Skip to content
Qzentra

Case fileAI Systems & Automation

AI Website Audit & Outreach Engine

An automated pipeline from a list of target URLs to personalized, segmented outreach: crawling, contact extraction, PageSpeed analysis, LLM-written messages, and automated follow-ups.

Enters through
01AI Systems & Automation
System type
Web data pipeline with AI-generated outbound
Also demonstrates
web data extraction · API integration · LLM generation · segmentation · outbound automation02Data Engineering
Verified fact
Every message grounded in a Google PageSpeed audit of the prospect’s site
System route: target URLs to automated follow-ups.Seven stages rising from the data layer through APIs and AI to automation. A pipeline: nothing returns to an earlier layer.06  Application05  Automation04  AI03  APIs02  Infrastructure01  Data1234567
  1. 1Target URLsData
  2. 2Crawling and contact/social extractionData
  3. 3Google PageSpeed APIAPIs
  4. 4Technical analysisAI
  5. 5LLM-generated personalized outreachAI
  6. 6SegmentationAutomation
  7. 7Automated follow-upsAutomation

01 Context

The challenge

The requirement was outreach that referred to a real, current issue on each prospect’s own website, generated automatically from a list of targets, segmented, and followed up without manual tracking.

That meant treating each website as a data source: crawl it, extract who to contact, measure it objectively, and only then write.

What made it non-trivial

  • Websites are unstructured; contact and social details had to be extracted reliably
  • Technical analysis had to be objective and repeatable, not an opinion
  • Generated messages had to be grounded in the audit rather than generic
  • Segmentation and follow-up had to run without a person tracking each thread

02 The system

What Qzentra engineered

Qzentra built a pipeline that starts from target URLs. Each site is crawled; contact and social details are extracted; the Google PageSpeed API measures the site; and the results are analysed into technical findings.

From those findings, an LLM layer using OpenAI and Gemini generates outreach personalized to each prospect. Prospects are segmented, and follow-ups run automatically.

What Qzentra was responsible for

  1. Crawling and contact/social extraction
  2. Google PageSpeed API integration and technical analysis
  3. LLM-generated personalized outreach with OpenAI and Gemini
  4. Segmentation and outbound campaign logic
  5. Automated follow-ups

03 The route

How the system works

Read left to right. Every stage consumes the output of the one before it, which is what makes this a pipeline rather than a workflow.

  1. Target URLs

    Data

    The pipeline starts from a list of target websites.

  2. Crawling and contact/social extraction

    Data

    Each site is crawled. Contact details and social profiles are extracted as structured data.

  3. Google PageSpeed API

    APIs

    The site is measured through the Google PageSpeed API, giving repeatable technical metrics.

  4. Technical analysis

    AI

    The PageSpeed results are analysed into findings that can be written about.

  5. LLM-generated personalized outreach

    AI

    OpenAI and Gemini generate outreach for each prospect from the findings for their own site.

  6. Segmentation

    Automation

    Prospects are segmented, which decides the outreach sequence they enter.

  7. Automated follow-ups

    Automation

    Follow-ups run on a schedule without anyone tracking individual threads.

04 Decisions

Engineering decisions

  1. Audit before writing

    Each message is generated from the PageSpeed findings for that site, so personalization is factual rather than a template with a name inserted.

  2. Objective measurement through an API

    Google PageSpeed provides repeatable technical metrics; the analysis step interprets them. The pipeline does not rely on a model to assess a site.

  3. Extraction as a data step

    Contact and social details are extracted from the crawl as structured data before any outreach logic runs.

  4. Segment, then send

    Segmentation decides which outreach sequence a prospect enters, and follow-ups are timed automatically from there.

05 Verified outcomes

What was achieved

  • Technical website analysis is produced automatically for every target URL
  • Contact and social details are extracted as structured data
  • Outreach is personalized to each site’s audit findings
  • Segmentation and follow-ups run without manual tracking

06 Technology and capabilities

Technology, after the architecture

Named last, with the role each played. The capabilities are what the project demonstrates beyond its primary service.

  • Web crawling and extractionContact and social data from each site
  • Google PageSpeed APIRepeatable technical site analysis
  • OpenAIOutreach generation
  • GeminiOutreach generation
  • Workflow automationSegmentation and timed follow-ups

Capabilities demonstrated

  • web data extraction
  • API integration
  • workflow automation
  • dynamic content generation
  • segmentation
  • outbound campaign logic
  • follow-up automation

07 Related work

Other case files

Systems that share a layer, a capability, or a decision with this one.

  1. 01Case fileAI Systems & Automation

    Multimodal AI Sales Funnel

    An end-to-end sales funnel for a crypto-mining hardware enterprise: Instagram and WhatsApp leads qualified, answered from a catalog of 500+ ASIC miners, scored, and handed to sales.

    Verified factAirtable-backed catalog of 500+ ASIC miners

    Enters through01AI Systems & Automation

    Read the case file
    Multimodal AI Sales Funnel: route through the six layers.06  Application05  Automation04  AI03  APIs02  Infrastructure01  Data12345678
    1. 1Instagram DMsAPIs
    2. 2Lead qualificationAutomation
    3. 3WhatsAppAPIs
    4. 4Multimodal AI (text, voice notes, images)AI
    5. 5Product knowledge retrievalData
    6. 6Lead scoringAutomation
    7. 7Human sales handoffApplication
    8. 8Automated follow-up or cold-lead archivingAutomation

  2. 04Case fileData Engineering

    Enterprise Data Query Architecture

    Query architectures for analytical workloads across data platforms and data lakes, using distributed and embedded engines over Azure Data Lake storage.

    Enters through02Data Engineering

    Read the case file
    Enterprise Data Query Architecture: route through the six layers.06  Application05  Automation04  AI03  APIs02  Infrastructure01  Data1234
    1. 1lake storageData
    2. 2query enginesInfrastructure
    3. 3federationAPIs
    4. 4analyticsApplication

08 Discuss a project

Have data on the web that should be driving a workflow?

Describe the sources, what has to be extracted or measured, and what should happen with the result. Pipelines like this are scoped from the data outward.

Discuss a ProjectAll work