Episode Categories:
Resources:
Join the discussion on LinkedIn: 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
Codex Sites and AI-Native File Sharing: Specific Business Insights from Everyday AI
The business landscape is rapidly shifting from static, version-laden file sharing to dynamic, AI-powered workflows. The latest episode of Everyday AI presented a live, hands-on demonstration of OpenAI's Codex Sites—highlighting exactly how this feature addresses long-standing bottlenecks with file versioning, team-wide data access, and actionable business intelligence.
This article breaks down the specific advantages and operational takeaways for business leaders seeking practical AI adoption, based directly on the workflow and real-world use cases explored in the podcast.
AI-Driven File Sharing: Moving Beyond Static Documents and Version Chaos
The conversation focused on two pinpointed problems in enterprise environments:
The impracticality of sharing custom-coded tools or dashboards that require constant manual updates or risky, unsecured sharing;
The persistent inefficiency of static file sharing, where deliverables—PowerPoints, PDFs, spreadsheets—quickly become obsolete and pile up as “final_v2_really_FINAL_v3” on desktops and inboxes.
A key theme that emerged was the transition from static work artifacts to living, constantly updated, interactive resources. Codex Sites specifically targets these obstacles: it auto-generates shareable, live dashboards and tools from a single prompt, rendering static documents and chaotic versioning largely unnecessary. Version control becomes irrelevant because the information, analytics, and outputs in a Codex Site are always current and accessible based on user permissions within an enterprise business plan 02:13.
Codex Sites Business Applications: Detailed Use Cases and Live Demonstration
Several points were raised, including a step-by-step demonstration of building an internal "Audience Momentum Dashboard" using Codex Sites. The process clearly illustrated business-specific functionality:
Codex Sites can ingest data scattered across disparate platforms—such as an email newsletter provider, podcast analytics, and online community platforms—by connecting directly via skills and browser automation 08:08.
Unlike many other agents or dashboards, Codex Sites is capable of actively navigating browser tabs, pulling filtered and deeply-buried metrics, and performing on-the-fly data analysis, such as calculating 21-day performance averages or spotting above-trend podcast episodes 13:19.
The resulting dashboard pulls together over 50 unique data points—requiring hundreds of manual clicks and navigation through 30+ separate screens if done manually—into a consolidated, shareable, interactive interface 34:06.
The site can be set to auto-refresh at specific intervals (hourly, daily, etc.), ensuring executives and team members always see up-to-date numbers and trends without repeated manual intervention 11:35.
These capabilities allow operators and analysts—without deep technical skills—to build applications directly inside OpenAI’s workspace, eliminating the need for deployment pipelines, engineering tickets, or stand-alone, versioned files sent as attachments 19:03.
Codex Sites and Team Collaboration: Internal Security and Application Hosting
One concept discussed was the platform’s approach to file sharing and collaboration:
Codex Sites auto-generates internal URLs, requiring authentication via the organization’s AI platform credentials. This secures data access and enforces company-wide data governance without third-party apps or external file sharing 24:24.
The platform supports live role-based access, so only authorized team members view or interact with dynamic business intelligence. This solves for both speed and security—non-technical leaders access only the insights needed, without raw access to sensitive internal databases 12:16.
Codex Sites currently only supports sharing within the same business or enterprise plan infrastructure. While this limits immediate external sharing, it ensures internal teams operate in a high-trust, high-velocity environment—a crucial requirement in regulated or data-sensitive industries.
Codex Sites Versus App Builders: Specific Strengths and Limitations
The discussion explored whether Codex Sites could displace full-stack app builders like Replit or Lovable. Several key distinctions arose:
Codex Sites trades broad feature depth (editors, standalone database management, etc.) for seamless workplace integration and direct access to existing Codex data and automations 20:24.
For teams already using ChatGPT and Codex, Codex Sites means “building in context”—the same workspace houses conversations, data requests, automations, and now, deployable dashboards and collaborative tools. For many companies, this reduces complexity and onboarding time.
While Codex Sites is not as robust for custom, large-scale apps, it excels in building lightweight, actionable tools and dashboards—especially those enabling nontechnical teams to access and use operational or market insights rapidly 21:19.
AI-Native Artifacts: The Practical Future of Reporting and Deliverables
A key theme that emerged was the diminishing utility of static deliverables in high-velocity businesses. With AI-native tools, like Codex Sites, work products become living artifacts—always reflective of the latest data, updates, and stakeholder input.
This shift includes:
Eliminating “snapshot in time” reporting and instead presenting boards, planners, or repositories where metrics and tasks update automatically as data changes.
Supporting faster, more accurate decision-making by ensuring stakeholders see one version of the truth.
Encouraging faster feedback cycles, where stakeholders can use new annotation features for rapid changes to dashboards or visualizations 22:23.
