AI agents

AI agents in daily operations

Every new piece of work triggered the same chain of tasks: write it up, send it out, prepare the posts, chase the reply. Done by hand, that chain set the ceiling on how much could ship.

  • Writing agent
  • Distribution
  • Social posts
  • Follow-ups

Private system, demo on request

A generated article as published on a local TV news site (Fort Wayne's NBC MarketMinute) through the automatic distribution: outlet header, title and opening paragraphs

One input in; write-up, distribution, social posts and follow-ups out. No one touches it in between.

What's inside

The parts that do the work.

A post on the Xraised LinkedIn page with the generated social slides

01

Social posts, generated and published

Slides, captions and short videos come out of the same source and go out on their channel on schedule.

The follow-ups page with drafts to approve (fictional demo data)

02

Follow-ups that wait for a human

In the CRM, an agent notices when the client has not written back, drafts the next nudge and queues it for approval. Nothing is sent on its own.

An interview page on xraised.com with the player, category and title

03

One input, everything else derived

A single source, here a recorded conversation, is the only input; the write-up, the distribution and the posts are derived from it.

Built with

Python · FastAPI · PostgreSQL · APScheduler · Playwright · ffmpeg · Next.js · Railway

Integrations: Anthropic Claude, ElevenLabs, Newswire API, Buffer, Microsoft 365.

Need custom systems and crm like this?

Software built around how your business actually works — not the other way round.