twenty80.studio
Book a call

Case study · Product

30,000 posts a month.
Four clicks.

A SaaS platform in a regulated industry scheduled its customers’ marketing by hand, and was about to license an expensive, fully agentic tool for it. We built it into their own product instead, with automation, rules and a little AI.

Before

  • Every post copied and pasted.
  • Everyone got the same content.
  • No time to tailor or curate anything.

The other option

They were about to buy an agent for everything.

  1. The fully agentic tool

    • $200K licence, plus usage fees
    • AI on every single booking
    • Every booking, a new chance for error

  2. What we built

    • Automation and rules do the routine work
    • AI only for writing rules and curating
    • Every booking follows the same checked rules

We audited it, then split it three ways.

  1. Mechanical work

    Automation

    Scheduling became a pipeline. A rerun never double-books, and every post leaves a trail.

  2. Policy

    Rules

    Staff write rules in plain words and an agent turns them into exact ones. Applying them needs no AI: same input, same schedule.

  3. Judgement

    AI

    One agent per customer picks and tailors content from approved material. A new premium tier to sell.

What changed

9
full-time people’s work taken off the team: 18,000 hours a year
250×
cheaper to run than the tool they nearly bought
+1
product line: a personal programme for every customer
  • Every post scheduled by rule and checked against approvals: no manual entry, no double bookings, nothing sent unapproved.
  • No AI on the routine path, so fewer ways to fail.

Stop piloting.Start compounding.

It starts with the map: where AI pays in your company, what it costs, and the first build worth doing.

Or write to [email address]