Considerations for Deployment: Limits and Best Practices
Business leaders considering Codex Sites should weigh several operational realities:
The platform is currently available only for business and enterprise ChatGPT accounts; team members must log in for access, ensuring company-level data governance 06:13.
Codex Sites are live by default; it’s advisable to create ‘draft’ versions for iterative work or experimentation, rather than tinkering with production dashboards during critical business hours 24:13.
The practical benefit lies not in exporting complex prompts, but in leveraging a library of proven, stackable use cases. Building with OpenAI’s documented skills and integrations is more efficient than starting from scratch 37:18.
Conclusion: Live, Single-Source Insights and Accelerated Workflow
Codex Sites bridges the traditional gap between manual dashboard creation and instant, AI-native business workflows. This tool is already enabling marketing, analytics, and operations teams—especially those without dedicated engineering support—to move faster, waste less time reconciling file versions, and act on live marketplace signals.
For organizations ready to reduce friction, improve clarity, and move away from static reporting, Codex Sites is not just a technological upgrade. It is a logic-driven step toward more secure, self-serve, and business-relevant applications built directly from prompt to output—enabling teams to focus on actions, not artifacts.
Topics Covered in This Episode:
- Codex Sites vs Static File Sharing
- Live Dashboards and Automated Workflows
- Building Internal Apps With Codex Sites
- Real-Time Data Integration in Codex
- Agent Layer and Role-Based Access Control
- Codex Sites vs Replit, Lovable, Bolt
- Dynamic Business Insights and Collaboration
- Codex Sites Secure Team Sharing Limitations
- Automations and Custom Skills in Codex
- Future of AI Native Business Tools
Episode Transcript
Jordan Wilson [00:00:17]:
I think there's a gap between the reality of Vibe coding yourself useful tools and dashboards and the practicality of the enterprise reality of sharing it with your team. I mean, yeah, you can vibe code yourself some autonomous dashboard thingy, but what good is it if it's using old data and requires human duct tape or if your team can actually take advantage of it in a secure way? That's problem number one. Problem number two, the file sharing nature of big business is still stuck in the pre AI era. I mean, you have static files that are shared back and forth via email or versions of a deck that are stacked up on a desktop that are final one, final two, final three, really final. Right? That doesn't exactly scream AI native. So OpenAI's new codex sites feature sets off to solve both of those problems. But I get the issue. You've probably had, like, zero time to use the new codex sites.
Jordan Wilson [00:01:14]:
There's huge drops every day. You wanna go use that new Claude Mythos Fable five model. You wanna check out all the new codex plugins. You're itching to use the new Copilot scout. You don't have time to spend hours with codec sites. Don't worry. I did the hard work for you, and that's why we do this thing on Wednesdays called putting AI to work on Wednesdays. So we're gonna see if codecs is really a replet and lovable killer, and it if it'll change the way that we share files forever, or if it's just another drop in the AI features bucket that will be forgot about in a couple of weeks.
Jordan Wilson [00:01:48]:
Well, there's only one way to find out. Let's get to it and do this thing live. Alright. Welcome. If you're new here, my name is Jordan Wilson, but let's talk about the big picture. Static work is now becoming live software. Whether you've seen this or not, the business world is shifting. It's getting easier and easier to make living, breathing versions of your company's most important work.
Jordan Wilson [00:02:13]:
And I think we're obviously getting far away from the days of just sharing old and, different version files to sharing information and intelligence for your company that updates automatically. Codec sites shows how nontechnical teams can build and ship usable apps and work, you know, kind of these different workflow, you know, dashboards, all these different things just from a single prompt. And I think the real shift shift is not just coding faster, it's shifting your mindset to work differently. So on today's show, stick with me for the next twenty five minutes, and you're gonna learn why static documents, decks, and version file chaos may be finally nearing their end. You're gonna know how any team member can build and ship a live app from a single prompt today. You're gonna see whether OpenAI's codec sites is a real replant and lovable competitor or something entirely different, and you're going to know what the agent layer inside of codec sites means for how your team runs its tools. Let's get to it. Welcome to Everyday AI.
Jordan Wilson [00:03:16]:
My name is Jordan Wilson. This thing's for you. It's your daily livestream podcast and free daily newsletter helping business leaders like you and me keep up with the nonstop avalanche of all these new AI features, these new AI tools. I take the time to run them all. I tell you what's important, what's not. You use that information to become the smartest person in AI in your company, and you grow your business and career. Sound like a good trade off? Sweet. All you gotta do is make sure to go to our website at youreverydayai.com.
