How to Connect Salesforce to Claude in 1 Minute (AI-Powered CRM Analysis)

Every sales leader has lived this moment.

Quarter-end is approaching. The leadership asks which deals are actually going to close, not which ones the team thinks will close. You pull up Salesforce, run three reports, export two CSVs, and spend an hour in a spreadsheet trying to reconcile pipeline stage against close date, owner, and last activity date.

The data is all there.
The problem is that Salesforce shows you what happened, but not what it means.

• Which rep’s pipeline has the most stale opportunities sitting in “Proposal Sent” with no activity in 30 days?
• Which lead source is producing the highest average deal value, not just the most leads?
• Which accounts are churning silently, showing zero open opportunities after a year of active buying?

Answering all these questions takes hours in Salesforce’s native reports.

With Windsor MCP, your Salesforce data streams directly into Claude — automatically, with no code — and you can get deep insights into any area of your CRM in just seconds.

🚀 Automate your Salesforce analysis in Claude with Windsor MCP. Try it free for 30 days.

Instead of exporting reports or building custom SQL queries, you can connect Salesforce to Claude using the Windsor.ai connector and start asking any questions about your pipeline, leads, and revenue in natural language, and even generate such insightful visual reports on top of your CRM data.

Get started: Connect your Salesforce data to Claude with Windsor MCP

Windsor MCP acts as the data layer between Salesforce and Claude, pulling live CRM data from the Salesforce API and delivering it to your Claude environment already structured and normalized.

All you need to do is follow three simple steps.👇

📖 Full walkthrough: windsor.ai/documentation/windsor-mcp/how-to-integrate-data-into-claude/.

Prerequisites

Before you begin, make sure you have access to:

• A Windsor.ai account (free trial or paid plan)
• A Salesforce account with API access
• A Claude account

Then follow these steps to connect Salesforce to Claude:

1. Go to onboard.windsor.ai and select Salesforce as your data source.

2. Authenticate with your Salesforce credentials and approve the connection. Select the account(s) that you’d like to sync to Claude.

3. In Claude, open the Windsor.ai connector page and add it with the Connect button.

💡 Quick tip: Set permissions to “Always allow” so your Salesforce data will be live in every Claude conversation automatically.

You’re set. Your Salesforce data is now connected to Claude and ready to query.

Salesforce prompt ideas: What you can do with your CRM data in Claude

The use cases and prompts below help sales leaders, RevOps teams, and CRM analysts get faster answers from Salesforce using the fields Windsor streams to Claude.

If you’re a sales leader managing the forecast and pipeline

Challenge: Your data exists, but generating a clear, real-time forecast still takes hours every week. Which deals will close this quarter, and which are quietly dying?

Pipeline health check: stale opportunities by stage and rep

Deals go stale, but which ones, and who owns them?

Prompt:

Show me all open opportunities approaching close this quarter that haven't had recent activity. Group by rep and rank by total pipeline value. Highlight reps with multiple stale deals so I can coach them this week.

Forecast risk: deals with optimistic close dates and no recent activity

Some deals look like they’re closing soon but show little recent engagement.

Prompt:

List opportunities in late-stage sales that are set to close soon but haven’t seen activity lately. What is the total value of deals at risk?

Win/loss analysis by lead source and competitor

Understand which channels are performing the best and why deals are lost.

Prompt:

Analyze closed deals from the past six months. Calculate win rates and average deal size by lead source. Identify the top reasons for lost deals.

If you’re running RevOps or sales operations

Goal: Make the pipeline visible and the sales process predictable.

Pipeline velocity: how fast are deals moving?

Track the speed at which opportunities progress through each stage.

Prompt:

Calculate average time in each stage for deals that closed recently. Identify the stages with the highest drop-off rate.

Lead conversion funnel

The Salesforce lead-to-revenue funnel spans multiple objects. Windsor lets you connect them in a single prompt and measure how leads convert into opportunities and revenue.

Prompt:

Show lead-to-revenue conversion rates over the last 12 months, broken down by lead source. Which sources produce the highest-quality pipeline?

Account health: identify silent churn risk

Find valuable accounts with no active opportunities before they churn.

Prompt:

List customer accounts with no recent wins or active opportunities, sorted by revenue. Flag high-value accounts with no activity.

Activity coverage: are your reps engaging accounts?

Check whether your reps are touching opportunities consistently.

Prompt:

Highlight accounts with open opportunities that haven’t had recent logged activity. Flag reps with multiple unengaged deals.

If you’re a sales rep or account executive

Your Salesforce data can tell you where to focus, but right now, it takes too long to get that read every morning. Here’s how Claude gives you a daily briefing in 10 seconds.

Daily deal briefing

Prompt:

Show all my open deals sorted by expected close date. Highlight any I haven’t engaged with recently. What should I prioritize today?

Preparing for a QBR or executive meeting

Prompt:

Summarize a given account’s opportunities and activities over the past two years. Include revenue, pipeline, and any open risk. Create a concise 3-sentence account summary for executives.

