Altify's MCP Servers

Ask your AI assistant about your pipeline and act on the answer without leaving the chat.

Once Altify MCP is set up, you and your users can ask questions like:

  • "What are the biggest risks across my open opportunities this quarter?"
  • "Prep me for tomorrow's call with Acme. Who's involved and where do we stand?"
  • "What are the next best actions to move the Globex deal forward?"
  • "Draft a follow-up task on the Initech account plan and save it."

(See MCP Server Prompt Libraries for lots more examples).

Your assistant pulls the answer straight from your Altify and Salesforce data across Opportunity Manager, Account Planning, and Sales Process Manager, and it can update or create records when you ask it to. It all happens in plain-language conversation between your users and their AI assistant.

Note: Altify v9.18 is the minimum supported version for Altify's MCP Servers.

What Altify MCP is

MCP (Model Context Protocol) is the standard that lets an AI assistant securely connect to your data. Altify provides four MCP servers, each covering a different area. Your AI assistant connects to them, and they retrieve and update your Altify and Salesforce data on your behalf, within your permissions.

  • Altify Salesforce MCP searches and retrieves opportunities, accounts, and contacts using flexible queries, always within your Salesforce permissions.
  • Altify Analysis MCP applies Altify's selling methodology to your data to produce assessments, vulnerability analysis, call plans, and next best actions, grounded in your records rather than invented.
  • Altify Retrieve MCP reads detailed Altify account, opportunity, and sales process data for your assistant to use.
  • Altify Write MCP writes updates back to your Altify accounts, opportunities, and sales process so plans stay current.

You don't call these servers by name. You just ask your assistant a question, and it uses whichever servers it needs. For a full breakdown of each server and its tools, see Altify MCP Server and Tool Descriptions.

Before you begin

Setup crosses a few different systems, so be sure to have the following.

  • Salesforce admin access. You'll install a package and create a sign-in app in Salesforce Setup. If you're not a Salesforce administrator, you'll need one for the install steps.
  • An API key for your AI provider, for the Analysis server only. The Altify Analysis MCP server uses a large language model to do its work. Rather than dumping a large blob of raw data into your assistant, its tools gather and contextualize the relevant Altify data first, then return a focused result. That processing runs through your own AI provider account (OpenAI, Anthropic, Google, Azure, or Vertex), so you'll need an active account, an API key, and the model ID you want to use. The other three servers (Salesforce, Retrieve, and Write) query and update Salesforce directly and don't need this key.
  • Admin access to your AI assistant. An account on ChatGPT, Claude, or Microsoft Copilot Studio that lets you add a connector or app.
  • Your org type. Know whether you're setting up a production or sandbox org. The steps are the same. You'll just use the matching server addresses.

The two things you'll set up in Salesforce

Setup involves two separate pieces in Salesforce. It's worth knowing the difference up front, because they do different jobs. You'll set up the package first, then the External Connected App.

The Altify MCP Connector package is the package you install into your org. It creates the permission sets that control who can use Altify MCP, and it's where you add your AI provider API key and set token usage limits.

An External Connected App is an app you create separately. It lets your users authenticate to Salesforce through MCP using OAuth, so their assistant connects securely as them and respects their permissions.

Who does what

This flow touches three roles, which may be three different people:

  • Your Salesforce admin installs the Connector package, creates the External Connected App, and assigns access.
  • Whoever owns your AI provider account supplies the API key used by the Analysis server.
  • Each end user connects their own AI assistant once setup is complete.

Data and privacy

  • Your permissions always apply. Every request runs with your own Salesforce login, so you can only see or change data your existing Salesforce permissions already allow. Field-level security, sharing rules, and record access are enforced by Salesforce on every call.
  • Salesforce stays in control of sign-in. Salesforce is the sole identity provider. Your existing single sign-on, multi-factor authentication, and login policies apply unchanged, and Altify never sees or handles your login credentials.
  • Nothing new connects out from your Salesforce org. The Connector package adds no outbound connections. All access happens through Salesforce's standard, secure APIs, initiated by your AI assistant, using a token you've granted.
  • No data is stored by the MCP servers. Access happens per request and nothing is written to disk by the servers beyond the in-org audit log. Your AI provider API key is encrypted at rest in Salesforce and decrypted only briefly in memory when needed.
  • AI usage is capped and audited. Altify Analysis server tools call only the AI provider your admin configured, with your own API key. A daily token cap is enforced before any call, and every tool use is logged in your org.

How setup works

Work through the following steps. Each step links to its own page.

  1. Install and set up the Altify MCP Connector: Install the package, assign permission sets to admins and users, add your AI provider API key, and set daily token usage limits.
  2. Create the External Connected App: Set up the app that lets your users sign in securely through OAuth.
  3. Connect your AI assistant. Add Altify in the assistant you use: ChatGPT, Claude, or Microsoft Copilot Studio. You only need the guide for the one you use.

Optional: tailor the output to your methodology

By default, the Analysis server frames its coaching and recommendations in the Altify selling methodology: the four phases ("Is there an opportunity?", "Can we compete?", "Can we win?", and "Is it worth winning?") and the PRIME action framework. If your team sells with MEDDICC, Challenger, or a homegrown framework, you can change the language the AI uses once you're up and running.

For further information, see Methodology Customization for Altify MCP. This step is optional and isn't needed to get connected.