So What? What Plug-ins are Available for Claude Cowork

Well, hey everyone. Happy Thursday. Welcome to So What, the Marketing Analytics and Insights Live Show. I am Katie, joined by both Chris and John. John has internet this week.

I have internet. This is life on the outer rim. We tend to lose communication now and then. Well, we're glad to have you back, John. Um this week, so two weeks ago uh on the live stream, we covered Claude Co-Work.

You can find that episode uh if you go to trustinsights.ai slash YouTube, so what playlist, Claude Co-Work. Last week we covered Claude Code. And this week, we're coming back to Claude Co-Work because in the two weeks' time that since we've covered it, things have changed and more stuff has been added. And I think for once, I've actually used the things that Chris is going to be showing on the live stream. But we'll see.

We'll see. So I guess first question. Um, John, I know you were using some generative AI earlier to work on some responses, but have you done any digging into Claude Cowork or Claude Desktop or Claude at all? No, like the past two weeks, you know, it's been all the rage. I've we covered it in marketing of our coffee with Chris, and I know you guys have covered it.

So uh yeah, I'm interested to see all the hooks and plugins this week, but I have not uh again. I let Chris run around and show off all the cool stuff, and then I'll jump in and take the cool three things rather than spend any time digging. All right. Awesome. So Chris, I'm nervous.

Where do we start? All right, let's do some table settings because there's some new vocabulary. What has changed in the two weeks since we did a show? Is I'm nervous. Co-work has added support for plugins.

Um and so this is there's there's three four components that we need to be uh vocal piece of vocabulary we need to be aware aware of in the clawed ecosystem. There is our agents, there are skills, there are commands and hooks. Yeah, commands and skills are pretty much the same thing, they are a shortcut for how to do a specific process. Command is the older version, skill is the newer version. Okay, a hook is a trigger.

If you think about like a trigger in like your marketing automation software or your spam filter, whatever, a trigger basically says, hey, when this happens, do something. Uh, if you use tag manager, Google Tag Manager triggers, very similar idea. Agents are little workers, little self-contained things that do a specific thing, like the co-CEO uh of Katie or the voice of the customer. If you bundle together agents, skills, and hooks into a package, that package is called a plugin. And it in the claw ecosystem plugins operate in two places: Claude Cowork and Claude Code.

They are cross-compatible, which is important because if you're using certain plugins in clawed code and you really like them, you can use them in clawed co-work with no major additional changes. Now, the reason why this has caught so much attention is because this adds in all of those skills and agents and stuff that we love in Claude Code, but for people who don't want to fire up a terminal window, it's kind of inaccessible. Also, Anthropic bundled with the launch, 11 of their own Claude plugins for co-work in a variety of areas. So let's go ahead and take a look. I'm gonna shift over to our uh co-work environment.

Remember, there's three environments: there's chat, which is regular conversation, there's cowork, and then there's claud code on the web, which we specifically ignored because it's not particularly good. So in co-work, there's a new little plugins window down here, uh, option down the lower left-hand menu. If we tap on plugins, you can see here are the 11 plugins that Claude now comes with that you can install: bio research, customer support, data, enterprise search, finance, legal, marketing, product management, productivity, and sales. So these are the out-of-the-box ones that Anthropic made for users. So for fun, let's install one.

Let's install the marketing one. And so uh no, I have not used this. I was using the new new connectors this morning. I got very excited thinking I was getting ahead of the game, but the answer is no, I haven't. So this is all still new to me.

I'm trying. So now we've installed the marketing plugin, and you can see there's now six new commands: brand review, draft content, campaign plan, email sequence, competitive brief, performance report. If you want to, and you should do this, tap back on the plugins menu, and you can see what is inside. So, if what are the commands, brand review, and then it shows you the prompt that it uses. So this is this is an important thing for people to know.

If you're not, please don't use these and just say, okay, Anthropic made it, it must be right. Like actually go and look and see what's in the box. So these are the the commands, the connectors that you could install with it if you wanted to. Uh, and you could see this one has uh Slack, Canva, Figma, HubSpot, et cetera. So you could install these connectors that would then give the marketing plugin the ability to talk to systems.