If you’re connecting Salesforce to ad data (marketing + sales alignment)

The biggest blind spot in B2B marketing is that Google Ads or LinkedIn shows you leads and form fills, but Salesforce is where you find out which ones turned into money.

Connect Salesforce alongside your ad platforms in Windsor, then ask Claude the following.

Closed-loop attribution: which campaigns produce revenue?

Prompt:

Match connected marketing campaigns to closed deals. Show which campaigns generate revenue versus just lead volume.

LinkedIn Ads: persona insights

Prompt:

Based on our LinkedIn Ads data, analyze which job titles and seniority levels are converting to closed deals in Salesforce. Are we targeting the right personas?

If you’re managing customer success or account renewals

Goal: Identify churn risks and prioritize renewals with actionable insight.

Salesforce holds the case history, renewal dates, and account health signals your CSMs need, but getting a coherent view of a single account requires jumping between Opportunities, Cases, Contacts, and Activities.

Renewal risk report: accounts up for renewal in the next 60 days

Prompt:

Show upcoming renewals and any accounts with high open case volume. Flag the riskiest renewals for immediate CSM action.

Case volume trends: which accounts are generating the most support load?

For this, you’ll need to connect Salesforce with your customer support system, like Zoho, Zendesk, etc.

Prompt:

Analyze accounts by case volume, resolution time, and priority. Identify accounts at high risk of churn.

🤖 For more advanced prompt ideas across various data sources, browse the full Windsor AI Prompt Library.

How Windsor MCP works with Salesforce and Claude

Instead of exporting Salesforce reports manually or writing SQL queries, Windsor connects directly to the Salesforce API and streams structured CRM data into your Claude conversations.

The process works like this:

1. Windsor connects to your Salesforce property via the Salesforce REST API.
2. It retrieves structured data from standard Salesforce objects: Opportunities, Leads, Contacts, Accounts, Activities (Tasks and Events), Cases, and Campaigns.
3. The Windsor MCP connector makes this data available inside Claude.
4. Claude queries and analyzes the data using natural language.

Core capabilities of Windsor MCP for Salesforce

• Live CRM data. Windsor pulls data directly from the Salesforce API, so Claude always analyzes the most current pipeline, activity, and deal data without repetitive exports.
• Access to the most essential metrics. Windsor exposes a broad range of Salesforce fields across all standard objects: opportunity amounts, stage history, lead status, contact data, account details, activity records, case data, and more.
• Cross-object analysis. Salesforce stores data in separate objects: Leads, Opportunities, Accounts, Contacts, Cases, and Activities. Windsor joins these for Claude, so you can ask questions that span the full customer lifecycle, from first lead touch to closed deal to renewal.
• Cross-platform data blending. Because Windsor connects to 325+ marketing and business data sources, you can combine Salesforce data with advertising platforms (Google Ads, LinkedIn Ads, Meta Ads) for closed-loop attribution, marketing automation (Pardot, HubSpot, Marketo) for lead quality analysis, finance systems (Stripe, QuickBooks, NetSuite) for revenue reconciliation, or customer success tools (Zendesk, Intercom) for churn risk signals.

Key Salesforce data fields Windsor sends to Claude

Windsor gives Claude access to 30+ metrics and 260+ dimensions across Salesforce’s standard objects. Below are the most commonly used fields for real CRM analysis, organized by object.

Opportunity fields — Core of pipeline analysis

• opportunity_name — Deal identifier
• opportunity_stage_name — Stage in the sales process (Prospecting → Closed Won/Lost)
• opportunity_amount — Expected deal value
• opportunity_close_date — Projected close date
• opportunity_type — New Business, Renewal, Upsell
• opportunity_probability — Chance of closing (%)
• opportunity_lead_source — Lead source attributed to the opportunity
• opportunity_description — Deal description (can include loss reason)
• opportunity_last_activity_date — Last logged activity
• opportunity_created_date — When the opportunity was created
• opportunity_owner_id — Opportunity owner (rep)
• opportunity_account_id — Account linked to the opportunity

Lead fields — Top-of-funnel intelligence

• lead_first_name, lead_last_name — Lead identity
• lead_email, lead_phone — Contact info
• lead_company — Company associated with the lead
• lead_title — Job title/persona analysis
• lead_lead_source — How the lead was acquired
• lead_status — Lead status (New, Working, Converted, Disqualified)
• lead_created_date — Lead creation date
• lead_converted_date — Date lead converted to an opportunity
• lead_converted_opportunity_id — Opportunity the lead converted into
• lead_industry — Industry / vertical

Contact fields — Customer base

• contact_firstname, contact_lastname — Contact name
• contact_email, contact_phone, contact_mobilephone — Communication channels
• contact_title, contact_department — Role and function
• contact_accountid — Linked account
• contact_ownerid — Owner/rep
• contact_lastactivitydate — Most recent activity logged

