Analyze Unstuck Engine data with Claude

Analyzing your Unstuck Engine data with Claude Sonnet 4.5 transforms raw intent signals into strategic insights you can act on immediately. Claude Sonnet 4.5 is Anthropic's most capable AI model for data analysis, research, and complex reasoning—released in September 2025 with breakthrough capabilities for working autonomously on extended tasks. This article shows you exactly how to upload your Unstuck Engine exports to Claude and extract actionable intelligence that accelerates your sales and marketing motions.

Unstuck Engine captures real-time intent signals from 20+ sources, scoring leads across multiple ICP and persona models. Your exported data contains rich firmographic, technographic, behavioral, and engagement information—but making sense of hundreds or thousands of records manually is impossible. Claude Sonnet 4.5 processes your entire dataset in seconds, identifying patterns, trends, and opportunities that would take hours or days to uncover manually.

example analysis from Claude
​

Speed and accuracy: Claude Sonnet 4.5 analyzes thousands of records instantly, extracting insights with state-of-the-art accuracy on complex analytical tasks.

Deep pattern recognition: The model identifies hidden correlations between ICP scores, persona matches, signal types, industries, and engagement stages that human analysis would miss.

Actionable recommendations: Claude doesn't just analyze data—it suggests specific next steps tailored to your business objectives, from outreach strategies to campaign optimizations.

Extended autonomous operation: Claude Sonnet 4.5 can work independently for 30+ hours on complex analysis tasks, maintaining clarity and focus throughout.

No coding required: Get sophisticated data analysis through natural conversation, without writing a single line of Python or SQL.

When you export data from Unstuck Engine, you receive a CSV file containing 47 columns of lead intelligence. Each row represents a single intent signal event—like a LinkedIn profile view, connection request, or website visit—matched to a person and company, then scored against your ICP and persona models.

Critical data columns

Your export includes lead-level data across four main categories that Claude can analyze:

Signal information: Timestamp, signal type (LinkedIn activity, website visit, etc.), signal source, and intent classification (1st-party, 2nd-party, 3rd-party).

ICP scoring: ICP code, ICP score, maximum possible score, match reasoning, and version number for tracking model changes.

Persona scoring: Persona codes (can match multiple), persona scores, maximum possible scores, and detailed reasoning for each match or mismatch.

Person details: LinkedIn profile URL, name, job title, headline, connection count, follower count, location, skills, languages, and professional history.

Company details: Company name, website, LinkedIn URL, industry, employee count, founding year, headquarters location, description, and follower count.

Raw JSON data: The raw_data column contains the complete unprocessed signal data in JSON format for advanced analysis.

The raw_data column is a goldmine for deep analysis. It contains nested JSON with complete LinkedIn profiles, education history, job experience with durations, company funding information, and the full qualification logic showing exactly why each lead matched or failed to match your ICP and persona criteria.

Navigate to your Unstuck Engine dashboard and export your lead data as a CSV file. The export includes all signals captured within your selected time range with complete ICP and persona scoring.

Open Claude at claude.ai and start a new conversation. Click the attachment icon or drag and drop your CSV file directly into the chat. Claude supports CSV files of any size and can process thousands of rows instantly.

Note: Claude Sonnet 4.5 has a 200,000 token context window, which translates to approximately 150,000 words or several thousand lead records with full details.

Once uploaded, Claude has instant access to your entire dataset and can answer questions, identify patterns, and generate insights through natural conversation. No setup or configuration required.

Instead of asking "show me the data," ask questions tied to specific goals. Claude excels when you provide context about what you're trying to achieve.

Effective question examples

Which accounts show the strongest buying intent based on multiple 1st-party signals in the last 30 days?

What industries or company sizes have the highest ICP scores, and how should this inform our targeting strategy?

Analyze persona matches—are we attracting more Product Champions or Decision Makers, and what does this mean for our outreach approach?

Which signal types correlate most strongly with high-scoring leads that matched our top-tier ICP?

Compare engagement patterns between accounts that matched our primary ICP versus secondary ICPs.

Identify accounts with high ICP fit but low engagement—these are prime targets for focused ABM campaigns.

What geographic markets show the strongest concentration of qualified leads, and where should we expand sales coverage?

Start broad, then drill down. After Claude provides an initial analysis, ask follow-up questions to explore specific findings, request different visualizations of the data, or test hypotheses about what's driving results.

Ask Claude to translate insights into concrete next steps. Examples include prioritized account lists with specific outreach recommendations, campaign strategy adjustments based on data patterns, or ICP model refinements to improve targeting accuracy.

For complex analytical tasks, ask Claude to use extended thinking mode for deeper analysis. Simply say "think through this carefully" or "use extended reasoning" and Claude will engage step-by-step problem-solving that's visible to you.

Ask Claude to segment your leads across multiple dimensions simultaneously—for example, grouping by ICP tier, engagement stage, industry, and geographic region to identify your highest-value microsegments.

