MCP Server for AI Agents

AI-powered access to your CRM data

What this is

The Solve360 MCP server is a secure bridge between AI assistants (like Claude Desktop, ChatGPT Desktop or other MCP-compatible clients) and your Solve360 account.

It allows an AI to query, search, manipulate, and analyze your CRM data in real time, using natural language — no need to learn how Solve360 works under the hood, just ask what you need and get answers.

Technically, it is:

• A Model Context Protocol (MCP) server
• Backed by OAuth2 authentication
• Acting as a controlled proxy to the Solve360 API
• Designed to return structured, human-readable responses

Why use it

CRM systems are powerful but rigid. They dictate how you’re expected to work with them.

That typically means:

• Learning and navigating a specific UI paradigm
• Relying on predefined views, dashboards, and reports
• Manually filtering data to answer simple questions
• Or falling back to API work to get anything truly custom

This MCP server changes that.

Instead of …

“Open report → filter → export → interpret manually”

You can just ask …

“Which deals are most likely to close this month?”

“What should I focus on today?”

“Show me overdue tasks across my team”

And get an immediate, structured answer. You don’t even have to type your questions as most AI tools have built-in speech recognition abilities. Talk to your CRM data!

Key value

• Natural language interface to CRM data
• Your custom workflow to process CRM data
• Faster decision-making (no manual report building)
• Context-aware insights (not just raw data)
• Friendly for non-technical users
• Works inside AI tools you already use

Core capabilities

1. Data retrieval

• Search Solve360 records and activities by complex criteria, including zone-based search
• Fetch full record details including complete activity history for contacts, companies, tickets, and project blogs/your custom type
• Receive formatted, user-friendly summaries instead of raw JSON data

Examples:

"Find contacts whose Business email starts with ‘vs’ and Organization name is not empty"

"Find companies tagged with Restaurant or Dining and assigned to Chris Maine"

"Show contacts created in the last month and located within 400 miles from me"

2. Activity and workflow insights

Build reports:

• Calendar with events and tasks
• Time-tracking insights
• Activities requiring your attention
• Deals
• Follow-ups
• Activities matching different criteria

Examples:

"What are my overdue tasks?"

"Find companies tagged with "Show all upcoming tasks and events I have in my calendar for this week" or Dining and assigned to Chris Maine"

"Show me a won/lost deals summary"

3. Data creation & modification

Beyond reading data, the server can perform controlled write operations in the CRM, including:

• Creating new records (contacts, companies, tickets, etc.)
• Updating existing records
• Adding or modifying activities (tasks, notes, follow-ups, deals, etc.)

This enables workflows like:

“Create a follow-up for this contact tomorrow”

“Update deal stage to 60%”

“Add a note to this company”

Write operations are:

• Executed through the same secure proxy layer
• Constrained by defined schemas
• Validated before reaching the API

This turns the MCP server from a reporting layer into an operational interface.

4. Data normalization & presentation layer

Transforms raw API responses into usable output:

• Timezone-aware date formatting
• Standardized numeric values
• Human-readable labels
• Reconstructed hierarchies
• Filtering of irrelevant/internal data

5. AI-Oriented tooling design

Designed specifically for LLM interaction:

• Strict tool schemas
• Enforced constraints (batching, timezone handling, etc.)
• Predictable structured outputs

This makes it reliable as both:

• A user-facing interface
• And a machine-to-machine integration layer

Example use cases

Everyday examples showing how users can ask simple questions and get structured, actionable CRM output.

Advantages

1. Secure by design

• OAuth2 authentication
• No direct API exposure
• Internal “gatekeeper” proxy layer

2. AI-native interface

• Designed specifically for LLM interaction
• Clear tool schemas and constraints
• Minimizes invalid or unsafe API usage

3. Efficient data access

• Keeps Solve360 API complexity hidden from end users
• Batch requests reduce API overhead
• Smart caching
• Minimal redundant calls

4. Clean output

• Structured, readable responses
• Context-aware formatting

Limitations

1. Requires MCP-compatible client

• Works best in environments like Claude Desktop or ChatGPT
• Not a standalone UI

2. Context matters

AI responses depend on:

• Clarity of user prompts
• Available data
• Correct interpretation of intent

3. Write operations carry inherent risk

AI misinterpretation of ambiguous requests may result in incorrect record updates, wrong field values, or unintended changes that require manual correction.

4. No real-time push

• Operates on request/response model
• Not event-driven (no live updates without polling)

When this is a great fit

Use this MCP server if you want:

• Conversational access to CRM data
• Faster insights without building reports
• AI-assisted sales or operations workflows
• A clean integration layer between AI and Solve360

Installation

Prerequisites

You need:

• A Solve360 account on a subscription plan that includes MCP
• An MCP-compatible client (e.g. Claude Desktop, ChatGPT Desktop or similar with a plan including MCP servers support)
• Download the Solve360 SKILL.MD file to your computer 
  
  Download

During the introductory period, the Solve360 MCP server is available to all clients to try regardless of their current plan.

In Claude Desktop (Pro/Max plan)

1. Navigate to Customize > Connectors
2. Click ”+”, then “Add custom connector”
3. Enter the Solve360 MCP server URL: https://mcp.solve360.com
4. Click “Add” to finish
5. Click “Connect” to authenticate
6. Ask Claude how it can help you with Solve360 :)

Skill

In addition to the MCP server itself, there is a dedicated Claude Skill for working with Solve360 data. It is not necessary but might help AI to use MCP tools more efficiently.

The Skill acts as an operational layer on top of the MCP tools:

• Teaching Claude how to use the tools correctly
• Enforcing CRM-specific rules and constraints
• Reducing mistakes and invalid operations
• Improving reliability and consistency of results

Think of it as:

• A “specialized operating manual” for Claude
• Combined with workflow automation and safety rules

How to Add the Skill to Claude Cowork

1. Navigate to Customize > Skills
2. Click ”+”, then Create skill > Upload a skill
3. Select the SKILL.md previously saved on your computer

Once present, Claude automatically detects and uses the Skill during conversations.

No additional installation is required.

Self-modifying feature

The Solve360 MCP skill has a unique ability: it learns and adapts to your preferences over time. When you define a preference or rule during a conversation, the skill automatically updates itself so all future interactions follow your new rules.

How it works:

1. You define a preference (mention Solve for better results)
   
   Tell the skill to always or never do something specific. Examples:
   
   
   • “Always show me Solve deal values in CAD, not USD"
   • "Always show only billable time records from Solve"
   • "Default to my timezone (America/Edmonton) for Solve without asking"
     
     
2. The skill records it
   
   Your preference is saved in the User Rules section of the underlying SKILL.md file. This is a living document of all your customizations.
   
   As the skill runs in a secure environment you will be asked to give access to its folder:
   
   
   
   Click “Allow” to let the skill remember your preference.
   
   
3. The skill applies it
   
   All future interactions follow your new rule automatically, you don't have to repeat it.
   
   

The Skill does not replace the MCP server. The Skill contains instructions and rules only.

Summary

The Solve360 MCP server turns your CRM into something you can talk to.

Instead of navigating dashboards and filters, you ask questions, and get structured, actionable answers.

Solve360 MCP Assistant workflow examples

Real-world examples showing how teams use the assistant to surface risks, automate follow-ups, and keep CRM records current.