AI-powered CRM analysis: 25 ready-to-use prompts for ChatGPT and Claude

Using MCP Server Uspacy, ChatGPT, Claude, or another AI assistant can work with the CRM data available to it. The assistant correlates deals, clients, comments, tasks, and activities to identify recurring issues, anomalies, and hidden opportunities.

All prompts collected here can be copied and used immediately. If needed, simply adjust the time period, the number of results, or the prioritization criteria.

The AI can only access information available through MCP Server Uspacy and within the permissions granted to the specific user. The quality of its insights also depends on consistent and accurate CRM data management.

A prompt like "Show me my deals" will simply return a list of records. An analytical prompt should encourage the assistant to compare data, identify exceptions, and explain its conclusions.

A well-written prompt should specify the time period, the objects to analyze, the expected outcome, and any constraints. For example, you can instruct the assistant not to make any changes in Uspacy without your explicit approval.

Copy this basic prompt:
— Review all Uspacy data available to me for the past three months. Identify non-obvious patterns, risks, and opportunities that impact sales. For each finding, provide the specific deals, clients, comments, tasks, or fields that support it. Prioritize the insights by importance. Do not make any changes in Uspacy.

To separate facts from assumptions, use this version:
— Analyze the Uspacy data available to me. For each finding, clearly indicate: confirmed facts, your hypothesis, your confidence level, and the information that is missing to verify it. Do not draw conclusions that are not supported by specific CRM records.

This format reduces the risk of receiving convincing but unverified generalizations.

A sales funnel may appear full even though many deals have been inactive for a long time. AI helps distinguish genuine opportunities from records that only make the report look more promising than it actually is.

Start by identifying inactive deals:
— Analyze all open deals available to me in Uspacy. Identify deals that have not been updated for a long time or do not have a recorded next step, meaning no scheduled activities. Categorize them into three groups: those that require attention today, those that should be reviewed further, and those that are likely no longer relevant. Explain the reason for each deal.

To identify problematic stages:
— Compare the number of open deals across all available stages of the sales funnel. Identify stages with a disproportionate accumulation of records. Show how long deals have remained unchanged at each stage. Suggest possible explanations, but clearly label them as hypotheses.

To monitor high-value opportunities:
— Identify the 10 open deals with the highest values. Check whether they have recent comments, active tasks, and scheduled activities. Show the deals that do not have a clear next step and explain the risk associated with each one.

To identify inconsistencies:
— Identify deals whose data does not match their current stage. Check whether key fields are completed and whether comments, tasks, and scheduled activities are present. Show exactly what appears to be inconsistent. Do not edit anything.

To validate the sales forecast:
— Analyze open deals by value, stage, and owner. Determine whether the sales forecast depends on a small number of high-value deals or on a single sales manager. Highlight the areas of highest concentration and explain the potential risks.

These insights depend on the data your team maintains in the CRM. Deal values, loss reasons, lead sources, and dates make the analysis significantly more accurate.

The highest-risk client is not always the one who complains. More often, it's the one who simply stops responding, agreeing to the next step, or returning with new orders.

To identify clients your team is losing touch with:
— Identify clients with open deals that have had no comments, tasks, or scheduled activities for an extended period. Prioritize them based on deal value, current stage, length of inactivity, and interaction history. Explain the reasoning behind the ranking.

For repeat sales opportunities:
— Identify contacts and companies associated with multiple closed deals but no new open opportunities or recent activity. Create a list of clients to re-engage. For each client, explain which data supports this recommendation.

To analyze client objections:
— Analyze the comments available in deals from the past six months. Group recurring client objections by topic, such as price, timelines, functionality, budget, approvals, or other reasons. Provide specific examples for each group.

To assess dependency on major clients:
— Group the available deals by company or contact. Identify the clients that account for the largest share of the total deal value. Determine where there is a risk of excessive dependency and support your conclusions with specific figures.

To identify expansion opportunities:
— Identify clients with multiple successful deals or a long history of interactions. Show which clients are good candidates for a repeat purchase, an additional service, or a new product. Do not assume client needs. Base your recommendations solely on deal history and recorded comments.

