Micah Johnson (00:02)
Everyone keeps asking us, should I use Cowork or should I use n8n And the answer is yes. And in this episode, we're going to talk about what should be built in Cowork and what should be built in n8n
Alane Boyd (00:20)
Yeah, Micah, we've been getting this question individually a lot, either from clients or even just casual conversations that we're having with other business owners. It's happening more often and then things are moving so quickly that people are starting to even wonder like, well, does this replace things? Do I need to just start shifting everything to this? Where do I put what? What belongs in Cowork? What belongs in n8n or similar platform?
Micah Johnson (00:46)
Yeah. And I think this is a really important episode to be recording today right now, Alane, because this is one of the gaps that we see widening every day. Cowork has so much value. We've been talking about it. We've been using about it. We've been doing webinars about it. We've got workshops coming up about it. You need to be using tools like Cowork, but that doesn't negate the other tools yet. And I think there's just a lot of
confusion that I hope this episode can help straighten out. And maybe just off the top of this episode, we start with what does each tool actually do?
Alane Boyd (01:26)
Yeah, and I think about Cowork as like when AI and chatting with AI came out, like everybody really said, like, this is my assistant. And I see Cowork as the most elevated version of that with an AI chat kind of option, because then you can automate some things. It can pull from your from your user and log in. So it really is like your assistant just right there at your desk with you.
Micah Johnson (01:53)
Yeah, I'm going to try to come up with a terrible analogy, so forgive me everybody, but if chat is like, you know, your brother's son's best friend who came to work with you for a summer to knock some stuff out and help you like that, then Cowork is like your personal executive assistant with 10 years of experience.
Alane Boyd (02:16)
And then n8n is where I see it as we've got a process developed, multiple platforms being involved, and we've got automation and AI agents as a part of that process. So, Cowork for more just my support team, and then n8n is my company's support team.
Micah Johnson (02:37)
Yeah, that's a great way to put it is if you're thinking about Cowork is helping you get your work done, your responsibilities done. That's a perfect bucket to put Cowork in or envelope or whatever container you wanna put it in. And then n8n is your automation, your let's integrate between APIs and leverage web hooks and have deterministic automation and AI and...
Alane Boyd (02:51)
Mm-hmm.
Micah Johnson (03:04)
all the cool stuff that we've been able to do for years with that, that Cowork just can't do.
Alane Boyd (03:12)
And I don't want to misrepresent this, Micah when we were talking about this episode and the difference between where we see Cowork things happen and where we see n8n's place is I can see a lot of individuals going, well, what I do is everything that I need to do. And it's a misrepresentation of really the process for the business. So if you're a task hoarder or you don't want to delegate and you're thinking, all of these things do involve me. So I'm just going to use Cowork.
That's not what I'm really talking about. It's the difference between what truly needs you involved and your judgment because you are leading this and the agent is where you've already figured that out and you just need execution between software and tasks.
Micah Johnson (03:57)
Absolutely. And so I think, you know, the kind of secondary question that we get a lot on this thread is why can't I just use one for everything? And we've touched a little bit on this already, just in the description of each platform. But the short answer is that neither platform has enough capability, as strong as Cowork is, and it is amazing as n8n is, they both have their strong suits right now.
One cannot replace the other without doing yourself and your company a disservice for what you're trying to do through automation and creating systems.
So, Cowork is really strong at reasoning, but it's pure AI, which means you can't refine data before it makes it to that. If Cowork is connecting up to a bunch of platforms and it's making a bunch of MCP calls and getting a whole bunch of data, it's got to reason it out itself. And the output from that is only as good as Cowork can handle. And the models are getting stronger and it is good.
but your usage may balloon. And right now that matters. Next year, maybe we'll do an episode on why context windows no longer matter. But for now, it's an issue and you can overwhelm the AI and you can get frustrated with it. And it's just, it can't connect to everything you need it to connect to. It can't do all the things. So you need it to be.
your reasoning partner, your execution partner, it can run on your machine, it can create slide decks, it can write documents, it can do all that time consuming busy work, but it's busy work. It's not making judgment calls. You still need to make judgment calls, just like you wouldn't leave n8n up to making judgment calls for your business. The same applies to AI and Cowork right now.