Jordan Wilson [00:03:43]:
Sign up for the free daily newsletter. That's your cheat code. And we're gonna be highlighting not just the big portions of today's show that you need to know, but also all the other AI news you need. No filler, no BS, just the facts, Jack. So codec sites, are they gonna be the next big wave? Right? Is this gonna be something as big as, like, Claude Code was at the end of twenty twenty five and early twenty twenty six? I think the potential is definitely there, but you're gonna have to decide, with me. And, yes, on Wednesdays, we do these more hands on demos. Sometimes they go really well. Sometimes they blow up my face.
Jordan Wilson [00:04:18]:
So we'll see how this one goes. But this might be one of those ones you might wanna catch the video version of this, so make sure you go to your everyday ai.com. We have the video versions on there. Also, we did a two part series of a codex breakdown. So if I'm showing you something and you're like, wait. What is this codex thing? I already spent a ton of time, probably thirty plus hours putting these two shows together. So make sure you go listen to episode seven eighty one. That was part two in our two part series of, kind of advanced codex for newbies.
Jordan Wilson [00:04:51]:
Also, put together one of those extremely valuable things that you you're gonna see other people selling this stuff for, like, $500. I'm giving it away for free. So we put together a codex site stack. This is 50 real use cases for, how you can do exactly what we're gonna be showing you how to do on today's show. So all you have to do is go repost this episode on LinkedIn. So if you're watching live, that's easy. Two clicks. If you're on the podcast, we always put the link to today's LinkedIn show, in the show notes.
Jordan Wilson [00:05:21]:
So make sure you check out the show notes. Go click that, button, repost it, and I'm gonna send it your way. Alright. Let's start live. What could go wrong? My gosh. This one actually might be, a little bit crazy, because it's gonna be a little tough. Alright. So what could go wrong? Sharing a live agentic system.
Jordan Wilson [00:05:42]:
Hopefully, I don't dox myself. That won't be good. Alright. Let's see how it goes. So I am in codex. Codex, if you don't know, it is a much more powerful version of chat g p t that's made for the desktop. So it has access to read and write all the files that you give it access to. It It can use the terminal.
Jordan Wilson [00:06:02]:
It can control your browser. It can control your whole computer. So it's like if chat gbt was more of an actual human using chat gbt, that's what codecs is. So I have a prompt here, but let me start here. Sites is not available for everyone. That part's a bummer. So it is only right now available for business customers. So I am on my ChatGPT business plan, and that's an important, differentiator here because the limits are not as good as my max plan.
Jordan Wilson [00:06:28]:
So I don't even know. I did one test of this before. I just barely snuck it in with 97%, using 97% of my limit. So I am trying to do something somewhat complex on a, on a plan that is not my my daily driver. Right? My daily driver plan is my my, you know, $200 a month plan, which I could do this all day, except the new codec sites feature is not available. So it's only available, right now at least on Teams plan. So the business plan, the enterprise plan, etcetera. So I have my prompt here.
Jordan Wilson [00:07:00]:
I'm gonna click enter. Alright. Hopefully, this works. If not, I do have a finished version ready to show, but I'm hoping we can do this thing live and watch it work. So, I'm gonna explain my prompt as I go, but podcast audience, everyone else, I just went into codecs. I had the prompt ready on my screen. Hit enter. Let me tell you just very quickly what we're gonna do.
Jordan Wilson [00:07:21]:
I'm not gonna read this prompt verbatim. Alright. But so I am telling it to use sites. So I'm saying, build me a simple, elegant internal dashboard for everyday AI called audience momentum dashboard. Alright. And then I'm gonna quickly explain some of the different things that I'm having it do and also why. And you can see here for our livestream audience, there are certain things in different colors. So I'm using two different skills.
Jordan Wilson [00:07:50]:
So, a skill is essentially a repeatable workflow. And then I'm telling it a couple of times to use, oh, no. Actually, just one time to use a different app. Okay? So I'm having it control Chrome as well. So I'm having it use a set of prepackaged skills, and I'm also having it use Chrome. Aside from that, I'm also having it pull in data from two different MCP servers. So if you don't know what m oh, let me move this. There we go.
Jordan Wilson [00:08:17]:
I knew this is gonna shoot up. Alright. It's already starting to work. So you guys can watch it. Maybe we'll be able to watch it work, live here a little bit. Alright. So yeah. It's actually running live.
Jordan Wilson [00:08:29]:
It's it's, you know, using my my browser right now. It actually started to take over, my my StreamYard tab, so I had to pull it, into the other, screen that I'm sharing. So I'm sharing my full screen, so things might be flashing, on on and off a lot. Anyways, what I'm having to do is something that I would have to do manually. And I wish I had the time to do this manually, but I don't because it takes a lot of time. So I'm having it create this audience momentum dashboard, which for me is something that takes a lot of clicks. It takes going into a lot of screens. But I do have it connected to my MCP servers for my email newsletter and the MCP server for my circle community.
