Naro Agent Prompting Guide
Welcome to the Naro agent prompting guide! This guide will help you get the most out of your AI assistant by showing you how to structure your requests for maximum effectiveness.
The Naro agent is like having a highly skilled analyst at your disposal who can search through your organization’s knowledge, identify patterns, and help you create compelling content. The key to success is knowing how to ask for what you need.
Why this guide matters: The agent has powerful capabilities, but vague requests like “analyze our calls” won’t unlock its full potential. This guide will show you how to structure requests that lead to actionable insights and high-quality outputs.
Cheat sheet
| Agent Drivers | Details |
| Data sources | Sales calls (terms include calls, meetings, conversations, transcripts) Customer insights (terms include artifacts, topics, pain points, objections, questions, use cases, success stories, feature requests, competitor mentions) Content library (terms include shareable content, internal content, case study, product sheet, etc. ) Sales insights (term is sales insights – these insights gleaned from sales calls) Content briefs (terms include content brief, brief, or [brief name]) Content scoring (term is content score, this is the score of how well supported a conversation topic is by existing content) |
| Capabilities | Find / search across data sources Organize and group findings Web research Analyze results, find trends and gaps Summarize data sources, clusters, or findings Draft content Identify strategic insights / make recommendations |
| Terms | “Cluster”: Groups similar items into themes “Assess”: Evaluates content coverage and identifies gaps “Extract quotes”: Pulls actual customer language “Draft”: Creates new content based on findings “Compare”: Analyzes differences between groups or time periods “Prioritize”: Ranks items by importance or urgency “Synthesize”: Combines multiple sources into unified insights “For each”: Apply the requested task to each source (i.e. each call, each cluster) |
Understanding Your Data
Before diving into prompts, it’s helpful to understand what types of information the agent can access:
Sales Calls
- What they are: Recorded conversations with prospects and customers
- What’s captured: Full transcripts, participant details, discussion topics
- How to reference: “calls”, “meetings”, “conversations”
Artifacts (Customer Insights)
These are specific insights extracted from calls:
- Pain Points: Operational challenges customers face (“We waste hours on manual data entry”)
- Objections: Hesitations about your solution (“This seems too expensive”)
- Questions: Things customers ask (“Does this integrate with Salesforce?”)
- Use Cases: How customers plan to use your product (“We’d use this for employee onboarding”)
- Success Stories: Positive outcomes (“We reduced errors by 90%”)
- Feature Requests: Desired functionality (“We need bulk export capabilities”)
- Competitor Mentions: References to other vendors (“We’re also evaluating CompetitorX”)
Content Library
Your organization’s sales and marketing materials:
- Educational: Guides, documentation, FAQs, whitepapers, ebooks
- Sales Materials: Case studies, testimonials, product sheets, presentations
- Multimedia: Videos, demos, webinars, podcasts
- Research: Analyst reports, market research, competitive analysis
Important distinctions:
- Shareable Content: Customer-facing materials you can send to prospects
- Internal Content: Internal-only resources for your team’s use
Sales Insights
Reusable facts and knowledge extracted from conversations:
- Product capabilities and features
- Competitive differentiators
- Implementation details
- ROI metrics and success metrics
Content Briefs
Planning documents that outline upcoming content to be created, including target audience, key messages, and resources.
Content Scoring
Artifacts are automatically scored (0-4) based on how well your existing content addresses them:
- 0: No relevant content exists
- 1: Barely covered with vague information
- 2: Partially addressed but missing key details
- 3: Mostly covered but could use more depth
- 4: Fully supported with specific, detailed content
What the Agent Can Do
The Naro agent excels at these key capabilities:
1. Finding Information
- Search for specific topics across all your data
- Filter by date ranges, participants, or content types
- Find exact phrases or explore concepts
- Identify relationships between different pieces of information
2. Organizing and Grouping
- Cluster similar items into meaningful themes
- Identify patterns across multiple data points
- Group content by topic, urgency, or any other criteria
- Create structured views of unstructured data
3. Research
- Search the web based on topic
- Review specific links or websites provided
- Compare or incorporate identified information to existing conversation
4. Analysis and Synthesis
- Summarize large amounts of information
- Extract key quotes and evidence
- Identify trends and patterns
- Assess how well your content addresses customer needs
- Compare different perspectives or time periods
5. Content Creation
- Draft blog posts, emails, and sales materials
- Create responses to common objections
- Generate talk tracks for sales teams
- Develop compelling narratives from data
6. Strategic Insights
- Identify content gaps in your library
- Highlight emerging customer concerns
- Track competitive mentions and positioning
- Surface opportunities for new content creation
Effective Prompting Patterns
The most effective requests follow this pattern: Gather → Organize → Analyze → Create
Here’s how to structure your requests for best results:
Pattern 1: Find, Cluster, and Analyze
Goal: Understand patterns in customer feedback Example Prompt:
“Find all customer pain points from the last month, cluster them into themes, and provide a summary of each theme with recommendations for our product team.”