These prompts help you monitor risks while uncovering opportunities for growth.

The goal of this analysis is not to rank "good" and "bad" sales managers. Instead, AI should highlight workload imbalances, uneven work distribution, and inconsistencies in how the CRM is maintained.

To analyze workload distribution:
— Compare the number of open deals, active tasks, and overdue tasks assigned to the users available to me. Identify who may be overloaded. Support your conclusions with facts, and do not evaluate performance based solely on the number of records.

To monitor next steps:
— Identify owners who have open deals without current tasks, scheduled activities, or recent comments. Show the number of such deals and the five highest-risk examples for each owner.

To assess data quality:
— Analyze the open deals managed by different sales managers. Compare the completeness of key fields, the presence of comments, and recorded follow-up actions. Highlight recurring gaps and provide specific examples. Do not draw conclusions about employees' overall performance.

To identify deals where the team has lost context:
— Identify open deals where the owner has changed or where multiple employees have been involved. Check whether, after the handoff, an up-to-date comment, a next task, and a clear description of the current status were added. Show the cases where context may have been lost and specify what information is missing. Do not make any changes in Uspacy.

To assess dependency on a single employee:
— Determine whether key clients, high-value deals, or a significant number of active processes are concentrated under a single employee. Provide the supporting data and explain the operational risks that would arise if this person were temporarily unavailable.

This gives managers not just a general impression, but a list of specific topics to discuss with the team.

Tasks reveal more than just the team's workload. They can expose delays, dependencies on specific employees, and activities that consume time without producing meaningful results.

AI correlates deadlines, owners, clients, and related processes. As a result, isolated overdue tasks become clear management signals.

To identify tasks with unrealistic deadlines:
— Compare the planned and actual completion times for similar tasks over the past three months. Identify the types of work that consistently take longer than expected. Show the average variance and provide examples.

This prompt highlights planning issues rather than problems with individual employees.

To identify the causes of repeated deadline extensions:
— Identify tasks whose deadlines have been postponed multiple times. Group them by task type, owner, and related process. Highlight recurring patterns, and label any possible explanations as hypotheses.

The issue is not always a lack of discipline. It may result from unclear requirements, dependencies on colleagues, or waiting for a client's response.

To identify tasks without a clear outcome:
— Identify completed tasks that do not have a closing comment, a follow-up action, or any other recorded outcome. Show examples and specify what information is missing.

This scenario reveals a loss of context. The task is formally completed, but the team does not know what to do next.

To identify tasks that signal future problems:
— Analyze active and overdue tasks from the past three months. Identify topics or types of work after which related processes are more likely to stall. Highlight recurring patterns and provide specific examples.

These prompts turn a task list into a source of actionable management insights. They reveal where the team is losing context, making planning mistakes, or encountering systemic obstacles.

Once hidden bottlenecks have been identified, this type of analysis should become part of your daily or weekly operating routine.

A one-time analysis helps identify a problem. Regular analysis helps you spot it before it affects your sales targets.

For a weekly review:
— Analyze the Uspacy data available to me from the past seven days. Prepare a report in five sections: key changes, at-risk deals, new opportunities, data quality issues, and questions for the team. Support each conclusion with specific examples. Do not make any changes.

To prioritize your day:
— Review my open deals, tasks, and scheduled activities. Identify the five situations that require the most attention today. Consider deadlines, deal value, length of inactivity, and the risk of losing the client. Explain the priority order.

To validate a management hypothesis:
— I believe the main sales challenge is an insufficient number of new leads. Evaluate this assumption using the Uspacy data available to me. Present arguments both supporting and challenging this view. Suggest alternative explanations and identify what additional data would be needed to reach a definitive conclusion.

To uncover overlooked questions:
— Analyze the available CRM data as an independent consultant. What are the three important questions I haven't asked but should? For each question, show the signals in Uspacy that led you to it.

Start with three recurring workflows: your daily priority review, a weekly review, and sales funnel analysis. Once these prompts become part of your management routine, add client risk monitoring and CRM data quality reviews.