Jordan Wilson [00:09:10]:
That's our platform that we have the everyday AI inner circle community. There's, like, 3,000 members who go on there, network with other people, talk about all kinds of AI stuff. It's awesome. It's free. You should check it out. Anyways, to get the type of data that I need from both Beehive, which is my email newsletter provider, to get it from Buzzsprout, which is my podcast distribution platform, and to get it from Circle, to get just the minimum level. Right? It's gonna take me hundreds of clicks in usually a couple of hours. Alright? Because sometimes you have to really dig down and go really deep.
Jordan Wilson [00:09:46]:
Alright. So that's essentially what, what codecs is doing now. Alright. So it's using a couple, a couple of my different tabs here. So you'll see it actually is is, using this, tab in my browser, right now, going through some of my podcast stats. Right? So I'm not moving this. So the computer used inside codex is is goated. If if if you've ever used the clod one, it makes the clogged one look like a preschooler, and this one has, like, 30 PhDs.
Jordan Wilson [00:10:16]:
So it has its own cursor. It operates at both the operating system and the user interface level, so it's really cool. So, essentially, what I'm gonna have going here is it's gonna be pulling all of these stats that take me so long to compile. Because here's the reality and what I want you to think of your own problem when you're doing this. Software is both great and it's terrible. Right? Because a lot of times to get the most important things, some of this information might be available on your home page of whatever dashboard you're thinking of. But sometimes you have to apply all these filters. You have to click around.
Jordan Wilson [00:10:50]:
Right? You have to refresh a bunch of things. You have to stack different, you know, things to get the exact metrics or the exact outcome you need. And some, you you know, some pieces of software that are a little more enterprise, you might be able to save some of these metrics. Some, you can't. So in my case, right, to get the level of intel that I need. Right? So if I'm like, okay. Should I be focusing more on, you know, teaching people about tool calling or teaching people about connectors, talking more about AI and education. Right? I don't have time to go through dozens of web pages and all of these different metrics.
Jordan Wilson [00:11:26]:
So that's what I'm having codex hopefully build me. Here's the other thing in the like, what I wanna call out right now. This is something that will be able to live. Right? When you pull in all of these things through codec sites, I can set up an automation that refreshes it every hour, every day, and I don't have to touch a thing. So I can just build this once inside of codec sites, then I can run an automation that just gets it going every single day and updates it every single day. And the great thing is, I can then share this with my team. They don't need access to any of those other tools. Right? This is when I had a a marketing company.
Jordan Wilson [00:12:04]:
This is the main thing that I would do. This is, like, one of my main jobs is I would have to pull together, different stats for different clients, other people on my team, and not everyone had, you you know, access to all of these platforms. Right? When you talk about role play, role based access control, you know, maybe you don't don't want every single person in your organization to have access to, you know, your email list with a 100,000 customers, but maybe you want them to be able to tap into the insights that, you know, that data has. But it might just take way too long. Alright. So just real quick. A couple other things I'm asking it for. So I am using, the sites feature, and I'm gonna show you guys a little bit more about that here in a second.
Jordan Wilson [00:12:46]:
I built my own skill called the buzzsprout podcast stats. So I'm having it use one of my skills that I've built. I'm I'm telling it to use Chrome with one of my certain Chrome profiles, and then I'm using a design taste front end skill. So I'm using one of my own skills, one of the OpenAI skills. I'm using computer use. I'm using Chrome. And then I'm also pulling from multiple MCP servers, that I've already set up as well. Those don't show show up as a different color.
Jordan Wilson [00:13:14]:
I'm just telling it to go in and use those things. And I'm having it do a lot of data analysis on the fly. Right? I'm having it run a, essentially, a twenty one day podcast stats sprint. So it's looking at twenty one different days of data. It's creating averages. You know? So, hey. Here's your average working Wednesday show. So I know if I put one out, if this one is gonna be above the pace below the pace, so I can look.
Jordan Wilson [00:13:40]:
Is there good evergreen shows from months ago that are still performing above that twenty first twenty one day average? So not only am I just asking for, you know, data from these different platforms, I'm asking you to do a lot of data analysis work as well. So I don't have to go and compare these things to averages, to totals, and all of that. I just have a super smart, model, g p d five five. Do this for me. Alright. So I'm running through here. I'm checking on the status. It's still going.
Jordan Wilson [00:14:09]:
It's still building. All good. So let me just quickly read a little bit about OpenAI. This is from their website, saying, here's what codec sites is in their own words. They're saying, starting in preview for business and enterprise customers, codecs can now create and share interactive hosted websites and apps. Sites are a new kind of canvas for your ideas. Codecs can take your ideas, analysis, and plans, and turn them into dashboards, planners, review workspaces, project boards, galleries, and lightweight tools. Today, sites can be shared with anyone in your workspace via a URL, giving teams a shared place to explore work, contribute input, track progress, and make decisions together.