What happens: The agent will:
- Search for all pain point artifacts
- Group them into meaningful clusters (themes)
- Summarize each cluster
- Provide actionable recommendations
Pattern 2: Research and Create Content
Goal: Create content that addresses customer concerns Example Prompt:
“Search for security-related objections from the last quarter, cluster them by specific concern, extract key quotes from each cluster, then help me draft a security FAQ that addresses the top concerns.”
What happens: The agent will:
- Find all security objections
- Group them by specific security concern
- Pull actual customer quotes as evidence
- Draft FAQ content addressing each concern
Pattern 3: Assess and Improve
Goal: Identify content gaps Example Prompt:
“Find all use cases mentioned by enterprise customers in the last 3 months and assess how well our current content supports them. Focus on identifying the biggest gaps we need to fill.”
What happens: The agent will:
- Search for enterprise use cases
- Evaluate existing content coverage (0-4 scoring)
- Identify specific gaps
- Recommend content to create
Pattern 4: Competitive Intelligence
Goal: Understand competitive landscape Example Prompt:
“Find all competitor mentions from the last two months, cluster them by competitor and topic discussed, then create a competitive battle card highlighting our advantages for each main competitor.”
What happens: The agent will:
- Search for competitor mentions
- Organize by competitor and topic
- Analyze positioning discussions
- Create structured battle cards
Pattern 5: Time-Based Analysis
Goal: Track trends over time Example Prompt:
“Compare customer objections from Q1 versus Q2. Show me what new objections emerged, which ones disappeared, and how the frequency of different objection types changed. Create a trend report with recommendations.”
What happens: The agent will:
- Search two different time periods
- Identify changes and trends
- Analyze frequency shifts
- Create actionable report
Example Tasks by Department
For Product Marketing
Customer Voice Report:
“Find all feature requests from the last month, cluster them into capability areas, summarize each cluster with the business impact, and create a prioritized list for the product team. Include customer quotes for the top 5 requests.”
Positioning Analysis:
“Search for all competitor mentions where pricing was discussed, cluster by competitor and price point mentioned, then help me create updated positioning statements that address each competitive scenario.”
Use Case Development:
“Find success stories from the last quarter, cluster them by industry and use case, extract metrics and quotes from each cluster, then draft three detailed use case documents for our website.”
For Content Marketing
Blog Topic Mining:
“Search for customer questions from recent demos, cluster them by topic area, identify which clusters have the least content coverage, then create a prioritized list of blog topics with outlines for the top 5.”
Customer Story Development:
“Find all success stories and positive outcomes from the last 6 months, cluster by industry and impact type, extract compelling quotes and metrics, then draft 3 customer story frameworks ready for follow-up interviews.”
SEO Content Gaps:
“Look for technical questions asked in calls over the last quarter, cluster them by technical topic, assess our documentation coverage for each cluster, then create detailed briefs for documentation we need to write.”
For Sales Enablement
Objection Handling Guide:
“Find all objections from the last two months, cluster them by objection type and severity, analyze how our reps currently handle each type, then create an updated objection handling guide with talk tracks for each cluster.”
Competitive Battle Cards:
“Search for all competitor mentions from enterprise deals, cluster by competitor and differentiator discussed, extract specific comparison points, then create detailed battle cards for our top 3 competitors.”
Industry-Specific Playbooks:
“Find all healthcare industry calls from the last quarter, identify common pain points, questions, and use cases, cluster them by topic, then create a healthcare industry playbook with relevant case studies and talk tracks.”
