Skip to content
Build · AI automation

Your Team's First AI Automation

For teams with a process that takes hours every week. We build the automation with the team and hand it over with tests, human approval, and people trained to maintain it.

Duration
4 to 6 weeks
Audience
Operations, content, support, and back office
Outcome
A working automation and a team that can maintain it

U.S. publisher · NDA

7 workflows

in the editorial workflow, with human approval between steps

Pulso

88 editions

through Oct 7, 2026, with every link opened and the numbers checked before going live

Lapidador

23 techniques

independent product in production, with built-in payments

Who it's for

  • A repetitive, well-known process: screening, reports, standard replies, content.
  • The team already uses AI and wants to move from individual use to a team workflow.
  • A stalled pilot needs something working, with clear limits.

How it works

  1. Week 1

    Map

    Today's process and what can be removed before automating.

  2. Week 2

    Prototype and decision

    A prototype on real data. The company decides to continue, adjust, or stop.

  3. Weeks 3 to 5

    Build with the team

    Human approval, test cases, and checks on numbers and sources.

  4. Final week

    Hand over

    Training for the operators, documentation, and a record of cost and failures.

Decision point in week 2

If the prototype does not show value on real data, the project stops there and the company pays only for that phase.

Either way, the company keeps

  • The process map, showing what is worth automating and what is worth removing
  • The prototype and a record of what worked

What you get

  • Process map
  • Automation with human approval
  • Quality test cases
  • Trained team and documentation
  • Code, accounts, and keys in the company's name

In production, not just a demo

U.S. publisher (under NDA)

Seven AI workflows in the editorial workflow of a U.S. fiction publisher, from the first manuscript screening to marketing materials.

Orchestrated in n8n and Monday, with human approval between steps: AI speeds up the work, and the decision stays with people.

Company name and internal details omitted under a confidentiality agreement.

Related case studies

Frequently asked questions

Have these automations run in production?

Yes. For a U.S. publisher, under NDA, seven workflows were deployed with human approval between steps.

Why build together with the team?

An automation only the vendor understands tends to be abandoned. Gartner predicts that by 2028, 70% of enterprises will abandon AI agents built by vendor engineers.

What is out of scope?

Systems without a documented API, data migration, and ongoing operation. Maintenance can be added as monthly support.

Contact

To talk about this service, write to oi@dorly.com.br.

Other services