Nexoryn Executive Intelligence Platform Case Study
An AI-powered executive briefing agent. It pulls daily sales data from Google Sheets, aggregates live crypto, currency, weather, GitHub, and tech-news signals in parallel, has an LLM write a structured executive summary and recommendation, then renders and emails a fully branded HTML dashboard, hands-free, on schedule.
Service
Automation
Industry
SaaS & Tech
Category
AI Automation
Built with: Groq / Llama, n8n, Google Sheets, Email (SMTP).
Overview
The Problem
Executives need a daily business pulse, but compiling sales figures, market data, and external context by hand is slow and inconsistent
Sales data sits in spreadsheets, disconnected from the crypto, currency, and industry signals that inform real decisions
No reliable reporting cadence, insight quality depends on whoever has time to assemble the numbers that day
Raw metrics without narrative force leaders to interpret the 'so what' themselves
The Solution
Automatically pulls daily sales data from Google Sheets and computes core KPIs, revenue, transaction count, average sale, highest and lowest sale
Aggregates external context in parallel: crypto prices, FX rates, local weather, GitHub activity, and top tech news
AI (Groq-hosted Llama) analyzes the combined dataset and writes a structured executive summary with key insights and a recommendation
Report is rendered into a branded HTML dashboard and delivered straight to inboxes on schedule, no manual compilation required
Solution Workflow
Schedule
Fetch
Merge
Compute
Analyze
Deliver
Full Technical Breakdown
1. Scheduled Multi-Source Data Intake
A Schedule Trigger fires at a set interval and fans out to six parallel API calls at once, Google Sheets for sales records, CoinGecko for crypto prices, ExchangeRate API for FX rates, Open-Meteo for local weather, GitHub for repository stats, and Hacker News (Algolia) for top tech stories.
2. Parallel Branch Merging
A Merge node combines all six branches by position into a single item, so every downstream node receives the complete dataset in one pass instead of chaining sequential lookups.
3. KPI Calculation Engine
A JavaScript Code node processes raw Google Sheets rows to calculate total revenue, transaction count, average sale, and highest/lowest sale, then normalizes every external source into one structured JSON object ready for analysis.
4. AI-Generated Executive Analysis
The structured metrics are passed to a Basic LLM Chain running on Groq Chat Model, prompted to act as a senior business intelligence analyst. Output is constrained to a fixed structure, Executive Summary, key insights, and one recommendation, concise, plain-text, decision-ready.
5. Dynamic HTML Report Generation
An Edit Fields node maps computed KPIs and AI output into a fully styled HTML email with inline CSS, KPI cards, a live-data table, and a dedicated AI analysis section, so every value is pulled directly from the workflow run.
6. Automated Email Delivery
The finished HTML report is sent automatically via the Email Distributor node, giving stakeholders a polished, ready-to-read daily brief with zero manual compilation.
Results
Key Features
Multi-source data aggregation: Combines internal business data with live external signals, crypto markets, currency exchange, weather, developer activity, and tech news, into a single unified report.
AI-powered insight generation: Uses an LLM to interpret raw metrics and produce a structured executive summary, key insights, and a clear recommendation, narrative context, not just numbers.
Automated KPI calculation: Custom JavaScript computes total revenue, transaction count, average sale, and highest/lowest sale directly from live spreadsheet data, no manual spreadsheet math.
Scheduled, hands-free reporting: Runs automatically on a recurring schedule with zero manual triggering, the report is generated and delivered without anyone lifting a finger.
Branded HTML dashboard: Renders as a fully designed HTML dashboard with KPI cards, a data table, and a highlighted AI analysis section, not a plain-text email.
Parallel data processing: All six data sources are fetched simultaneously and merged by position, keeping the workflow fast even as more sources are added.
Before
Manually checking spreadsheets, market prices, and news separately each morning, no consistent reporting format, insights dependent on whoever has time to compile them, zero automated narrative or recommendation.
After
One consolidated dashboard delivered automatically every day, sales KPIs calculated instantly, market and industry context included alongside business data, AI-written executive summary with actionable recommendations, consistent branded format every time.
What This Proves
This system demonstrates that AI can turn disparate, disconnected data sources, internal and external, into a single coherent narrative, giving executives a daily briefing that would otherwise require a dedicated analyst to compile by hand.
Tech Stack
AI Layer
Groq (Llama) — Fast-inference LLM for executive summary generation
n8n — Workflow orchestration engine
Data Layer
Google Sheets — Sales and transaction data source
CoinGecko API — Live cryptocurrency prices
ExchangeRate API — USD conversion rates
Open-Meteo — Local weather data
GitHub API — Repository activity metrics
Hacker News (Algolia) — Top tech industry stories
Communication Layer
Email (SMTP) — Automated report delivery
HTML / CSS — Custom-styled dashboard rendering
Scalability & Flexibility
LLM Provider Flexibility: Currently runs on Groq-hosted Llama; can swap to OpenAI GPT-4o, Anthropic Claude, or Gemini for deeper analysis or longer-form reporting.
Data Source Expansion: Built on a parallel-fetch-and-merge pattern, so new sources (stock indices, competitor pricing, social sentiment, internal CRM data) can be added as additional branches without reworking core logic.
Delivery Channel Flexibility: Currently delivered via email; can extend to Slack, Microsoft Teams, or a live web dashboard using the same computed dataset.
Report Frequency Control: Schedule trigger can be adjusted for hourly, daily, or weekly cadence, or triggered on-demand via webhook for ad-hoc executive requests.
Multi-Recipient Distribution: Can expand from a single recipient to a distribution list or role-based routing (e.g., finance gets a KPI-focused cut, ops gets a different one).
Historical Tracking: Report data can be logged to a database or spreadsheet on each run to build a historical trendline for revenue, market conditions, and AI-flagged insights over time.
Custom Alerting: Can add threshold-based logic to trigger urgent notifications when KPIs (e.g., revenue drop, crypto volatility) cross defined limits, rather than waiting for the scheduled report.
Template Versioning: HTML layout, KPI set, and AI prompt structure can be versioned so different stakeholders receive tailored report formats from the same pipeline.