Like if you have hrefs and you have HubSpot and you have Slack, you can install, you turn on those connectors, and then as you're using the marketing plugin, it can talk back and forth through the systems and say, hey, I want to let the sales the marketing team know that I put the new live stream graphic in Canva and it's ready for you, right? Because you can see can button Figma in there. And then you could see the the skills, uh, what skills it has. So commands, connectors, and skills. That's what's in the box.

And so now, if I wanted to do any of those tasks, I could say, like, here's an email, help me design an email sequence. Here's an idea I had. Let's design an email sequence for it. Walk me through it. Same rules apply.

You still want to give it a folder to work in. You still want to give it um, hey, wait a minute. That's weird. They revved Opus to 4.6. That was not there an hour ago.

Well, and so that's actually helpful because offline I was gonna let you know that I'm getting an error message uh in my cloud saying the model used in this conversation is long no longer available. Switch to continue chatting, but nothing happens when I try to switch the model. So it only lists those the two older ones. Um but anyway, so as of preparation for the live stream, things were different than they are now. Surprise.

Well, here's what we know behind the scenes. This this is just some some inside baseball. Uh Anthropic released a bunch of commercials that are going to be airing at the Super Bowl, and it is widely suspected they're going to release new models either just before or during the Super Bowl as their way of sort of sticking it to OpenAI and ChatGPT. Uh and the commercials actually very funny. They're on YouTube if you want to go watch them.

Uh, they're all making fun of chat GPT ads. So back to co-work. We now have the ability to use these plugins. And you can install multiple plugins in the same way that uh in Claude Code, you can install multiple plugins. So if I wanted to, I could say, let me let's add let's browse some more.

Uh John's here. Let's install the sales plugins. And Katie, you like productivity, so let's install the productivity plugins as well. So let's see what's in sales. Sales has call summary, how to do a sales forecast, do a pipeline review.

So you could actually connect HubSpot to this and say, let's do a pipeline review. And it there's a different ways of doing that. The connectors, it would it's going to ask for things like clay, close, hubspot, notion, zoom info, fireflies, right? So it can review if you want to do a call review, a sales call review, you could say do this, and then post the results to Slack. And then the skills account research, call prep, competitive intelligence, create an asset, daily briefing, draft outreach.

So that's what's available now in Claude Cowork if you want to do these tasks uh in sales. And then finally, if we go to productivity, we have uh just getting started and updated, and then memory management, task management, just generic stuff like, hey, I I need to remember how to do things. I want I want to uh improve my productivity through the day. So is that all out of the box stuff? That's what it comes with just out of the box.

Yeah. I have some questions. I know you're surprised by this. So I'm looking at this and I'm like, oh my gosh, that's so exciting. They are basically, you know, for businesses like ours where we really can't add headcount at the moment, like it's like adding that headcount without us having to do it.

But as you're walking through it, I'm like, oh wait, it wants to access all of my Hubspot data. It wants to access all of my Asana data. It wants to access basically my entire ecosystem of how I'm running the business in order to make me more productive. So can we just acknowledge for a second like yes that sounds great out of the box but that's also a little terrifying. Yes.

And you have to yeah you have to trust in the service level agreement and the terms of service that you signed when you when you purchased your clawed account um that that that is you know that it's it's pre that it's as ironclad as it can get with a big tech company. Hmm okay there's also a warning so like the legal one they do say with the league one this is not a substitute for an actual lawyer. So this is more things like review contracts, triage in NDA, et cetera, prep for briefing with my meetings and things. But they are they also, you know, throughout the the the thing says, do not provide legal advice because they don't want to get sued. Well, I look at that and I'm like, oh, that could be interesting.

Cause one of the things, John, I know that sort of like can bog down your time is going through NDAs and MSAs that are sent over by clients to make sure that they are, you know, all the languages correct and well represented for them to be mutually beneficial. It sounds like this is something that we could use to help sort of maybe uh make that kind of task more efficient, knowing that John, the human who has the most experience and expertise would still want to review what Claude Co-Work has said. Okay, here's where you should red line. Here's the language you should change based on your standard templates. And you'll notice the first part of this prompt is playbook review based review methodology.

If no playbook is available, ask the user to make one. Check for a compliment configured playbook that defines the organization's standard positions, acceptable ranges, and escalation triggers for each major clause type. So for example, when you're doing a sales contract, if they says, hey, we we pay net 120 and our playbook says net 30 is the max, this will flag and say, uh-uh, we you your playbook says net 30, but it's contingent upon us having the playbook, which we currently don't think we have a legal playbook. We have a sales playbook. I don't think we have a contract playbook yet.

