Clip
Summary
The team showed an early demo of controlling Evo Workflow Automation directly from a Teamwork chat. Benny listed active workflows, started a deployment run, asked for status updates mid-execution, and even created a new workflow — all from conversational prompts. Some of the API wiring is still being built (toggle-active needed a retry), but the core capability is taking shape. The goal is a world where you manage your automation from the same chat window you use to think.
Screenshots
Transcript
Hello, I'm going to show now the connexion between Evo Teamwork to Evo Workflow Automation and how we can control the workflow automation from Teamwork. So the first prompt I already gave it, and I ask what are my workflows in workflow automation? I see that I have two workflows.
One, it's a longer one that cheques if the smoke test went well. Let's say I'm a team lead and I need to make sure that all this went well. And then also I have another one that cheques for updates from my team.
And now let's start to run the deployment and smoke test one. So it's hopefully going to... get that workflow and ask it to start the run. We're going to see this in here in the execution.
Yeah, it already started. I'm going to wait a bit. It's going to come back to this bit.
And then we see the workflows that we have currently here. I'm going to try to create from the teamwork another another. workflow here. I'm not going to wait for that because we are running out of time.
So look at my task that it's ask for status update and create one based on that and then to have some follow up. task in that. So this one has two, one node, and the next one should have another two nodes there. So let's see what it's going to do right now.
It's looking for that one, trying to understand my existing workflow and trying to create one that has another node or another step. In that. So, let's see him.
It goes a little bit back and forth. It tried one and it was not successful. This is because we don't have yet a skill that guides how the API should work.
Yes, so now we have this one here. Let's cheque if that actually was created. Yeah, I can see it.
It's here. I can see it's not active. Let me ask to activate it.
Toggle it active and let's see this should make it here that it's active if this goes through. Let's double cheque that, let's see that. Let's see if this worked.
Just a second, I need to refresh. No, this doesn't work. I'm not going to insist on it.
This is going to go probably from the first, from the second try. Or let me ask actually. Cheque it.
Probably, yeah, it's gonna see that it's not. It tried one and it came back that actually it is, or was there a, yeah, no, it's still not active. It should, yeah, it recognised it's not.
Again, this is a little bit of back and forth because the APIs, we don't have a skill for that API. Let's see if this went through. Yes, yeah, I can see now this is active.
And now we still have this long running task here. Let's cheque a status update. on this task. Let me copy it here.
Hall. How is this? So this is checking for a long running task and it should see that it's not done yet.
Probably it would tell us what steps are we in, etc. And we can double cheque that from our UI to see if that went well. It's going to find the ID.
Yeah, so it found correctly that it's running. It's currently. It even tells us what steps it runs so far.
Oh, no, these are the actually history executions that we had today. Let's ask what step are we at? And we should...
See what steps are, or how is the execution going? This is a multi-step longer task. It's going a little bit through the API.
Yes, so it got to, we know it at what step it is at step 4. This was it on my side. If you have questions, just put them on the chat.
Thank you. And over to you, Michael.