For Customer Success
Implementation Guide Updates:
“Search for questions and challenges mentioned during implementation calls, cluster by implementation phase and topic, assess how well our current guides address each cluster, then create updated implementation documentation for the biggest gaps.”
Success Metrics Collection:
“Find all mentions of metrics, ROI, or business impact from the last 6 months, cluster by metric type and use case, create a metrics library with context for how each metric was achieved.”
Creative Marketing Use Cases
Build Your First FAQ from Real Customer Questions
The Challenge: Creating an FAQ that actually addresses what customers ask.
The Prompt:
“Find all questions asked by prospects in demos and discovery calls from the last 3 months, cluster them by topic, extract the exact questions from each cluster, then help me create a comprehensive FAQ page for our website. For each question, include the actual customer phrasing and craft an answer based on how our best reps respond.”
What This Creates: An authentic FAQ using real customer language, not what you think they ask.
Transform Call Transcripts into Customer Stories
The Prompt:
“Search for success stories and positive outcomes from calls in the last quarter where customers mentioned specific metrics or ROI. For each one, extract the challenge they faced, the solution we provided, and the results they achieved. Then draft 3 customer story templates I can use for case study interviews, including suggested interview questions based on what wasn’t fully covered in the call.”
The Result: Ready-to-use case study frameworks with targeted follow-up questions.
Create Industry-Specific Landing Pages
The Prompt:
“Find all calls and artifacts from healthcare companies in the last 6 months. Cluster their pain points, questions, and use cases by theme. For each major theme, extract relevant quotes and success metrics. Then help me create content for a healthcare industry landing page that speaks directly to their specific challenges using their own language.”
Why It Works: Uses actual healthcare customer language and concerns, not generic industry assumptions.
Develop Competitive Comparison Content
The Prompt:
“Find all competitor mentions from the last quarter where prospects compared us to [Competitor Name]. Cluster by comparison topic (features, pricing, support, etc.). For each cluster, extract what prospects said about both us and the competitor. Create a comparison guide that addresses each point with specific advantages we offer, using evidence from successful deals.”
The Output: Fact-based competitive content using real prospect feedback.
Generate Social Proof Campaigns
The Prompt:
“Search for all positive customer quotes and success metrics from the last 6 months. Cluster by impact type (time savings, cost reduction, efficiency gains, etc.). For each cluster, extract the best 3-5 quotes with speaker attribution. Create a social media campaign plan with 20 posts showcasing different customer wins, formatted for LinkedIn and Twitter.”
What You Get: Authentic social content with real customer voices.
Build Objection-Handling Email Templates
The Prompt:
“Find the top 10 objections from lost deals in the last quarter. For each objection, search for similar objections that were successfully overcome in won deals. Extract how our reps handled them successfully, then create email templates that address each objection using proven talk tracks and relevant customer success stories.”
The Result: Battle-tested email templates based on what actually works.
Create Persona-Based Content Strategies
The Prompt:
“Find all calls from the last 6 months and group them by participant title (CEO, VP Sales, IT Director, etc.). For each title group, cluster their main concerns, questions, and priorities. Create persona profiles with actual quotes, then develop content recommendations for each persona including blog topics, content formats they prefer, and key messages that resonate.”
Why It’s Powerful: Personas based on real conversations, not assumptions.
Develop SEO Content from Customer Language
The Prompt:
“Find all questions customers asked using the exact phrase ‘how to’ or ‘what is’ in the last 6 months. Group similar questions together and count frequency. For the top 20 question patterns, create SEO-optimized blog post outlines using the exact customer phrasing as headers, and include related questions they also ask in the same conversations.”
The Outcome: SEO content targeting actual customer search intent.
Build Sales Enablement Battlecards
The Prompt:
“For our top 3 competitors, find all mentions from the last quarter. For each competitor, identify: 1) Why prospects considered them, 2) What concerns they had about them, 3) Why they chose us instead (from won deals) or them instead (from lost deals). Create battlecards with specific talk tracks for each scenario, including trap-setting questions and proof points.”
What This Delivers: Battlecards with real-world proven strategies.
Generate Webinar Content from Hot Topics
The Prompt:
“Find all pain points and questions from the last 60 days, cluster by theme, and identify the top 5 most frequent clusters. For each cluster, extract key quotes, identify which industries care most about this topic, and create a webinar outline including title, agenda, and speaker notes based on how our team successfully addresses these topics in calls.”
