EnterDone builds the agent that understands your company

The thesis: context is not promised, it is built

The whole market has discovered the word context — and rightly so: an AI model without the company's context is a brilliant intern on their first day, forever. The difference is that EnterDone does not treat context as a prompting trick. It treats it as product:

The agent learns from the house's conversations, files, brands, clients and systems — and every fact it keeps carries its origin: who said it, when, where. When it answers, it answers with sources. When it does not know, it says so, and says where it looked. When two sources disagree, it points at the disagreement instead of picking one in the dark. Memory without provenance is gossip; EnterDone's is a record.

And context respects access: the same question, asked by people in different roles, is answered within each one's reach. The whole company's context, delivered through each person's own door.

The second leg: AI proposes, code decides

Giving an agent context and power is only acceptable with deterministic governance. At EnterDone, approval, spending, access and delivery are decided by auditable code — signed authority, a chained audit trail that cannot be rewritten, an emergency brake, a credential vault the model never sees. It is all described, piece by piece, on the governance and security page.

What exists — today, in operation

EnterDone runs every day at a real company in Brazil, on an isolated instance (one server per company, in São Paulo). The agent talks with the team in the company chat, reads and operates systems through a cloud computer — one per person, which calls the person only when it needs them —, uses the connectors management has allowed, runs routines agreed in plain language and delivers with receipts. New companies join by invitation.

Who builds it

Suleiman Ribeiro, founder and data protection officer. An advertising professional, film director and producer, he has run Cacilda Filmes, in São Paulo, since 2017 — producing for brands such as Grupo Boticário, Consórcio Magalu, Lojas Marabraz, XP Inc. and Escola Conquer and, through it, for Coca-Cola FEMSA; today he works mostly in corporate education. Since 2023 he has studied language models, intelligent systems and software engineering in depth. He is not an engineer by training: he built EnterDone as the operator of its first customer company, with a public method — nothing ships without proof, and every decision is on record. Criativo.io is the umbrella for the products he develops; EnterDone is the main one.

EnterDone was built the way it works inside your company: by agents, within rules written in code, with a receipt for every step. What backs that is not one person — it is what can be checked: daily operation at a real company, test batteries with veto power before every release, the signed trail, the declared agent on the web and the vendor list, name by name. The engineering team that reviews and maintains the product is introduced here as each person authorizes their own name.

How to reach us

[email protected]. EnterDone's agent is also a declared citizen of the web — identity, public keys and browsing policy are on the agent's public page.