We have our standard templates, which could be which would basically, I would say we would feed those in as the playbook. So we already know what those terms should be for trust insights. Yep. Uh let's see if there's any other pla uh built ins that are worth playing with for fun. Again, for you if you want to do data analysis, if you want to do uh customer support and customer care, finance.

Let's take a look at under the hood at data analysis. So this is things like exploring data sets, building dashboards, creating visualizations, writing uh queries to your database. Um interestingly, so this is where you're starting to get into why plugins are more than just prompts. If you look here in the data context extractor, there's references that contain things like templates, and then there's a folder, a thing called scripts that includes actual working Python code to do some of the data analysis so that Claude doesn't have to try and reinvent the wheel every single time you use the plugin, um, which is pretty cool. Oh, Andrew Bruce Smith says Opus 4.6 has just gone live.

Uh so the Anthropica is in the process of releasing things right now. I remember so go ahead. I was saying I remember we were at um Maycon two years ago, and Paul Ritzer was about to get on stage, and they literally released whole GPT 4.0 was released the moment you know, minutes before he's getting on stage, he's like, Well, there goes my talk. Yeah. Um here's a question for you know companies like ours, but also maybe like larger companies who are using a shared login.

So we're all logged into Claude Desktop through our client services account, which is the account that owns the Claude agreement. As you're installing these plugins on your desktop account, would I see those reflected on my desktop account? Or do I then also have to go in and install whatever plugins I want to use? I believe you should see them on yours. Okay.

Oh yeah. 4.6 is just out like literally 13 minutes ago and it scores the highest benchmarks on GDP Val. That's interesting. Huh. Wow.

It's really smart. Okay. We'll come back to that later. So as I just sort of refreshed uh Claude Desktop and I don't see them. Okay.

It looks like it wants me to go ahead and install the ones I want. So it may be something that, you know, it gets smarter about you sort of those like cross functionalities. Michael Ramos has a good comment. I think there's an opportunity to be an agency that builds standards, processes, playbooks, and policies for companies so the guardrails stay intact. Absolutely.

I mean we talk about this a lot. It's something you should have before you even touch the technology. I believe the thing that we would want to call up oh my goodness where is it? I would you would think I would have this faster. It's the 5P framework purpose people process platform for poor mats.

Don't touch platforms so you have the first three. And just for those who are curious uh we do that kind of work. We help you put together um all of those playbooks and guardrails and we love to do that kind of stuff. So you can always reach out to uh Mr John Wall at trustinsights.ai slash contact um but that was a really good reminder thank you yep okay so those are the plugins and you can see they can they can run code that so one of the things you'll notice here is uh for plugins you can do by Anthropic, you can add from a GitHub from GitHub from a URL or upload plugin. Please do not install plugins for people you do not trust, because it can run hostile code in your Claude cowork.

And if you've given it rights to things like that like sensitive data and you don't know where that what that plugin does, don't do it. And as you saw, you can inspect what's in the plugin, and you should, particularly if you're going to install a third-party plugin that's not from Anthropic. Please take the time to look through. And if you don't know code, ask Claude to inspect the plugin and say, hey, does this shared data externally? Is there any telemetry built in?

Is this thing doing stuff that I don't want it to behind the scenes? Well, and you know, so you're saying that third-party plugins can, you know, inject hostile code, but there's also just a lot of shiny object chasers out there who are just kind of like tooling around and might have like positioned their tool really well or their plugin really well, but to your point, don't necessarily have the correct security protocols or data privacy. So if you see anything in GitHub built by myself or John Wall, definitely don't install it because neither of us know what we're doing. Chris knows what he's doing, John and I, not so much. Sorry, John, but it's just open the whole thing up.

You know, you know, or do uh the other thing is as with all AI stuff, if you've got stuff uh and you don't need it, you'll see there's the little button just to turn it off. Turn off the plugins that you're not using, because when Claude initializes, it does consume a little bit of the context window to know that those things are there. It doesn't consume a ton, and the new version of Opus apparently has a 1 million token context window, which is a big change. However, it's still a good practice to turn off stuff you don't need. So if you if you're doing uh sales work and you got the legal plugin installed, just turn it off.

