---
title: "What is Spud and how to get started?"
last_updated: "2026-03-26T09:09:35.730Z"
category: "Getting Started"
canonical_url: "https://maybeclients.com/h/mash-help-centre/what-is-spud-and-how-to-get-started"
help_centre: "MaSH! Help Centre"
---

# What is Spud and how to get started?

Spud is the name for MaSH!'s AI Agents, powered by Maybe* and living inside Microsoft Teams. Tag @Spud, and it gets to work. That is all you need to know to get started.


THE BASICS

What is Spud?

Spud is what we call our AI Agents. They are powered by Maybe* and live inside Microsoft Teams. There are over 20 Agents, each one configured for specific tasks, and they connect to your real tools, including Float, Box, and the web. To set any of them to work, just type @Spud in Teams.

---

What can Spud help me with?

A lot. There are over 250 tasks across the full Agent set, from administrative work like resource planning and file management in Float and Box, to creative tasks like generating videos, building synthetic audiences, and writing briefs. If you are not sure whether Spud can do it, ask. The worst outcome is it tells you it cannot.

---


GETTING STARTED

How do I access Spud in Teams?

Type @Spud in any Teams message. Once installed, that is all it takes. Spud determines which Agent is best for your request and gets started.

---

Do I need to install anything?

Yes, a one-time setup is needed. You should have received an email invitation with login details and steps to get connected. If you have not received it, contact xxx and they will get you sorted.

---

How do I start using it for the first time?

Three steps: accept your invitation email, install Spud in Teams, and connect your systems. Once that is done, type @Spud and start asking. A colleague who is already set up can walk you through it in a few minutes if you get stuck.

---


HOW TO USE SPUD

Talking to Spud

How do I start a conversation with an Agent?

Type @Spud in Teams and write what you need, just as you would ask a colleague. Spud works out the rest.

---

Do I need to use specific prompts or commands?

No. Talk to Spud in plain language. You do not need to learn commands, keywords, or special formatting. If your request needs more detail, Spud will ask.

---

How do I know which Agent to use?

You do not need to. @Spud routes your request to the right Agent automatically. Over time you will get a sense of which Agents are doing the work, but you never need to pick one manually.

---

What types of tasks can Spud complete?

Spud can handle a wide range across the business, resource and project management in Float, file retrieval in Box, content and social tasks, briefing, research, synthetic audience work, video generation, compliance checks, news summaries, and much more. Many tasks also run on a schedule and land in Teams automatically.

---


QUICK START

Accept your invite  →  Install Spud in Teams  →  Connect your systems  →  Type @Spud and ask anything

---


QUALITY & ACCURACY

Can I trust the output?

How accurate is Spud?

Spud is highly capable, but it is not infallible. Accuracy depends on the task type and the quality of information available to the Agent. For tasks connected to live data, like Float or Box, output is only as current as that data. For generative tasks, quality improves significantly when you give Spud more context.

---

Should I review the results before using them?

Yes, always. Treat Spud's output as a strong first draft, well-structured, informed, and useful, but apply your own judgment before anything goes to a client or gets published. You are the professional. Spud is here to reduce the work, not replace your expertise.

---

How does Spud generate its answers?

Spud's Agents are powered by Maybe*, which connects multiple large language models with your real business tools, Float, Box, the web, and more. When you ask a question, the relevant Agent pulls the right data, applies the right logic, and delivers a finished output inside Teams. You see the result, not the machinery.

---

What should I do if I get an answer that seems wrong?

First, check whether the underlying data is correct, many issues trace back to missing or outdated entries in Float or Box rather than the Agent itself. If the source data is fine and the output is still wrong, note exactly what you asked and what came back, and flag it to your Champions lead. That feedback directly improves the system.

---


BEST PRACTICE

Getting the best from Spud

A few habits that make a real difference:

- Give context. The more specific your request, the better the output. 'Write me a brief for BMW M' will produce something useful. 'Write a brief for BMW M targeting urban performance drivers aged 30 to 45, for a social-first campaign launching in Q3' will produce something excellent.
- Iterate. If the first response is not quite right, follow up. Tell Spud what to adjust, length, tone, focus, and it will refine.
- Ask before assuming it cannot. Over 250 tasks are configured. If you are not sure whether Spud can help, try it.
- For platform tasks, Float, Box, Instagram, name the platform in your message. Spud needs that signal to trigger the right integration.
- Review before sharing. Always apply your judgment before anything goes to a client or gets published.

KNOWN LIMITATIONS

What Spud cannot do

Knowing the limits is as useful as knowing the capabilities:

- Spud cannot access systems that have not been connected. If you ask about data in a tool that is not integrated, it will not find it.
- Platform-connected Agents, Float, Box, Brand Guardian, require you to name the platform in your message. Without it, the integration will not trigger.
- Generative output is only as good as the input. Vague requests produce generic results. Specific requests produce specific, useful results.
- Spud does not post, publish, or take action on your behalf unless a task has been explicitly configured and scheduled to do so.
- For legally sensitive or regulated content, Spud's output is a first pass, not clearance. Always involve your legal team for high-stakes work.

REMEMBER

Spud is here to remove the boring, repetitive, and time-consuming work, so you can focus on the thinking that actually needs you. Use it. Give it feedback. It gets better the more it is used.

---
