How to grow to be more than just the sum of your employees?

2.3.2026

For quite some time, I’ve been thinking: what if an AI agent could be more than a personal assistant – what if it could be a real member of an organization, working alongside teams toward shared goals? A coworker.

Then came OpenClaw – an autonomous AI agent platform that enables AI to control a computer environment to execute complex tasks on a user’s behalf.

That was the spark for us to start building our own “OpenClaw,” but designed for an enterprise context.

Meet Loco!

Loco is a work-in-progress agent living on its own Windows 11 virtual machine, with an Entra ID identity and the necessary Microsoft 365 licensing. At its core is OpenAI’s Codex CLI, providing generative intelligence and decision-making – wrapped inside a deterministic control layer responsible for orchestrating Codex sessions, providing tools, listening to inbound channels, maintaining memory, and managing scheduled tasks, and more.

This is the step organizations need to take to become a frontier firm.

Yes – empowering employees with personal assistants (like the Copilots in the Microsoft ecosystem) is important, and most organizations are already doing it. But that alone won’t be enough.

To truly transform how work gets done, you need agentic coworkers that can take responsibility for execution – operating independently, collaborating in teams, and communicating through familiar enterprise channels like Teams and Outlook, alongside both humans and other agents.

And of course, different agents would specialize in different capabilities.

One could connect to your CRM and external B2B sales intelligence systems to generate market insights and keep your organization continuously informed about customer developments and competitor movements. It knows your customers. It tracks your competitors. And it understands your business.

Another could act as the guardian of your company’s strategy and long-term goals – monitoring progress, ensuring alignment, proactively following up on key initiatives, and intelligently distributing tasks across both human and digital teammates.

And one could specialize in capturing and documenting the organization’s tacit knowledge in a wiki. This agent would not be the main actor in any team – instead, it would serve as a silent and patient “clerk”, continuously taking notes, structuring information, and making it accessible to both humans and other agents.

Agent specialization is accomplished by installing the necessary MCP servers and CLI’s on the agent’s virtual machine and by giving agent specific instructions.


Take a moment to reflect on that – and then review the screenshots below for inspiration.


# Use Case: Sales Agent market watcher

  1. Use the agent to fetch our top 5 customers’ market signals cross checking between Vainu (B2B sales intelligence system) and public Internet sources

  2. Subscribe alerts if something relevant changes


# Use Case: Subscribe to new Copilot Studio feature notifications

[Sorry about the screenshot language, our team works internally mostly in Finnish]

Please note that Loco is here participating in Teams chat that has nearly 10 members. So, it’s not about only having 1-to-1 chats with the agent – it’s also about having the agents collaborate as members of a larger team.

Here is the English translation of the texts in the screenshot:

User message (14:44):

Loco, create a ready-made wiki page where Arto can later document the Power Platform custom pipeline configuration so that deployment happens using a service principal.
At this stage, at least mention on the page that Arto Niemelä can provide more information.
You could also search online and add a couple of links to the page where this is explained.

Loco (14:46)
Wiki page created and pre-filled

I created a ready page for Arto’s upcoming documentation and added two Microsoft Learn links related to service principal deployment. ⏱️

📌 Location
/StudyTips/Power Platform/Custom Pipeline - Service Principal Deployment

🧱 Content at this stage

  • Clear purpose description

  • Instruction: “Ask Arto Niemelä for more information”

  • 2 Microsoft Learn links for guidance

  • TODO section for future additions

ℹ️ Note

I also had to create the parent page /StudyTips/Power Platform so that the new subpage could be added.
Would you like me to add a third link (e.g., a general overview of Power Platform pipelines) or adjust the page structure in preparation for Arto’s documentation? 🙂

User (14:47):

Provide the link to the page you created.

Loco (14:48)
🔗 Here is the link to the created wiki page

Custom Pipeline – Service Principal Deployment


# Use Case: Create Word document based on Power Platform solution package contents


# Use Case: Use Agent Cockpit to monitor and manage existing agents and to provision new ones

[This feature is still on the drawing board; the below screenshots are proposals from AI how things could look]


# Use Case: What's Yours?

Do you have a use case in your mind? Do you want to get it done?

Please, feel free to contact me at terho.antila@locoda.fi

Terho Antila is a Microsoft MVP and a top expert in Power Platform solutions. He has deep expertise in Copilot Studio, Power Apps, Power Automate, and Dataverse, as well as in their integration with Azure’s AI and automation services.

He is proficient in both low-code/no-code and pro-code implementations, enabling scalable and innovative solutions across diverse business environments. Terho’s technical expertise, combined with strategic thinking, makes him a valuable partner for organizations looking to fully leverage the Power Platform and Azure.