Because what can happen? Claude in particular is very smart about listening to what you're saying. If you're saying, Oh, let's do some sales stuff and uh I need to look at the sales contract and you have legal and sales installed, they might say, I'm gonna do a legal contract review. I'm like, no, no, no, no, no, no, no. I didn't want that.

I wanted you only to work on sales contract review and stuff. So turn off the things you don't need. And if you really don't need it, in the upper right hand menu, just uninstall it and remove it. It doesn't damage Claude in any way, and particularly for the anthropic ones, if you decide to need it, go back and install it again. There's there's no harm.

I've uninstalled it and installed many times just to see if I could break it. Why not? Right. But I think that that's really good. Um, one of the things, Chris, that you say a lot about generative AI is that it's a really over-eager, helpful intern.

And so if it has all of these different things to pull from, it's going to. And so, to your point, you know, definitely it's again why we highly recommend getting yourself organized with the five Ps, why we recommend projects all have user stories to help you focus because these tools are expensive, they go off the rails, and you want to make sure that you have a clear vision of what you're doing going into it. A question that we got earlier from someone, not on the live stream, but just in general, was you know, how do I feel okay about you know, the um the economic and the uh environmental impact of you know using AI and you know, Chris, you had given a more technical answer, but my advice to this person, knowing that she's not a technical user, is you need to be really thoughtful about how you're using it. You know, don't out you know, don't outsource your thinking to it, like really think about why it is that you're using it and what you want to get out of it. Because if you're just asking it random questions that you know, you're just like, oh, I'm curious about this, curious about this.

That's when that environmental impact starts to build up. But that's probably a topic for a different show. No, definitely, definitely a topic, but an important one and worth doing. All right, so let's turn the marketing one back on. Let's just see what this looks like.

Um, just so we can get a sense of it. So I want I'm gonna say, let's do a um a campaign plan review. Let's do a campaign plan review. And you start typing in uh the slash command, you'll notice that it you have both the plug-in name and this the skill name. So you want the shorter one, campaign plan in this case, because that's we're gonna do.

I'm gonna drag and drop in the inbox insights growth strategy, which we did a while ago on how do we get our newsletter to grow. And we're just gonna say, just review this. Let's uh just tell me what's it what do you think about this? And let's let's hope that the Anthropix APIs are working at all, having just launched a new model. So, as with everything, you can see it's gonna in co-work, it's gonna start by loading up the skill to understand it and then uh taking the pieces apart and uh taking the document out, doing its data extraction, running its commands, and it'll provide in theory.

If it was able to successfully get the document, it'll provide. Oh, it's gonna have questions for us, Katie. Okay. Uh well, let's see. Uh, what would you like the focus of this campaign review to be?

Full strategic audit, feasibility check gap analysis, or actionability review. Hmm. Let's do a full strategic audit. Full strategic audit. What format do you want it in?

A word doc, a chat summary, a spreadsheet, or something else? Uh, Word doc is fine. Word doc. All right. A full strategic audit delivered as a word document.

So it now has to load its docx skill uh as well uh to to get that out. Sure. Well, you know, it's funny because I've been using co-work for quite a few things, and you know, Chris, you've seen sort of what my output is. And it is always interesting when it goes through, and if you don't give it that direction, it's like, well, wait, what is it you want me to do with this? Here's some options.

And I really actually appreciate that it goes through and it's asking those questions because it makes me the end user go, what am I doing with this? And so I think that having those questions up front of like, is this a full strategic audit? Is just a feasibility? Like that's actually really helpful because those are two very different asks. Exactly.

So it's interesting, you can see here, as part of its thinking, it's doing a bunch of research on our strategy plan to validate are we even are we even uh strategizing about the right things? Um apparently one of the new things in this model, and this is true of the previous model as well, is it's really good at knowing what um what sub-agents and tasks to autonomously come up with, say, like you know, this is this would be a good thing to do here. Whereas in in other models and in previous versions of Claude, you have to say, and use your web search tools to do this thing. Hmm. Very interesting.

Okay, so now for some reason it's building its own JavaScript. I'm not sure why it's doing that. Um that might take a while. Okay. Uh so in the meantime, Chris, we got a question.