Jordan Wilson [00:14:48]:
Ask codex to create a site for an upcoming customer review, and it'll generate an interactive web page with relevant product updates, open questions, usage trends, and next steps for that account. Ask it to build the scenario planner from a financial model so leaders can compare assumptions instead of reading through tabs in a doc. Ask it to turn launch materials into a living hub where teams can find the latest messaging, milestones, owners, and decisions. Oh, as it's already done, it's starting to, you know, test things in a new browser. Ask it to turn launch materials into a living hub where teams can find the latest messaging, milestones, owners, and decisions, then ask codex to keep the site up to date as details change. Instead of adapting to the limits of a single tool or file, teams can create sites that fit that fit that work. Insights aren't static. They can also help track progress for a major project, help guide customer service reps, or act as a repository for your team's creative briefs.
Jordan Wilson [00:15:44]:
Alright. And then they give, some examples. So again, this is kind of the equivalent of creating bespoke software for your team that's connected to dynamic data that you can send anywhere. But it's not just creating software. Think of the things like spreadsheets, decks, you you know, all those things that you always have piling up in your downloads, your documents, your desktop folder. Version one, version two, version three, version four, version five. Right? The downside with the old way of file sharing is exactly that. The versioning is a headache.
Jordan Wilson [00:16:18]:
Right? I remember editing video way back in the day. I would have, like, v two final, v two final, really final, v two final, really final three. Right? And on and on and on. And if you're collaborating with a lot of people with static files, I think this is now looking at something like codec sites. It is now seeming like one of the biggest no brainers out there. Right? So it's not just, you know, vibe coding, you you know, pieces of software or dashboards or creating tools, calculators, whatever it is, right, that your company can use. You know, ROI calculators calculating the likelihood of the sales, you know, your own little Kanban tracking for, you know, an event. Right? Kinda like disposable software, but you can actually use it, with your team with live data.
Jordan Wilson [00:17:03]:
It's not just those things, but even things like presentations, something that you would normally send, you know, a Word document with a couple of charts in there. Okay. You can just create a site because it's much easier, to update that site because it's always live. It's always dynamic. Right? I can't count the number of times it's in the hundreds. Right? Where I've had to collaborate with internal or external stakeholders about, like, oh, here's a new version. Here's the new version. Make sure you check v three five.
Jordan Wilson [00:17:30]:
Right? Like, whatever it is, so much time back and forth is lost in this, and I think codec sites is a great way to counteract that. Alright. So we're gonna, check back in. Alright. Hopefully, it finishes with my, team rate limits, but let's just go over some other highlights here. I've already talked about the big problem with static documents is as soon as you hit send, you don't know if someone's gonna read it within an hour, within a day, within a week. Chances are, especially with how fast AI moves, that changes the pace that your business moves as well. So with that, static files now, their life in many cases, is like nothing.
Jordan Wilson [00:18:16]:
Right? As soon as you have to send that that deck, that PDF, right, you you force everyone through one linear experience regardless of their role, and it's just a frozen snapshot in time. So codec sites can turn any prompt into a live tool that your team can actually use. So whether that's a client portal, an internal business tool, a dashboard. Right? Codecs can just build the app with the new sites tool. So here's the other thing. It can host it. It can do auth. You can upload your database.
Jordan Wilson [00:18:47]:
Right? You don't have to necessarily even though Codecs has great plugins with things like in connection to things like GitHub, Vercel, Netlify, all these other things, you don't always need those. Right? You don't need a developer to get you started. Because now operators can build and ship dashboards without waiting for an engineering queue. If if you're a nontechnical leader, you can receive live current business context, right, updated in real time. I mean, you can literally have this thing updated automatically by codecs every single hour. Think if you have a spreadsheet that's receiving a lot of, automatic data. Right? If it's syncing from all these different sources, which is, you know, what a lot of, you you know, BI teams do a lot of marketing, advertising. You know, you have things connected to Google Sheets.
Jordan Wilson [00:19:37]:
Right? Why not take those five different sheets that have all these live flowing datas, pieces through them. Right? And you're sometimes, you you know, sharing this, downloading it, and sharing it to external stakeholders that can't have access, you know, to certain things. Right? To be able to turn that into a site now is huge. You do right now have to have the same, business plan account or enterprise plan account because you check these things by, well, by being connected to your OpenAI chat GBT account. So I do hope in the future, there's not a way right now, that you can share these with, you know, external or put them on your website. But I do hope and assume that eventually, OpenAI and the codex team will move that direction eventually. So is this a Replit lovable killer? Maybe. It depends on what you're using Replit or lovable for.
