For MSPs, the promise of AI is undoubtedly exciting. Getting AI to do something useful with the systems your business actually relies on, however, is another matter.
AI tools are great at answering questions, drafting content and brainstorming ideas. But it’s no secret that their value grows considerably when they can connect to the business data and tools behind your day-to-day operations.
That’s where the Model Context Protocol (MCP) comes in.
MCP is an open standard for connecting AI applications to external systems, including data sources, tools and workflows. In practical terms, an MCP server can give an AI application access to relevant information and tools outside the model itself. The official MCP documentation compares it to a USB-C connection for AI: a standardized way for applications and external systems to work together.
For Sherweb partners, that idea is becoming very practical. The Sherweb Cumulus MCP Server connects supported AI assistants and automation tools to Sherweb’s Partner Portal (AKA Cumulus), allowing partners to use natural-language requests to work with information including organizations, subscriptions, the product catalogue, billing, audit logs and renewal details. What information and actions are available depends on the user’s specific Partner Portal permissions.
In other words, instead of constantly moving between an AI assistant and your Sherweb environment, you can bring the two together.
Here are six ways that can make life easier for your MSP.
1. Find subscription information without digging for it
Sometimes the simplest questions consume more time than they should.
Which subscriptions does a customer have? What information is available about a particular organization? What’s coming up for renewal?
Connecting Sherweb’s Partner Portal through the MCP Server gives supported AI tools access to operational information such as tenants, subscriptions and renewal details, opening the door to asking focused, natural-language questions rather than manually searching for the same information.
The benefit isn’t AI for AI’s sake. It’s spending less time looking for information and more time using it.
2. Get quicker visibility into billing information
Billing is another area where information retrieval can turn into administrative work.
Sherweb’s MCP Server can make billing information available to supported AI hosts alongside other Partner Portal data. The capability is intended to help partners access information faster and reduce manual searching.
Think of the difference between navigating several screens looking for a piece of information and starting with a focused question that gets you closer to what you need. For MSP teams already balancing service delivery, sales and operations, reducing friction from small recurring tasks can make a big difference in the long run.
3. Check renewal information more easily
Renewals are another strong fit for natural-language access to business data. Utilizing the MCP Server, a partner can use an AI host to query relevant tenant renewal information available to their account.
This is an important distinction between simply asking a general-purpose AI model what to do about renewals and connecting that AI experience to your own operational information—and, where your permissions allow, act on it.
Instead of asking AI a hypothetical question, you can therefore ask it to work with relevant customer data without switching between tools.
4. Explore your product catalogue conversationally
Finding the right solution for a particular situation can also mean jumping between systems and screens.
Product catalogue information is also among the data accessible through the MCP Server. For an MSP, the value is straightforward: another common information-retrieval task can start with a conversation instead of a manual hunt.
5. Bring Sherweb’s Partner Portal data into the AI tools that suit your workflow
There isn’t one universal way MSPs work with AI, and there shouldn’t have to be.
Sherweb’s MCP Server supports the following host applications:
| Host application | Best for |
| VS Code with GitHub Copilot extension | Developers and technical users |
| Microsoft Copilot Studio | Low-code or no-code agent building |
| Cursor | Developers and AI-first IDE users |
| Claude Desktop or Claude Code | General assistant and developer scenarios |
| ChatGPT | General-purpose AI assistant |
Other MCP-compatible hosts may also work, although their configuration and behaviour can differ.
That flexibility matters because an MCP server sits between your AI experience and the connected system. MCP itself is designed around this client-server architecture rather than tying the connection to one AI application.
6. Build toward repeatable AI-powered workflows
Once AI has access to relevant information and tools, the next opportunity is automation.
Sherweb’s MCP Server uses read capabilities as a foundation for custom workflows and AI agents tailored to an MSP’s business needs. The goal is to turn repetitive processes into repeatable automations that can reduce manual effort and free up team resources—which is much closer to the operational promise of AI for MSPs.
Rather than treating an AI assistant as another tab where employees occasionally ask questions, MCP creates a way to connect AI to the systems where work happens.
Planning for future write access and permissions
MCP tool permissions are designed to align with the permission scopes available through the public API. Today, Sherweb partners can control which MCP tools are made available to users.
Based on feedback from early access participants, Sherweb will expand the permission model in an upcoming release to support write tools. Write access will rely on permission-based credentials and tool-level enforcement, helping ensure that users can only perform actions they’re authorized to complete.
For now, integrations should be designed around available read-only tools. Sherweb partners will be notified ahead of this major update, and provided with updated onboarding guidance as these capabilities become available.
How to get started with Sherweb’s MCP Server
You don’t need to automate your entire operation on day one!
We recommend starting out with a narrow, read-only request and including useful filters, such as an organization, date range or product, to make the response more precise.
Before connecting, you’ll need an active Sherweb Partner Portal account with API access, a compatible MCP host, API credentials and the appropriate permissions. Interactive sign-in scenarios may also require a redirect URI.
From there, the basic flow is:
- Open Security > APIs in the Partner Portal and set up the required API credential.
- Add the Sherweb Cumulus MCP Server to your chosen host.
- Authenticate using the appropriate flow.
- Confirm that your host can discover the Cumulus tools available to your account.
- Start with a focused, read-only question.
One important security reminder: treat Client Secrets and access tokens like passwords. Don’t paste them into support tickets, screenshots, recordings or chat messages.
For complete setup instructions, Sherweb partners can check out our associated knowledge-base article.
Put AI to work for your MSP
AI becomes much more useful when it can connect to the information and processes your MSP relies on.
Sherweb’s MCP Server is designed to create that bridge between supported AI tools and the Partner Portal, giving partners another way to find information, reduce repetitive work and explore more sophisticated automation opportunities.
It also fits a broader Sherweb goal: making it simpler for partners to do business and freeing up resources they can put toward their clients and growth.
Not yet a Sherweb partner? Discover what a Sherweb partnership can bring to your MSP and get started today.