Um, and you know, we won't do the whole example in the live stream, but I think you can speak to it, Chris. Is so let's assume marketing needs to be in legal compliance. We're turning off a legal plug and remove the safe harbor guardrail, or are we still looking or are we still working from the rag to stand on task and in compliance? So it's really interesting. Um Anthropic stuff, because they're Boris Cherney on the uh Claude Code team talked about this.

Anthropic does not use a rag. Um, they found that reg, which stands for retrieval augmented generation, delivers substandard performance um compared to just good old-fashioned search. So you'll see in all instances of Claude and Claude Code and and its various incarnations, uh, it just does straight up search. Now, in terms of safe harbor guardrails, it depends on your Claude subscription and what you are paying for and where your data is being processed. Um there you can, for additional fees, of course, pay for uh data processing in a certain section.

Of course, I have to put up the uh the required banner here. Uh, we are not lawyers, so please do consult with your legal team about how safe harbor applies to you and your company. But this is this is being processed by Anthropic in presumably the nearest data center you just for cost uh reasons. So um John, we were talking about um, so speaking on legal, but you know, we would never use this for legal advice. Would you consider using a system like this to do the first pass of an NDA or an MSA that's come across from a potential client to see where you need to red line?

Or is that something that you would still want to do yourself as a human? No, you know, my my take on that has changed. At first, I was just like, no, it's not worth doing because I still have to do the full comb through. But I've realized what it needs to do is I just it's what we were talking about with like with our playbook. I need to put together three or four pages of here's the things that I look for, you know, because and I realize because I just have that process baked in as I go through all the sections.

When I come to a section, I'm like, okay, here's the four gotchas that we I have to make sure aren't in this section. Well, if I just made a document of all those things, then I could say, hey, here's the uh trust insights patented NDA examiner skill, and then I can just call that. And then that, you know, the great thing with that is I don't know if it would get it right if it needs to be corrected, but it could easily go through and be like, okay, none of these problems are in these five sections, so you only have to go read these two sections, and that could save a ton of time there. So yeah, I did um I am confident in doing it. The the big thing is to make sure that you're you're easily using the process of it being able to categorize to get all the stuff that's okay out so that that then there's less work for me to do.

Because yeah, the idea of just saying, hey, check this out based on how our other agreements stuff can get missed in either direction, and it's not really a help because I still have to read the whole thing. But if I can have it say, hey, you don't need to read these three of the in fact it's not three of five, it's usually like three of 64. Um that can save a ton of time. I feel like I need an animation for these moments because folks, we just live on air had a breakthrough. We had our laggard on the innovation curve.

John's on board. John's on board with using AI. I just I need us to just acknowledge this moment. It's not that you don't believe in it. You obviously work with us, but for your own personal use, this is such a big deal.

And I love the way you're thinking about it. And I think it's just a big moment because John, you're a good representation of a lot of industry professionals of it's just easier to do it myself. But seeing that switch flip that you're like, oh, if I just do this, it's gonna save me so much time. And that's like that's such a huge moment. I just want to acknowledge it without sort of like, you know, putting you two on the spot.

But that's a big deal. Yeah, well, but it still comes down to the truth of I have to do it the first time and confirm that it actually works, you know, if it still comes back and it's like, yeah, you still need to read 27 of 27. It's you know, we're going nowhere. But yeah, it is it and the other thing that can't be downplayed with this stuff that I've been doing recently is that you know, I'm building these prompts based on the 65 skills that you and Chris have already put together in there, you know. So it's not like I'm just showing up doing it.

It's I just say, hey, obey the rules that these guys have spent two years putting together, which you know is now I look really smart. But the fact is you guys have done a ton of work to get me there. We we all rise together, John. That's true. Raises all it's for your org.

This is why you need to call Trust Insights because you don't want to be doing this yourself. Have us show up with our wagon full of 2,000 prompts that we've been beating on for three years, and save yourself literally hundreds of thousands of dollars. I think that's gonna be our new tagline. Hire our wagon of prompts. Wagon of prompts.

All right, so it finished its analysis, and I we won't go through the whole thing, but you can see this is a very long multi-page audit of our newsletter strategy. So we put together the newsletter strategy a little a while back. Um goal is to get to 40,000 subscribers, and it went through and section by section read through it and said, Hey, your campaign overview, strong, your target audience needs work, key messages needs work, channel strategy adequate. That's that's damning with faint praise. Content calendars adequate, content pieces needed strong, success metrics adequate, budget allocation needs work, risks and medications adequate, next steps strong.