Jordan Wilson [00:20:24]:
Right? So Replit, lovable, and Bolt. Right? These are just full stack app builders with editors, databases, and hosting built in. So codec sites is definitely slimmer, but it builds directly inside of the workspace in context that your organization already use. That's the big thing. People don't understand, and codecs actually made this the OpenAI team actually made this better with an update two weeks ago. So not only this is codecs, obviously, can always be connected to your dynamic data sources, but it now knows everything that happens in all of your chats, in all of your projects. Right? Codex can literally start its own, you you know, it can in one chat, it can go start its own chats. So it can control and understand everything that happens in your codex account like a human could.
Jordan Wilson [00:21:09]:
Right? So, yes, you know, you do have these more vibe coding or agentic engineering platforms like Replit, Lovable, and Bolt that are definitely more feature rich. They are much more mature and advanced. But if you're using something like Replit, Lovable, or Bolt, and you feel like you're only using, like, 10% or 20% of those services, and you're just like, I'm just using some of the basic features of this. It's more for the, hey, system. Go build something for me that I can make live. Right? And I don't necessarily understand the deployment side, the database side, all those things. Right? If you're just using surface level, I think codec sites could be better, especially and replace your usage in those instances, especially if your team is heavy on chat g p t or codecs. But the advantage is not just more features because codec sites does not have more features.
Jordan Wilson [00:22:01]:
It has fewer features now, but it's really with working with your existing code, codecs data without switching tools. That's the biggest thing. So not only that, but there are huge advantages to the actual features inside of codecs and then pairing those things with codec sites. So as an example, OpenAI just rolled out the six role based plug ins. You can obviously combine that with codec sites. It rolled out a great new annotation feature, right, where if you see something in codec sites, you can use the annotation feature, highlight something, leave a little message for codecs, and it will go fix it. Right? Whereas a lot of these other platforms don't have that. And sometimes making a couple small, whether they're structural changes, visual changes, UI, UX tweaks, A lot of times, because it's the nature of generative AI, it may change a little bit of everything.
Jordan Wilson [00:23:00]:
Right? So this is just a way to bring in, the different features of codex, in things like automations. That's the other big one. Being able to, connect your data and update something like this live. So this is early. Right? Codec sites is still very early. It's very young. But I think some of the best use cases are for marketing, agency, teams, advertising, business, you know, any BI teams, business intelligence. If you're in data analytics, if you're in any data role, and if you have a Teams chat gbt account, this is huge to be able to just build a live tracker that your team can use today.
Jordan Wilson [00:23:43]:
One other thing, a couple pieces of advice, these are live by default. So keep that in mind. You can tell, you know, there's some safeguards you can build in by, like, telling codecs, create a draft version of this before it's live. That's the other downside. Right? So if you are more of a tinkerer or super nontechnical and you're just getting started with codec sites, I wouldn't make this the most important thing for your organization and share it with everyone. And then you're going in and tinkering with this in the middle of the day, unless you set up, like, a draft version. I do assume that OpenAI is gonna come up with, you you know, a a little safeguard for this, but keep that in mind. And then like I said earlier, right now, it's only it gives you essentially an internal URL.
Jordan Wilson [00:24:24]:
And when you click on that URL, you have to log in with your ChatGPT account. So it's shared across those users within your ChatGPT organization to have that added level of security, and data governance. So demos over memos. Alright. I talked about this with, Google's chief evangelist, Richard, Richard Siroter a couple of months ago, and he said this thing that stuck with me. He said demos over memos. And I think that this is really indicative of where knowledge work is heading, especially for nontechnical people. I don't wanna read a a a a PowerPoint.
Jordan Wilson [00:25:02]:
I don't. Right? I don't wanna read a long word document with mountains of text. I want something interactive. I want something that's visual. I want something that's dynamic. I want something that's up to date. And I assume the rest of the business world will probably agree with me. You don't wanna be looking at four different versions of something that you downloaded and then having to go back and forth on Slack or Teams or email or set setting up a Zoom and be like, which version am I looking at? The pre again, I'm maybe this codec sites will set off a a whole new category of work, a a deliverable that makes sense, an artifact that is living and breathing and reflects the current state, the grounded truth of your company.
Jordan Wilson [00:25:47]:
I think, right, whether it's, something like codec sites or anything else. I I I do assume that this is the future deliverable that's gonna be picking up the most steam size, steam. I assume things like decks, spreadsheets, even old school BI dashboards are going to eventually start to lose their momentum because of the lack of flexibility, because of the static nature, or because, well, it takes too long because there's only two people in an entire 10,000, person organization who can make those things happen. So now you bring all of this feature, all of these features and functionality to nontechnical business leaders. I think the opportunity is huge. It is here to execute faster, and I think codec sites point toward fewer email attachments and fewer competing versions of the same file, which is more clarity, more speed, and more velocity for your average knowledge, worker. Alright. Let me check back in on our codex site.