Alane Boyd (05:54)
And we've really been seeing that they're so complimentary of each other too. We'll use Cowork to identify what the workflow could look like, what the build out in n8n could look like, and then build it out in n8n. So the two are really friends hanging out together.
Micah Johnson (06:14)
Yeah, absolutely. And I think that is the best way to think about this and why we wanted to make this episode right now, Alane, because you look at Cowork and it does feel like magic. You're like, oh man, I can set up skills and I can have scheduled tasks and I can do all of this stuff. But if you're thinking about it in a broader sense of how do I still automate my agency? How do I automate my business? How do I get this past just me? That's where Cowork currently
falls apart. But that's also where Cowork is strong enough to go, hmm, I could actually call an n8n workflow and trigger that. And we've got a couple examples coming up here. But just to kind of place that on the forefront is that the answer is not one or the other, because it can't cover everything. The answer is, should I use both? Yeah, absolutely. And I think the challenge
for a lot of this. And this is where we come in with a lot of clients too, on the strategic side is, if you have an objective and you're trying to build systems, what is the design of these systems? The fundamentals aren't going away, right? We see the fundamentals being necessary even more so now to say, all right, Cowork, you're doing this kind of work. n8n, you're doing this kind of work. Cowork, you need to call n8n.
when this happens or when you're done with this Skill, you're going to call n8n And n8n, you're going to deposit this document for Cowork to review on its next scheduled basis. There's a whole set of design patterns that we can talk about here, but maybe we should get onto some examples to kind paint the picture for everybody, Alane.
Alane Boyd (08:00)
Yeah, I mean, one of our favorite ways, and we've talked about this, is using it with Pipedrive and our email to help manage our sales follow-ups that we need to do, getting notes into Pipedrive, and where our Cowork is really great at going and finding that, updating the CRM. But when it comes to what happens after the sale, that's where we've seen, well, that's where an agent can come in.
when it moves to a certain status in the pipeline. That's where we've already identified the rest of the process. And that can be where, okay, Cowork got us to this point, and then n8n can create a Slack notification for the team that the contract was signed, or creating the new onboarding workflow, different things where n8n comes in, and we've already identified the process from that step.
Micah Johnson (08:28)
Yes.
Yeah. And I really, I want to, think this is just an awesome example, Alane, but I really want to articulate where Cowork falls off. So if you have an email thread happening, you can build a Skill for Cowork and say, "hey, look at this email thread, determine potential deal opportunities. Look at our products, look at our offerings, build this all out in Pipedrive", you know, interpret the opportunity essentially. Fantastic use for Cowork.
Alane Boyd (09:00)
⁓ Mmm.
you
Micah Johnson (09:20)
just like that would be a salesperson or a sales manager's responsibility. But then when the deal gets signed or the deal gets marked as one inside of the CRM, Cowork has no idea of knowing that. I would have to go in back to Cowork and say, hey, the deal is marked as won now process it. That right there is a huge issue because that...
Alane Boyd (09:44)
Mm-hmm.
Micah Johnson (09:45)
isn't scalable. We need the opposite to happen. We need when the deal gets marked as won something happens. Enter n8n. That's when the deal gets marked as won, that sends a webhook to n8n. n8n, like you're saying, Alane, processes the rest. So on one side it is, okay, cool, we've got this rolling, we're going to leverage Cowork. On the other side, it's, I just want stuff to happen for me when events are happening elsewhere in my platforms. And that's where n8n comes in.
Alane Boyd (10:16)
Yeah, it's like where do you really need the human and where don't you?
Micah Johnson (10:21)
Yeah. Yeah. And that's, that's an interesting way to think about it too. When you need a judgment call to be made, that is let's start in Cowork and then let's leverage that to be able to go in and say, all right, I'm going to get some feedback. I'm going to work with it. And then Cowork can take it the rest of the way I'm going to review. It's a lot of back and forth that that's hands on. It's not.
I built it and now it's running in the background. n8n is the running in the background.