The Result: Webinar content guaranteed to address current customer interests.
Create Newsletter Content from Recent Wins
The Prompt:
“Find all success stories, positive outcomes, and customer wins from the last month. Group by impact type and extract specific metrics and quotes. Create a monthly customer success newsletter with 5 sections: Metric of the Month, Customer Quote Spotlight, Use Case Deep Dive, Success Story Preview, and Implementation Tip based on what made these customers successful.”
Why It Works: Fresh, authentic content showcasing real customer success.
Develop Account-Based Marketing Content
The Prompt:
“Find all calls and interactions with [Target Company] or similar companies in their industry. Identify their specific pain points, priorities, and evaluation criteria. Search for success stories from similar companies. Create a personalized pitch deck outline with slides that directly address their unique situation using relevant examples and metrics from peer companies.”
The Output: Hyper-personalized ABM content based on similar customer wins.
Advanced Search Techniques
Based on real-world usage, here are powerful search patterns:
Finding Specific Content by Title
“What is the ID of content ‘[exact title]'”
This uses exact match (LEXICAL) search to find specific documents when you know the title.
Searching Calls by Keywords in Title
“Find calls with ‘sync’ in the title from June 2025”
Perfect for finding specific types of meetings (sync calls, QBRs, demos, etc.).
Getting the Most Recent Items
“Show me the IDs of the five most recent calls from [month]”
Useful for reviewing latest customer interactions or content updates.
Date Range Searches
“Find objections from April 18 to April 25, 2025”
Precise date filtering for time-sensitive analysis.
Filtering by Participant
“Find calls with [email@company.com] and list their objection artifacts”
Great for account-specific preparation or follow-up.
Complex Multi-Step Analysis
“Find up to 50 questions from May 2025 and cluster them. Summarize the clusters and list the calls they include.”
Handles large-scale analysis with multiple processing steps.
Tips for Better Results
Be Specific About Outputs
Instead of: “Analyze our content” Try: “Find all product demo videos, assess how well they address common technical questions, and create a plan for new demo content that fills the gaps”
Use Time Frames
Instead of: “Find customer feedback” Try: “Find all customer pain points from the last 30 days”
Specify Organization Methods
Instead of: “Group the results” Try: “Cluster the results by topic and urgency”
Ask for Specific Formats
Instead of: “Summarize this” Try: “Create an executive summary with key themes, supporting data, and three recommended actions”
Chain Actions Together
Instead of: “What do customers think?” Try: “Find customer objections from last quarter, cluster them by theme, summarize the top 3 clusters, and draft email responses for each”
Power Words That Help
Certain words help the agent understand exactly what you want:
- “Cluster”: Groups similar items into themes
- “Assess”: Evaluates content coverage and identifies gaps
- “Extract quotes”: Pulls actual customer language
- “Draft”: Creates new content based on findings
- “Compare”: Analyzes differences between groups or time periods
- “Prioritize”: Ranks items by importance or urgency
- “Synthesize”: Combines multiple sources into unified insights
- “For each”: Apply the requested task to each source (i.e. each call, each cluster)
Common Mistakes to Avoid
Too Vague
❌ “What are customers saying?” ✅ “Find all customer objections from last month, cluster them by concern type, and create a summary of each cluster”
Missing the Organization Step
❌ “Find all pain points and summarize them” ✅ “Find all pain points from Q4, cluster them by business impact, then summarize each cluster”
No Clear Output
❌ “Look at our competitor mentions” ✅ “Find competitor mentions from the last 60 days, cluster by competitor and topic, then create a competitive positioning document”
Forgetting Assessment
❌ “Find customer use cases” ✅ “Find customer use cases from enterprise accounts, assess how well our content supports each one, and identify the top 3 content gaps to fill”
Getting Started
Start with these simple patterns and build from there:
- Basic Pattern: “Find [what] from [when], cluster by [method], and summarize each cluster”
- Analysis Pattern: “Find [what] from [when], cluster by [method], analyze [specific aspect], and create [output]”
- Content Creation Pattern: “Find [what] from [when], cluster by [method], extract [evidence], and draft [content type]”
Remember: The agent works best when you’re specific about what you want to find, how you want it organized, what analysis you need, and what output you’re expecting. The more structured your request, the better your results will be!