Jordan Wilson [00:26:46]:
Let's see. Did did it work? Alright. It did. Fantastic. So AI moves too fast to follow, but you're expected to keep up. Otherwise, your career or company might lag behind while AI native competitors leap ahead. But you don't have ten hours a day to understand it all. That's what I do for you.
Jordan Wilson [00:27:11]:
But after 700 plus episodes of Everyday AI, the most common questions I get is, where do I start? That's why we created the Start Here series, an ongoing podcast series of more than a dozen episodes you can listen to in order. It covers the AI basics for beginners and sharpens the skills of AI champions pushing their companies forward. In the ongoing series, we explain complex trends in simple language that you can turn into action. There's three ways to jump in. Number one, go scroll back to the first one in episode six ninety one. Number two, tap the link in your show notes at any time for the start here series, or you can just go to starthereseries.com, which also gives you free access to our inner circle community where you can connect with other business leaders doing the same. The Start Here series will slow down the pace of AI so you can get ahead. Alright.
Jordan Wilson [00:28:09]:
Live stream audience. I'm bringing this back up so I can go through here. It only took eleven minutes, which actually isn't that bad. So I can go through, see exactly what it did. On the output, right, I even in my prompt, I had it spin up a sub agent, to check the work. That's probably why, it, you know, ended up eating some of my, some of my let's see. Okay. Not bad.
Jordan Wilson [00:28:36]:
This one did much better. It only took about took up about half of my usage. So I'm going through here. I'm checking. Okay. It went through, looked at my buzzsprout, looked at my circle, looked at my beehive. Perfect. Alright.
Jordan Wilson [00:28:54]:
So I'm scrolling through, and then it started to test visuals. It went through. It was testing it for mobile and desktop responsiveness, which is really cool. It spawned the sub agent. The sub agent went in there, fixed a couple of things, poked and prodded, and then it created the site. Awesome. So let's go ahead and take a look. Alright.
Jordan Wilson [00:29:17]:
So I'm gonna have to log in. Let me see if I can view the local host version of this. And I wanna have to live in log in here, live. Let's see if there's another way I can do this. All right. I'm gonna have to stop sharing my screen for just a second. As I put in my password, you love, love these demos. Love these demos.
Jordan Wilson [00:29:43]:
Alright. Let me finish logging in. Yeah. I have to, juggle couple different accounts here. Alright. Let me pull this on the screen, and there we go. So our livestream audience can see, but podcast audience, I have a dashboard here. I have a dashboard.
Jordan Wilson [00:30:01]:
Pretty nice. I have a URL that I can share within my, ChatGPT business account. Pretty great. Scrolling through, it's nice and smooth. It has this kind of parallax, feature, which is really nice. So, it pulled in. Right? So it said podcast demand is broad, beehive engagement is healthy, and circle segments point to a training in ROI need. So it has kind of like these little live indicators.
Jordan Wilson [00:30:31]:
So it has the total number of downloads over a twenty one day period, the number of our Beehive subscribers, our click through rate, open rate, the number of active community members in our circle, the number of recent posts and comments, in our community, all those things. Here we go. So now it's it's showing me some podcast trends. So it's showing me, okay. Cool. So, you you know, it's showing me, like, yesterday's, episode or no, Monday's episode, you know, showing it's above average. So it's saying that that one is off to a hot start, maybe because we covered, you know, Apple's w w w d c. So it's it's kinda showing me a couple episodes that are outpacing, where they would normally pace.
Jordan Wilson [00:31:18]:
And I know that now because it went into the last twenty one days, and it looked at daily download. So it knows, hey. Normally, a pod you know, a podcast episode that's two days old, you know, let's say it has 5,000 downloads. This one has 7,200. So it's outpacing. Right? Even though it's only a day old, it's outpacing where it would normally be. These are things that I normally don't know because I don't have time to go in there and create these little, you know, formulas or little algorithms that tell me these things. So it's going through there.
Jordan Wilson [00:31:47]:
It's finding, it's finding, you know, podcast episodes that are out to a good start. It's showing me some of the most popular episodes over the last twenty one days and some of the reasoning. Right? So, you know, Hayd said this one did well because of a strong mix of model news and practical tools. And then it's showing me also, which is pretty cool. It went through even though I didn't ask it to. It went through and found me some ever evergreen and back catalog movers. So it looks like, you know, this episode from five months ago, over a twenty one day pea over the most recent twenty one day period, still had, like, 700 downloads, and I have no clue why. Right? But, essentially, it's saying that it's part of the start here, start here back catalog is still drawing a lot of demand.