Alane Boyd (10:54)
one of the...
Micah Johnson (10:54)
Okay, so the next
example is a weekly retrospective. So we've started using this ourselves, Alane, where we're gonna have Cowork do a scheduled task that's going to actually review all of our session logs that we have Cowork save as we're working with it. It's going to analyze it, it's gonna pick out trends, it's essentially what's working, what's not, what do we need to improve, what are ideas for improvement?
Alane Boyd (11:00)
Mm-hmm.
Micah Johnson (11:18)
It's going to analyze all of that and write the email. But here's where gets stuck. It can't send the email. It doesn't have the MCP connection or the MCP server doesn't have the capability to send only drafts. So that means I've got to manually go in and do that. I don't want to do that. Do you? That sounds awful.
Alane Boyd (11:36)
I mean, that's a tedious little thing that's like, sure, it's not a lot of time, but that's something that we don't feel, you know, compelled to have to be that person. It is.
Micah Johnson (11:47)
but it's an interruption, right? Like
if Cowork can run this on a scheduled basis, I don't wanna be alert for Cowork to be done. So the solution to that is we have Cowork send n8n the information that it processed and then n8n uses the API for our email to send out the email. It's a simple workflow, but it bridges that gap and combines those two so that
Alane Boyd (11:55)
Mm-hmm.
Micah Johnson (12:14)
We don't have to sit there. We don't have to pay somebody to sit there. We don't accidentally miss it. It gets sent out at the same time every week.
Alane Boyd (12:22)
Mm-hmm. I mean really going back to like these two platforms are really good friends. They can work really well together.
Micah Johnson (12:30)
Yes, I do love that. It's like the PC and Mac guys, only they're like best friends. Yep, yep. Okay, so where do you start with all of this?
Alane Boyd (12:36)
Yeah, it's the opposite of that.
Micah Johnson (12:42)
I think the easiest place to start is that you don't have to worry about getting both systems going at once. I think in today's world, you could start with Cowork, get those Skills built. And Alane, you and I were chatting about this, not just earlier today, but earlier this week as well, where it's like, Cowork on an individual basis helps you start thinking in systems and processes for yourself. And if you get that worked out, then that's your
Alane Boyd (13:06)
Mm-hmm. Mm-hmm.
Micah Johnson (13:11)
gateway drug to starting to create n8n workflows that can work across multiple roles and multiple systems and multiple platforms.
Alane Boyd (13:20)
Especially for those of our listeners that are not super technical to go and build things out in n8n, it is a great place to start. And even building those Skills is the instructions that can be used for a future agent. You're building the instructions around a recurring thing in your business. And so this is the perfect starting place, I think, for somebody that wants to get started and then can build on this.
idea as they want to automate more processes in their business and with n8n.
Micah Johnson (13:51)
Yeah, I'm really glad you brought that up, Alane, because we have done that so many times now where we'll take a, when we say Skill, we're talking about like a Cowork Skill or a Claude Skill. So we take the details of that Skill and actually build that out as the system prompt for an AI agent in n8n because we've already tested it 50 times in different scenarios and know that it works.
Alane Boyd (14:16)
And the perfect thing to know too is what is your process? What do you want it to do? And get that ironed out. Because when you go to build it an agent, or in n8n as a workflow, it is a workflow you need to know your workflow.
Micah Johnson (14:30)
Yeah, I love this. think this is the perfect end. Hopefully you listen to this whole episode because this is the gem right here. Maybe we should do a quick clip edit at the beginning to secretly tease this so that people listen to the whole episode. But essentially, this concept, and maybe we need to do a whole episode on this by itself, Alane, but this concept of how do you work out all the nuance in a tool like Cowork?
so that you can build more efficient workflows across your teams? Wow. Mic drop, Alane.
Alane Boyd (15:03)
And
thanks, Micah. And we have a couple of ⁓ workshops coming up. We've got a Claude Cowork workshop coming up on April 16th, and we have our next AI agent training cohort coming up on April 20th. So if you're wanting to start diving into one of these or both of these, we'd love for you to join us. We're gonna have the links in the show notes.
