know, point 1% from, you know, each other.
Jordan Wilson [00:24:03]:
A lot of times, it's the scaffolding. Right? It's it's how the model actually works and how it's able to successfully or unsuccessfully use all these different tools that it has at its disposal. So, the large language models have largely gone from helpful AI assistance that were powered by older, slower models that hallucinated in a lot. Now they're the models themselves are agentic by nature and they can use tools and you can connect to your business's dynamic data. But there's risks, and you need to know them. Hallucinations, probably a word most people have heard of, but it's essentially a lie, right, or false statement or half truth that's put out there very confidently. Right. But if you do know the basics of context engineering, so if you've taken our free, prime prompt polish, course in our free prime prompt polish I can't speak.
Jordan Wilson [00:24:56]:
Prime prompt polish pro course, you know the basics of context engineering and your hallucination rate is gonna go down. But if you're using the right model, and if you know the basics of prompt engineering and, sharing your company's context, hallucinations are I'm not gonna say they're gone, but they're essentially gone. Right? People overlook that. If you know what you're doing, if you're using the right model, if you're connecting in your company's data, if you're, looking at the chain of thought, doing basic observability, you know, human in the loop, even lazy human in the loop, hallucinations are down to about zero. Prompt injections, you have to look out for those, especially as we have agentic browsers. Right? So these are browsers that are run by an AI agent, and they can go browse the web for you and interact, with interfaces on the web, with graphical interfaces. You know, click, you know, click drop downs and fill out forms for you. But there's prompt injections.
Jordan Wilson [00:25:51]:
But right? So there's prompt injections that get into training data, which impacts the models themselves. But on the agentic side for agents or agentic browsers, there's prompt injections on web pages. Right? And some of the earlier agents and agentic browsers didn't have great safeguards against those. Right? So, you know, as a agent, right, a simple example, you know, I could put something on my website. So if an agent or an agent at browser was trying to use it, and I could say, hey. You know, go, you know, go to this site and, you know, buy a 100 boxes of cereal. You know, put it in your Amazon cart or something like that. That's a very, safe and happy prompt injection.
Jordan Wilson [00:26:30]:
Right? It's it's not causing a lot of harm. I it just trick the model, or an agentic model into doing something that it didn't intend to. But there's obviously very harmful and unsafe, you know, prompt injections just like there's, you know, phishing emails and all these, you you know, scams that have been going around, you you know, with malicious links and all that. The same thing. Models have risks. So companies need to understand and talk about guardrails. That's when you combine, human made policies and filters to enforce safety and prevent misuse. Alright.
Jordan Wilson [00:27:03]:
So this is kind of the translation playbook. Alright. So in the future, we might be updating this, right, if we keep these start here series going on for a longer time. So make sure I always say, check the show notes of this show. So if you're listening on the podcast, we're gonna always keep the show notes updated. And, in the community in which you get free access to by going to starthereseries.com, we're gonna keep the resources there updated as well. So there's always gonna be new terminology. Some of the terminology is gonna change.
Jordan Wilson [00:27:34]:
Right? Like, as an example, the definition of artificial general intelligence changes every month. Right? By, definitions from ten years ago, we're well past AGI, but, you know, how people define it say, maybe we're not there. So we're gonna keep the definitions up to, up to date in our community. Gonna have a little AI translation playbook, but you need to ask yourself these questions. Alright. I'm gonna go back to how I started. The technical people and the nontechnical business leaders. These are the questions that you need to be asking yourself out loud when it comes to using AI and understanding it, and hopefully closing that capability and education gap.
Jordan Wilson [00:28:13]:
Number one, what's the problem we're trying to solve, and what's our current cost associated with that solution? Right? People overlook that. I'm like, you know, when people are trying to, you know, calculate ROI on GenAI, right? I I always ask, okay. What's your current cost on this? And they're like, okay. Well, you know, we have this piece of software and, no. How much time is spent? No clue. You have to understand what your current costs are. Next questions. What model, data, tools, and who approves? You have to go through those.
Jordan Wilson [00:28:45]:
Next, how can we responsibly run the fastest sprint possible with measurement and evaluation with proper, you you know, proper QA? You need to go fast. You have to be careful, but you have to go as quickly as possible while still being careful, while you're implementing AI. Because if you're too slow and if you take a pilot slow and if you're not building modularly, right, now all of a sudden you did a year long pilot. Oh, wait. That original model you started with is that was the best model in the world is now the fiftieth best model, and it's bad. And now you're putting yourself at a disadvantage. You can't treat AI implementation in the same way that we've treated tax, tech implementation for the past few decades. You need to ask what are the in the hallucination risks, the security risks, and the cost.
Jordan Wilson [00:29:35]:
And then last but not least, you always need to require receipts. The observability, traceability, expert driven loops. We don't need human, lazy human in the loop, logs defining the scope and getting your evaluations first. You have to understand what you're going after to be able to measure it, and go and, you know, understand the evaluations in the process. Alright. So now that's it. You now have the foundation for the rest of the start here series. Alright.
Jordan Wilson [00:30:03]:
We went over what the heck is generative AI, how do large language models work, and why it's the future of work. And now you have the language and the lingo down. Alright. So what I want you to do, sit down with your people. You need to get a thirty day plan. You need to be able to sprint. You need to understand how you're gonna deal with mistakes. And then later you can go into the rag, deep dive, ROI, security scale, all this stuff.
Jordan Wilson [00:30:24]:
But you have to set here's where we are. Here's where we wanna go. Figure out what's gonna happen in between, but you have to keep up to date with the language and the lingo. Otherwise, your plan is gonna fall apart. Alright. One way it's not gonna fall apart, go listen to episode six twenty eight, very much so related, to this episode. So that is what's the best large language model for your team, seven steps to evaluate and create ROI for AI. Alright.
Jordan Wilson [00:30:53]:
I hope this was helpful. Now you know the AI without the jargon, the AI language, every business leader needs to live by in 2026. Alright. If this was helpful, please go to starthereseries.com. That's gonna take you and give you free access to our inner circle community where not only can you go and check all of our start here series, on the web. Right? Or if on the mobile app, that's what you wanna do. But you're gonna be signed up for our free daily newsletter as well, and you can go take our free prime prompt polish, context engineering course. Alright.
Jordan Wilson [00:31:30]:
Thank you for tuning in. I hope to see you back tomorrow and every day for more everyday AI. Thanks, all.