So this is the skill and the package that it put together. And then you can see in each of the sections, it tells you its reasoning, what the strategy does well, what needs improvement, and so on and so forth. And so this is built right in to just the the basic document itself. And you can then take this. Now the logical next step would be okay, Claude, you have my original, you have your audit, write me a better one.

Right, and uh, I'll come back in 15 minutes when you're done. So those are the plugins that are built in. Now, here's where things can get fun. If you were paying attention to last week's live stream, you know that in Claude Code, there's a plugin development plugin, which is redundant. And in there, you can start to build your own plugins.

So last night for fun, because this is what I do for fun in the evenings. Kind of lame, but I wanted to write a press release for AI, not for humans. We know that AI sees things differently. And we know if you've taken our um generative AI use cases course, that if you're building press releases for AI, you're gonna do it in a different way than you would do it for humans. And to start, you need to have some idea of what it is that AI is looking for.

So we have our so from our we started with our AI readiness course, right? Which is uh the brand new AI readiness assessment, brand new thing that Katie came up with. And so I said, I want a plugin to help me write this press release, calibrate it against what generative AI sees. So my first step was to go into our uh private generative AI GEO, whatever tool, um, and I asked it um, what are some good AI readiness assessments? Right?

Very straightforward. It goes through, it it exports this massive amount of data that I then uh store on just as a file on my disk. And then I, of course, I uh have I built this plugin, and what we wanted, what I wanted to do is I want to try to semantically match press releases to what Google is doing in AI overviews. Google said these are the words and phrases and concepts that I am expecting, right? If we if I just scroll up here to read one of the traces, um, what are the specific questions, technical evaluation criteria, and scoring rubrics of the Microsoft Cisco and AI AWS AI readiness assessment tools to benchmark organizations?

So it's basically what is what is in the other ones out there. If I want my press release to resonate semantically, topically, with what Gemini is asking itself, I need to make sure that I'm using those concepts. And what better place to do that than in Claude Cowork? So in the plugin, there is uh the skill, and then there's a piece of Python code that I wrote that does the text metrics. And I said, here is, and I'll I can show you the contents of that folder, which is extremely boring.

Let's go into so I gave it just me foaming the mouth, two paragraphs. Hey, I'm promoting the new AI readiness assessment. Here's the URL, here's the benefit to people, some background boilerplate, and I copied and pasted just the assessment landing page, which contains a lot of the you know, what's in the assessment basically. And then I took the Gemini trace of what Gemini was asking itself over and over and over again as part of AI overviews, and said, Claude, use the plugin. And the plugin goes, Great, I'm gonna do two drafts of a press release.

It reads through it and it says, Hey, these are the terms based on what Gemini says. These are the terms that need to be in this press release, right? AI readiness assessment, AI project initiative, governance, ROI, framework, maturity. These are the nice to have, right? Strategy, leadership, executive.

It it reads that and then it says, okay, I'm gonna spit out two uh press releases, and then it scores them. It reads the uh it runs the analysis. Now let's see if I can show the scoring here. So it does the math on it and says uh candidate B, the second process scored highest. Now I said, Great now, make a third one and try and get your score even higher in terms of similarity with what Gemini says.

It did it came up with one and it it kept doing this. It went it went to 48. It was like, okay, 48 out of 100, that's pretty good. And I looked at it, I said, that's actually not bad at all. Um let's kill this and go back to my browser.

And so what went out on the wire last night was this gigantic press release that unlike previous versions of our press releases, which were just horrendous mangled piles of words, uh, because that's how we were doing GEO press releases a year ago. Now it actually reads coherently. So a human reads it, it doesn't think, wow, these guys are brain damaged. Well, and so this was the question I was gonna ask you, Chris is who is the audience of this press release that you generated? Because to your point, previously we had been writing, writing a bunch of organic search terms shoved together for the purposes of GEO.

But now it sounds like, you know, and so the audience was the machine, not the human. Now it sounds like the audience is both the machine and the human. Now it's still the machine. Still the machine, which I think is an important distinction. So when Chris says he's building a plug-in to write press releases, we are not in the business of public public relations, but we are in the business of helping people understand how to show up in generative AI engine optimization search, whatever.