Account fields — Company-level analysis

• account_name — Account/company name
• account_type — Customer, Prospect, Partner
• account_industry — Industry vertical
• account_annual_revenue — Size/tiering
• account_number_of_employees — ICP / fit analysis
• account_owner_id — Account owner/rep
• account_created_date — Creation date
• account_last_activity_date — Last activity date

Activity/task fields — Engagement signals

• task_subject — Task description
• task_activity_date — Due date
• task_call_type — Type of touchpoint (Call, Email, Meeting)
• task_is_high_priority — High-priority signal
• task_status — Task status (Completed, Not Started, In Progress)
• task_owner_id — Assigned rep
• task_account_id — Linked account
• task_created_date — Task creation date

Campaign fields — Marketing attribution

• campaign_name — Campaign name
• campaign_start_date, campaign_end_date — Timing
• campaign_budgeted_cost, campaign_actual_cost — Spend metrics
• campaign_number_of_leads, campaign_number_of_converted_leads — Lead volume and conversions
• campaign_number_of_opportunities, campaign_number_of_won_opportunities — Pipeline/revenue contribution
• campaign_owner_id — Owner of the campaign

Campaign member fields — Lead/Contact participation

• campaignmember_name — Lead or contact full name
• campaignmember_email — Email
• campaignmember_company_or_account — Company / Account
• campaignmember_status — Campaign engagement status
• campaignmember_lead_source — Lead source attribution
• campaignmember_first_responded_date — First engagement date

Conclusion

Salesforce is the most complete record of your revenue process. You can track every lead that entered the funnel, every deal that was won or lost, every account that went quiet. The problem has always been that turning that record into an insight still requires building a report, running a query, or exporting a CSV.

Windsor MCP gives Claude live access to your Salesforce data. You still need to know which questions to ask, but now asking them takes 60 seconds instead of hours.

For sales leaders, that means a real-time pipeline view without waiting on RevOps. For RevOps teams, it means answering ad-hoc questions instantly without spinning up a new report every time. For marketing teams, it means finally seeing which campaigns produce closed revenue, not just leads.

🚀 Connect your Salesforce data to Claude with Windsor MCP in 1 minute — Start your 30-day free trial now: https://onboard.windsor.ai/app/salesforce.

FAQs

What’s the fastest way to get Salesforce data into Claude?

Windsor MCP. It provides a 60-second setup at onboard.windsor.ai, then your Salesforce data is live in every Claude conversation. The alternative is exporting reports from Salesforce as CSVs and uploading them manually, which gives you stale data and no cross-object or cross-source joins. The Salesforce API route is possible but requires OAuth setup, SOQL knowledge, and ongoing maintenance. Windsor handles all of that automatically.

Which Salesforce objects does Windsor pull?

Windsor supports all standard Salesforce objects: Opportunities, Leads, Contacts, Accounts, Activities (Tasks and Events), Cases, Campaigns, and Campaign Members. It also supports the retrieval of custom fields and custom objects; connect your account at onboard.windsor.ai/app/salesforce to make custom fields available.

Can Claude analyze a pipeline across multiple stages and owners at once?

Yes. Because Windsor sends structured data across all Salesforce objects, Claude can query the pipeline by stage, by owner, by close date, by account, or by any combination of fields simultaneously without you needing to build separate reports for each slice.

Can I blend Salesforce data with my ad platforms in Claude?

Yes, and this is one of the highest-value use cases. Windsor connects to 325+ data sources. The most common Salesforce blends are with Google Ads and LinkedIn Ads (for closed-loop attribution using utm_campaign as the join key), HubSpot or Pardot (for lead quality and marketing-sales alignment), and Stripe or QuickBooks (for revenue reconciliation). Connect multiple sources in Windsor, and Claude can analyze across all of them in a single conversation.

Does Windsor support custom Salesforce fields and objects?

Yes. Standard fields are available immediately after connecting. Creating custom fields and custom objects is also supported within the app.

Can Claude write back changes to Salesforce?

No. Windsor MCP is read-only. Claude can analyze your pipeline, identify stale deals, flag at-risk renewals, and generate a prioritized list of actions, but all updates happen in Salesforce, made by you. This is intentional: no automated tool should modify CRM records without human review.

Does Windsor pull Salesforce opportunity stage history?

Windsor pulls the current stage data and the opportunity_last_activity_date field. For full stage history analysis (time-in-stage across every transition), connect your Salesforce instance and explore the OpportunityHistory object availability in your account configuration.

How does this compare to Salesforce’s built-in Einstein AI?

Einstein AI is embedded inside the Salesforce UI and focuses on surface-level predictions and in-app recommendations. Windsor MCP for Claude works differently: it pulls your raw Salesforce data into Claude’s reasoning engine, letting you ask any question in natural language and get a structured answer, including questions that span multiple objects, blend with external data sources, or require analysis that no pre-built dashboard covers. It’s more like a senior analyst who knows your CRM inside out than an in-app tip.