How canwe help?
Answers about Surogate, Digital Automation and the DenseMAX Appliance, and how to reach the people behind them.
Getting started
What Invergent is and where to begin.
What does Invergent make?
Three things that fit together. Surogate is the platform for agents, the processes they follow and the models behind them. Digital Automation is a no-code platform for complex digital workflows. The DenseMAX Appliance is the hardware they run on, delivered with Surogate installed.
Where should my team start?
Start with the one your teams need first. Most start with one process. We set up a pilot in our cloud or yours, then grow from there.
How do I get a demo?
Fill in the form on the contact page or write to sales@invergent.ai. In a personalised demo our experts show you how our products fit into your workflows, with your data under your control the whole time.
Do I need developers to build agents?
No. Surogate Agents lets your domain experts build, govern and run agents for whole teams without code. Skills are written in plain language and the platform turns them into something an agent can execute and an evaluation can score.
Surogate
Agents, Keel, the AI Gateway and our models.
What is in the Surogate family?
It starts with our own models. Our lab trains Surogate 3.7 Flash, Rune and Surogate Speech, and fine tunes them on your sessions. Surogate Agents builds and runs agents on them. Keel runs agents inside your processes, under rules your team owns. The AI Gateway puts every model behind one API, with keys, budgets and failover.
Which channels can agents work on?
Customer support runs on the web, WhatsApp and the phone. Internal help desks run in Slack and Teams. Agents reach only the tools and data you grant them.
What does Keel add to agents?
Keel runs whole processes end to end, with people where you want them. It keeps a record of every step and decision for your auditors, and you can change a process without disturbing cases already running. Typical uses are loan and mortgage approvals, insurance claims and customer onboarding checks.
Why route models through the AI Gateway?
You see what every team spends on AI and can cap it. Prompts go only to the providers you approve, EU data can stay with EU hosted providers, and work carries on when a provider goes down. It also serves coding assistants such as Codex and opencode.
Which models do you train?
Surogate 3.7 Flash is our language model for agent work, with 1M tokens of context, served from the EU. Surogate Rune is a decision model for text and images, 26B with 4B active. Jackrabbit is speech to text at 116M parameters, and Amami is text to speech that runs on a CPU.
Can a model be trained on our own sessions?
Yes. Sessions become datasets, datasets train models, evaluations score them and the experts go back into new sessions. You own the result, your data never leaves your network, and on your own hardware there is no per token bill.
Digital Automation
The no-code platform and its six modules.
What are the six modules?
Enterprise Portal for digital experiences. Advanced Analytics for answers from your data. Process Automation for processes, documents and rules in one flow. Identity and Access for sign in, MFA and SSO. Enterprise Integration for connecting and exposing services as managed APIs. Development and DevOps for building, deploying and running applications.
Do the modules work on their own?
Yes. Start with the module you need today. Each one runs on its own and shares identity, data and processes with the rest when you add them.
What can Process Automation handle?
Business processes, document lifecycles, decisions and unpredictable situations that need an event driven response. Three engines share one foundation: BPMN for processes, CMMN for cases and DMN for decisions.
Does Identity and Access cover customers and staff?
Yes. It delivers trusted sign in for customers, staff and partners, with multifactor authentication and single sign on from one identity solution.
How does Enterprise Integration connect our systems?
It lets you build, integrate, secure and expose your digital services as managed APIs, so every system is connected through one governed layer.
DenseMAX Appliance
The hardware, the specs and what ships on it.
What is inside a DenseMAX?
Up to eight NVIDIA RTX 5090 or RTX PRO 6000 Blackwell GPUs, two AMD EPYC processors, 24 DDR5 ECC RDIMM slots, a 400 Gbps InfiniBand card and five 2000 W redundant power supplies, in one serviceable, air cooled chassis.
What software ships on it?
Surogate comes installed: Agent Studio, the managed runtime, our own serving engine tuned for Blackwell, and the tools to train experts on your own traces with the GPUs in the chassis.
Does it need liquid cooling?
No. It is air cooled with front intake and optimised airflow. It runs at 10 °C to 35 °C, and GPUs, storage and power supplies are reachable with tool assisted access.
How do I get pricing and a configuration?
Write to sales@invergent.ai or book a demo. The team talks you through configurations, pricing and delivery. The spec sheet is available as a PDF on the appliance page.
Deployment and security
Where it runs and how your data stays yours.
Where can the platform run?
On a DenseMAX appliance on your own floor, in your own cloud account on AWS, GCP, Azure or Oracle Cloud inside your VPC, in a hybrid of both under one control plane, or fully air gapped where nothing leaves. Surogate also runs in our EU cloud.
Does our data leave our network?
Not unless you send it. Your data, your traces and the weights you train stay inside your perimeter, whether that is your VPC or your own metal.
How are agents kept inside their limits?
Every agent is a registered, versioned object with its own credentials and its own budget cap, and inherits nothing from whoever launched it. Redaction, PII handling, content filters, throttling and automated red teaming are applied while it runs.
Which compliance standards apply?
ISO 27001, GDPR and Schrems II, and HIPAA ready architectures. SOC 2 Type II is in progress.
Can we leave and take our work with us?
Yes. The platform is open by construction: Kubernetes underneath, the Model Context Protocol for tools, OpenAI compatible endpoints for serving and an open source training core. Your agents, datasets and trained models are portable.
Company, legal and support
Contracts, policies, partners and people.
How do I reach support?
Write to support@invergent.ai. For anything about plans, pricing or a new project, write to sales@invergent.ai.
Do you have a partner program?
Yes. Partners earn 30% of every client plan, every month, for as long as the client stays. Silver needs no fee, Gold starts with one client live, and Diamond starts with three clients live and pays a 40% share.
Where are your terms and policies?
The Terms of Use, Privacy Policy, Acceptable Use Policy and cookie settings are all linked from the footer of every page.
Are you hiring?
Open roles are listed on the careers page. Each one opens an email to apply.
Still stuck?
Write to us. A person reads every message.