When someone puts a you know a chat into Gemini, they want to know how often they're going to show up. This is what this does. Exactly. And this semantically, the way that AI sees in Google in particular, the way Google's Gemini sees language and understands what it means conceptually, checks the boxes on it without making it just a big pile, a big bag of words, right? Even the quote in here actually does sound like Katie, right?

Every week we talk to organizations that have invested heavily in AI tools but haven't addressed the fundamentals. They have data quality support, the data hygiene and data availability practices are inconsistent, their processes aren't documented, and their team lacks the AI literacy skills to manage and audit AI outputs. This AI readiness assessment helps surface those organizational readiness issues before they become expensive project failures. For example, we referenced the 5P framework. You can now in Claude Code make a plugin that is just a zip file of the different pieces, the prompts, the scripts, etc.

And now you hand that to your team members and say, install this in your Claude Cowork, and you're up and running. And is the fastest, easiest way to add new skills and new capabilities to your Claude Cowork if the 11 pack of 11 did not meet your needs. So, for example, we saw in these 11 sets of plugins here. Uh, let's go to the browse screen. We have research, support, data, search, finance, legal, marketing, product management, productivity, and sales.

Um, I don't see much in operations management there. I don't see anything on strategy in there. I don't see anything on public relations in there. So you and your organization could make versions of this for yourselves, and you should. You might even if it if you're really good, you might even make them and sell them to other people because you could say, Look, we are the I could see our friend Ginny Dietrich, for example, saying, who has who is the owner of the peso model?

Um, and if you use that without a permission, she will send a lawyer to punch you in the face. Uh-huh. She'll probably do it herself. She would prefer to. Yeah, no, she will do the punching herself.

Exactly. She could create a peso plugin for Claude Cowork that could audit your public relations strategy and say, here is how it does or does not align. And then any PR manager who is pressed for time can say, Here's our 2026 strategy, peso plugin, go. And it has the five or six skills, it has the frameworks that she does that she's known for, and it will say, Here's your audit, here's how you're doing. And because we know that these skills can connect to the various systems, it could say, Do you want me to come up with a project plan and put it in your Asana to get you aligned to the Peso methodology?

As long as the organization used Tasana, that that would be the way to do it. I feel like there's a lot of potential with these plugins, especially for companies that are looking to you know create their own. But again, with the caveat, it's software development. You need to follow the SDLC, which is the software development lifecycle. That AI has not made that go away.

If anything, AI has highlighted the need for structure and guardrails. And you know, we were talking with the uh person, uh, Michael commenting earlier about the legal guardrails. Like that doesn't go away. Yes, because AI can do it doesn't mean that you should just like, you know, okay, go ahead, do it. You still have to have that plan that you know, the requirements, you know, here's everything.

So there's a lot of potential to develop your own plugin, to develop your own, you know, custom GPTs and gems and everything with the caveat, it is still software development. Yep. And if you're using cowork, co-work is an agentic environment, which means that it can take your mistakes and make them faster. It will amplify them. Yes, it will make more mistakes faster.

Um, which is different than a GPT. That was actually a question we got uh in one of the Slack communities. Let someone's saying, What's the difference between this and GPTs? Yeah, GPTs are not autonomous. You are still doing all the work, you're copy pasting and moving stuff into them and stuff like that.

You have to go to Chat GPT and and and do stuff with systems like this. Uh the new OpenAI Codex, which dropped was that that was this week. Um, the new version of Claude Cowork, the new versions of Google Anti-Gravity. These are agent systems, which means that instead of you having to do all the typing and stuff, if you provide the materials up front, you set them off and you say, Go, and I'll come back when you're done, and and my work product will be done. So that's the big difference.

GPTs are still a lot of very manual work on your part, whereas agents now go and and do it for you, assuming that the underlying technology is good. Yeah. Um thoughts? Go ahead, Chris. I say Michael pointed out, don't you just say that doesn't do work best word first time.

The reason for that is because it's probable, it's a probabilistic tool. So you have to you give it a starting point, and then you can, if you're specific or your plugin is specific, but here's what better means, then it knows how to make it better. If you didn't so in our public relations plugin, there's a Python script that does semantic indexing, it's a four different mathematical measures. That gives Claude a goalpost. So I can say generate two versions, score them against this.

