Mashdun
AI GrowthWorkCapabilitiesIntegrationsProcessLabsMarketplaceBlogAbout
Get in Touch
Mashdun

Full-stack web developer, AI engineer & growth marketer. Building production-grade apps and intelligent solutions.

Navigation

  • Portfolio
  • Capabilities
  • Process
  • Labs
  • About

Resources

  • Blog / Notes
  • Marketplace
  • Contact
  • RSS Feed

Legal

  • Terms of Service
  • Privacy Policy
  • Cookie Policy
  • Do Not Sell or Share
  • Delete my data

© 2026 Mashdun. All rights reserved.

Built with Next.js, Tailwind CSS & Prisma

    All Labs
    Sales
    First-party CDP
    Enrichment Waterfall
    ICP Scoring
    AI SDR
    Build vs Buy

    AI SDR Pipeline

    An end-to-end acquisition system built on our own code and our own Postgres: a first-party event layer in the Segment Spec shape, an enrichment waterfall that deletes any field whose supporting quote is not in the page we fetched, explainable ICP scoring, rule-based routing, and drafted outbound that cites its facts. Vendor adapters (Apollo, Clay, Segment, RudderStack) plug in behind the same interfaces, switched on by env var — with a build-vs-buy panel pricing both sides from sourced numbers.

    AI SDR Pipeline

    First-party CDP
    Enrichment Waterfall
    ICP Scoring
    AI SDR
    Build vs Buy

    Loading the saved run…

    How it works

    1. Source target accounts from a first-party list; Apollo plugs in behind the same interface when a key exists.
    2. Fetch each company's own pages — homepage, contact, about — through the SSRF-guarded fetcher, recording every request's status.
    3. Read the postal address off the page and extract the other fields; keep only those whose quote is really in the fetched text.
    4. Hold any account with no page read or nothing citable. Score the rest on fit and first-party behaviour; unknown axes are left out, not zeroed.
    5. Tier by score band, thresholds on the page; an override can hold an account back, never promote it. Draft for the top A-tier accounts.

    Why the quotes matter

    An enrichment vendor hands you a field and you cannot ask it why. Here every field names the page and the sentence behind it, and a citation that cannot be found in the fetched text is deleted along with its field. A model asked to cite its sources will invent a plausible quote — checking it against the bytes is the only defence that holds.

    See the Outreach Composer these drafts run through