If your export spans weeks or months, Claude can analyze how lead quality, engagement patterns, and ICP distribution have evolved, helping you spot emerging opportunities or declining signal quality.

Upload your Unstuck Engine data alongside CRM data showing which leads converted to opportunities or customers. Claude can identify which combinations of signals, ICP scores, and persona matches predict conversion most reliably.

The raw_data column often includes competitive mentions and technographic details. Ask Claude to analyze which competitors appear most frequently in qualified accounts and what this reveals about competitive displacement opportunities.

Describe your unique business requirements and ask Claude to propose alternative scoring logic or weighting schemes based on patterns in your actual data, rather than theoretical assumptions.

"Analyze this Unstuck Engine export and identify the top 20 accounts I should contact immediately based on engagement recency and ICP fit."

"What percentage of our qualified leads come from 1st-party signals versus 2nd-party or 3rd-party signals, and how does this vary by ICP tier?"

"Show me which job titles are most common among our high-scoring persona matches—are we targeting the right roles?"

"Compare the characteristics of accounts that scored 30+ on ICP versus those that scored below 20. What differentiates high-fit from low-fit accounts?"

"Analyze engagement stage distribution. Are most of our qualified accounts still in stage E (no engagement yet) or have we successfully moved them up the funnel?"

"Which signal types drive the most high-value matches? Should we invest more in certain channels based on this data?"

"Based on the industries and company sizes in this data, recommend three distinct ABM campaign strategies tailored to different segments."

"Create a prioritized outreach sequence—which accounts should SDRs contact first, second, and third, and what personalization angles should they use based on the signal data?"

"Identify accounts with strong ICP fit but minimal engagement. What targeted campaigns could activate these dormant high-fit accounts?"

Ask Claude to create concise executive summaries of your lead data, highlighting key metrics, top-performing segments, and strategic recommendations in a format ready to share with leadership.

While Claude can't create interactive dashboards directly, it can design the structure and metrics for dashboards you'll build in tools like Tableau, Looker, or Google Data Studio. Claude analyzes your data and recommends exactly which KPIs to track and how to visualize them.

After Claude performs analysis, ask it to create formatted output you can export—like prioritized account lists in CSV format, markdown reports you can convert to PDFs, or structured data tables ready to import into your CRM or sales engagement platform.

Upload regularly: Analyze new exports weekly or monthly to spot trends and adjust strategies based on fresh data.

Combine data sources: Upload your Unstuck Engine export alongside data from your CRM, marketing automation platform, or analytics tools for cross-system insights Claude can synthesize.

Document insights: Ask Claude to create written summaries of key findings you can save and reference later, building institutional knowledge over time.

Test hypotheses: Use Claude to validate or challenge assumptions about your ICP, persona models, and go-to-market strategy using your actual performance data.

Iterate on questions: The most powerful insights often come from asking follow-up questions based on Claude's initial analysis. Don't stop at the first answer.

Pro tip: Claude Sonnet 4.5 has memory capabilities. In longer analysis sessions, it remembers previous context, allowing you to build complex, multi-step analyses without re-explaining your business or data structure each time.

Yes. Claude Sonnet 4.5 can analyze CSV files containing thousands of rows instantly. The model's 200,000 token context window accommodates most standard exports. For exceptionally large datasets (tens of thousands of records), you may need to split the file or ask Claude to analyze specific segments.

No coding or technical expertise required. Simply upload your CSV and ask questions in plain English. Claude handles all the data processing behind the scenes.

Claude Sonnet 4.5 achieves state-of-the-art performance on data analysis and reasoning tasks. However, always validate critical insights against your domain knowledge and business context. Claude is a powerful analytical partner, not a replacement for human judgment.

No. Claude currently requires manual file upload and cannot connect directly to Unstuck Engine or other platforms. An MCP (Model Context Protocol) integration is planned for future releases that would enable direct data access.

Anthropic implements enterprise-grade security measures. Uploaded files are processed securely, and conversations are private to your account. Review Anthropic's privacy policy at anthropic.com for complete details.

While Claude Sonnet 4.5 is recommended for complex analysis tasks, Claude Opus 4.1 offers even deeper reasoning capabilities for the most demanding analytical challenges. However, Sonnet 4.5 provides the best balance of capability, speed, and cost for most Unstuck Engine analysis workflows.

Yes, via the Anthropic API. Developers can build automated workflows that upload Unstuck Engine exports to Claude via API calls, execute predefined analysis scripts, and deliver formatted results to your team automatically. Contact your technical team or Anthropic's developer documentation at docs.anthropic.com for implementation guidance.

Claude pricing varies by subscription tier. Individual users can access Claude Sonnet 4.5 via the free tier at claude.ai with usage limits, or subscribe to Claude Pro or Claude Max for higher limits and priority access. API usage is billed per token. See anthropic.com/pricing for current rates.