And now it says, Oh, now you've told me what better is, I can keep trying and I can know how to make it better. But you have to be clear, or you have to have a mechanism in your plugin to say this is what better means. John, thoughts. Well, I had to laugh when you said when you ended with, you know, it's about the power of the platform and stability of the platform. And we opened the show with getting the engine ripped out of the plane and replaced as we were taking off.

So yeah, I think process is very important here to avoid uh wasn't what would they call it? A uh dramatic malfunction. So, Katie, what plugins are you going to have me build? That I'm gonna have you build. Yes.

Um I can very clearly see a co-CEO plugin getting installed in Clog Cork. I think we're definitely gonna do that. I don't know that that one would be publicly available unless we made it a little more, I don't know. Um, the thing that we've been getting a lot of asks for that would still have to be our own proprietary uh frameworks and sort of methodologies is a lot of AI alignment. And so that would be, you know, the foundation would be the five Ps, and there would be some other pieces that would go into that, but I could definitely see, you know, building something like that for you know, someone to interact with be like, well, I'm struggling.

I took, you know, the trust insights AI readiness assessment, and I don't know what to do next. Okay, uh, you know, trust insights AI alignment plugin. Help me put a plan together for what to do. And, you know, full disclaimer, it's gonna be a lot. These things no, and I say it because the majority of the issues are not around, you know, well, you don't have the right AI tool.

It's around people in process, and those are big buckets, those are big things. Um, we were talking with a client earlier, and you asked a very simple question of, you know, well, when someone makes a request, is there a ticket for that? And they said, No, sometimes they come by my desk and sometimes they come, you know, they send an email. And it's like that's the problem. It's not the platform where the tickets are being created or not created, it's people in process.

And so just as a I don't know, I like to go on these tangents. Someone needs to reel me in. I'm gonna stop. I was saying, I was asking what you want to build just for our internal use. I know.

And here we go. Uh for our internal use, I definitely think, you know, I have I had definitely have some ideas. Um, you know, a lot of them are going to be around productivity. Yep. I definitely see co-CEO coming.

I definitely see uh ask our ICP coming to blog co-work because those are two things where you could say, here's a campaign plan. Hey, co-CEO, audit it. Hey, ICP, tell us what you do and don't like about it, and then co-CEO, you know, figure it out. And I think to John's point, um, because we are going to put this on his to-do list, is once we have our legal playbook of, you know, what we expect to look for, I think that plug-in would be helpful. Um, you know, to take some of that burden off of John to manually read everything so he can be that last set of eyes to be like, what didn't, you know, the machines catch?

Because it still needs to go through human intervention. Yeah. Oh, absolutely. Finance, law, and health are the magic three. Never ever ever use AI outputs without a human being, a human expert reviewing them.

I mean, I would argue that's true for all AI outputs, but that's me. I am more uh risk averse. Uh yes. Well, let's put it this way: the trashy romance novel doesn't have the same legal risk as the the contract you're about to sign for $100,000. That's the uh the your money or your life data.

Exactly right, exactly right. Um, all right. So that's gonna do it for this week's episode on Claude Cowork plugins. Uh it is gonna be a very busy the next two weeks. Uh and I say that because we're coming up on Lunar New Year.

Lunar New Year, uh, for those of you are not aware, in the AI world means that a lot of model makers, particularly in China, Japan, and Korea, are racing to get their new technologies in markets so that they're positioned for the new year for prosperity, which means there's pressure on the Western companies like OpenAI, Anthropic, and Google to also come out with announcements so that they don't seem like they're getting left behind. So the next two weeks, there's gonna be a lot of noise. Be aware of that. There's gonna be a lot of noise in the AI space. Drop by our analytics for marketers slack group if you have questions like, hey, what does this announcement mean?

And stuff like that. We're gonna be talking about things uh there. And always, always make sure you're subscribed to the Trust Insights newsletter, and we will talk to you all next time. Thanks for watching today. Be sure to subscribe to our show wherever you're watching it.

For more resources and to learn more, check out the Trust Insights Podcast at TrustInsights.ai slash TI podcast. At our weekly email newsletter at trustinsights.ai slash newsletter. Got questions about what you saw in today's episode? Join our free analytics for marketers Slack group at TrustInsights.ai slash analytics for marketers. See you next time.