AI has to get into the systems.
AI agents now read and write in company systems. That only works on a base prepared to receive them.
The company's system of record, with AI built in.
A vertical ERP with tax built in and the governance large companies require.
- Pre-revenue
- In homologation
- Pre-seed open
Three shifts are arriving together.
The buyer's question changed.
In our reading, the large-company buyer no longer asks whether the system has AI. They ask whether the AI respects permissions.
Dense tax rules became a barrier.
Whoever operates well in a country with dense tax rules takes time to catch up with.
Aether was designed inside these shifts. Brazil is the first market because it brings all three together at high density.
A system of record with AI and governance built in.
Aether was designed to pass a large company's vendor assessment: permissions, approvals and change history are already part of the platform.
The honest comparison is with the new generation of vertical ERPs. The difference is durability: the more of the customer's operation lives on the platform, the more expensive it is to leave.
In one sentence
A vertical System of Record with embedded AI and governance ready for vendor assessment.
Unlike
- a copilot on top of an old ERP
- a chat layer pasted over company data
- an integration stitched on afterward
One platform and two agents.
The platform is where the operation lives. One agent builds the system and the other operates it with the user.
Platform
Generates screens from a description of the business, shows the same information in several ways, serves several companies of one group, runs automations without programming, exports reports to PDF and offers a REST API.
Nous, the operations agent
Chat in Portuguese that queries data, generates reports and takes the user to the right screen, inside the workflow. It only shows what the user may see and only runs an action with their confirmation.
Forge, the build agent
Builds the vertical modules on top of the platform. It is the Aether team's tool.
The defense is in the combination.
Four layers that reinforce each other. Copying any one of them is already hard work.
Tax regulation protects
Brazilian tax obligations are in the core, and NF-e already carries the Tax Reform calculation. Replicating that takes calendar time with regulators, and that time cannot be bought.
Governance from the first module
Each company sees only its own data, the person who requests is not the person who approves, approval rigor follows risk and the change history cannot be erased. Every new module is born with data isolation.
Two agents inside the platform
Nous operates and Forge builds, both with governed access to data and business rules, with no external integration on top.
It grows by modules
Each tax obligation is a separate module, and language and time zone are already part of the core.
The real state.
Pre-revenue. The product is in homologation and being rolled out in a real operation, and customer entry is by invitation, starting from a diagnostic.
The next six months are about proving value: documenting measurable operational gains and completing the initial portfolio of three slots.
Who leads.
Uemerson Santana Founder and architect
Architect of the platform since day one, with AI to multiply the productivity of the build.
The round's terms come out in conversation.
Pre-seed round open, on a SAFE.
The raise amount, cap, discount, conversion terms and use of capital are shared with qualified investors after first contact. The page does not publish them.
In due diligence, transparency is total. Until the round closes, the parameters stay off the page.
Risks we already know.
We discuss each one in detail in due diligence.
Long sales cycle
It comes with the category. Entry starts with one module of the operation, which lowers the initial friction.
A single market
Brazil is the first market by choice of the initial customer. Language and time zone are already part of the core, but other jurisdictions are not yet in operation.
Dependence on AI providers
The platform accepts more than one AI provider, but the dependence exists.
A conversation with the founder
Thesis materials and round parameters follow by email after first contact. The founder replies within 24h.