Jordan Wilson [00:32:35]:
Cool. So I know that that episode that's now really old is still getting a lot of downloads. So I might, as an example, be able to look at this and say, hey. Maybe I should create a fresher version or a new and updated version of that episode from six months ago or five months ago that's still driving a ton of demand because that means people are searching for it. Or, you know, I'd probably have, I would go in and, update that little prompt that I did and always have it investigate these things and give me a, you know, three step action plan or something like that. So I I wouldn't use this as is. Right? This is just for a demo purpose, and I was trying to get it in my usage limits to be able to show it live. But once I'm able to use my full, you know, token, allotment, I would really build this out a lot more to not just find these things, but then build out actionable next steps.
Jordan Wilson [00:33:23]:
But right now, this is something that's gonna save me hours a week. Right? Same thing here. It's looking at my newsletter focus. It's it's breaking down some of my more recent newsletters. You you know, each the number of sent open and click for all these recent ones, created some nice graphics for me, kind of our acquisition mix over the last seven days, you know, where these people where our new newsletter subscribers are coming from, the engagement trend over the last couple of days. So some of these things I know are readily available in my Beehive dashboard. Some of them I'm looking at, and I'm like, these require a couple clicks deep. Not only that, but it's bringing together.
Jordan Wilson [00:34:02]:
So it looks like I'm trying to do a rough estimate here. At least 50 pieces of data that would probably force me to go into at least 30 different screens and hundreds of clicks. Right? So that's that's the reality of something that you can start thinking of today. What is that business problem that you are either forced to do, but it takes way too much of your manual time that you would like to be able to share that with your colleagues, or something that you don't have the time, or something that you or someone on your team does not have the current technical capabilities to be able to pull all of this together in a way that just visually makes sense. Because as you can see from the livestream audience that's looking at this dashboard, it's really slick. Right? This is one shot. Very simple. Right? I didn't create some, you know, overly complex prompt.
Jordan Wilson [00:34:53]:
I did a good job at building it out and telling it exactly what I needed using a couple built in skills to make sure that this worked. But this with not a lot of effort is something that our team could use to save hours every single week, and it also gives us something, in terms of insights that we would not be able to do because no one on my team right now can go in and, you know, calculate, these these metrics or this little algorithm it used to find, you know, twenty one day pacing to say, hey. These episodes are more evergreen than others. I can't get that in my current in any of my current dashboards unless I download all of this information and put it into a large language model like Claude or Chat GPT or Gemini or whatever. Alright. So that's a wrap on today's show. So big picture, I think codec sites is really getting rid of two big problems. Number one, the file versioning and sharing these, you know, 10 different versions of document that are floating around everywhere.
Jordan Wilson [00:35:52]:
No one knows what version. The old version gets updated. Oh my gosh. We just wasted hours. Hopefully, it get we're moving to the demos of our memos, era of being an AI native workforce. And I think that's the big one is having a new type of live artifact that is actually AI native that's connected to your data. Right. So that's problem number one.
Jordan Wilson [00:36:13]:
And then problem number two, it's just kind of bridging the gap, between this this vibe coding capability. Right? Anyone can go in and create kind of helpful but disposable software by uploading, static documents. So being able to connect this to your team's dynamic data in a way that's already approved by, you know, your c suite or your IT or whoever to have that level of data security and governance and to be able to build something that's actually useful without having to wait months for someone from engineering, to put it together and to have that level of flexibility. This is huge, y'all. So as a reminder, go check out episode seven eighty one. That's the second part of a two part series of codex. So if I mentioned anything and you're like, okay. Let me take this codex thing a little bit more seriously.
Jordan Wilson [00:37:02]:
Go check that out. And remember, go find today's episode. I put together a ridiculously helpful. Right? You're not gonna learn too much from my prompt, and I didn't even read it off because it was kinda long. I gave you guys the literal stack to do all of this. Alright? 50 different real use cases taking into account all of the codec skills that are publicly available, taking into account all of the codecs apps that are currently available, looking at all of their capabilities, and how they can actually grow your business. I gave you one use case that's really helpful for me, but I have a stack of 50 of them that is ready to go for you. All you have to do, go repost this show on LinkedIn.
Jordan Wilson [00:37:44]:
So if you are watching on the LinkedIn livestream, that's very easy. If you're listening on the podcast, check out the show notes. We always leave a link to the LinkedIn post. So that's a wrap. Speaking of links, make sure to go to your everydayai.com. Hopefully, you enjoy putting AI to work on Wednesdays like we always do. Let me know if this is helpful. Thanks for tuning in.
Jordan Wilson [00:38:04]:
Hope to see you back tomorrow and every day for more everyday AI. Thanks, y'all.