A system that understands your business and runs it toward outcomes

HelloTwin builds a living Semantic Digital Twin of your business and runs Digital Authorities (“Twins”) on top of it, one for each function, accountable for your long-term goals. Two layers, one Semantic Operating System. No data team, no guesswork.

Layer 1

A living Semantic Digital Twin of your business

HelloTwin connects to your apps to build a living Semantic Digital Twin: a governed model of your business objects, their relationships and their history. Start with a prebuilt model for B2B software and AI businesses. Confirm and extend your company-specific definitions in minutes, without a data team.

Semantic Digital Twin

The Semantic Digital Twin is a four-layer architecture that turns raw application data into governed business meaning:

Data Layer: Connects to your existing apps (CRM, Billing, Ticketing, HR, and more) and structures the data for reliable, flexible usage.

Semantic Layer: Adds business meaning independent of data structure. The terms and metrics are still expressed the way each application expresses them.

Ontology Layer: Maps app-centric semantics into company-wide entities. Resolves which records across applications are the same customer, contract or account, and holds your terms, metrics, rules and goals.

State Layer: Captures the changing state of your business. Activities, changes and events are continuously updated and kept as history, so you can see what a number was as well as what it is.

Together these layers give your Twins identity, history, terms, metrics, rules and goals: one governed meaning to work from, not raw data.

Layer 2

Digital Authorities turn understanding into decisions you can defend

Each Twin is a Digital Authority accountable for a business function, such as marketing, sales or product. Twins deliberate, evaluate and prepare the decision, grounded in governed meaning, not guesswork.

Semantic Operating System

You set the goals. Digital Authorities track them, detect material drift, explain its causes and prepare evidence and proposed responses. People remain accountable for decisions today. Authority can expand over time, from governed answers to recommendations, coordination and execution within explicit mandates.

Goals: You define what matters: “increase pipeline conversion by 30% in Q3.” Goals are framed over governed metrics. No ambiguity, no drift.

Deliberation: The goal-assigned Twin nominates other functional Twins as specialized workers (Researcher, Critic, Analyst) to investigate the problem from their functional view. Contradictions are preserved, not hidden.

Evaluation: Evidence is synthesized into option comparisons and risk assessments. The goal-assigned Twin reviews the full picture, grounded in the continuously updated situation held in the Semantic Digital Twin.

Decision: The goal-assigned Twin puts the decision in front of the person accountable for it: what it recommends, why, and what it ruled out. They accept, modify or reject, and a full decision trace records the evidence, the reasoning and the outcome.

Execution within mandates: As authority expands, approved work can be delegated to built-in or third-party AI agents through a governed execution layer. Agents run in sandboxes with scoped access, inside the mandate they were given, and never bypass governance.

Governed calculations are deterministic and traceable. Recommendations and decisions preserve their evidence and full decision trace.

Live software of both layers

See the product in action

Watch both layers of the Semantic Operating System in action: the living Semantic Digital Twin and the Digital Authorities running on it.

Digital Authority
Digital Authority – Twin
Watch the recording →
Semantic Digital Twin
The living Semantic Digital Twin
Watch the recording →
Common questions

Frequently asked questions

How HelloTwin fits into your existing stack, governs business meaning and keeps people in control.

How quickly can we get started?

Start with one function and one important outcome. Choose from 750 preconfigured connectors, connect the applications you already use through read-only access, and start working with your data immediately.

HelloTwin applies a prebuilt operating model for B2B software and AI businesses, so you do not have to wait for a warehouse, data team or lengthy modelling project before seeing value. Confirm and extend your company-specific terms, metrics, rules and goals as you go.

How is this different from connecting AI through MCP?

MCP gives AI access to application records. It does not create shared identity, preserve historical state or consistently apply your company-specific terms, metrics, rules and goals.

HelloTwin provides that governed business meaning so AI does not have to infer what the records mean.

Does HelloTwin replace or write to our existing systems?

Your CRM, billing, accounting, product and other operational systems remain authoritative. The private beta starts read-only: Twins track goals, detect risks and opportunities, explain what changed and prepare the evidence for decisions.

Over time, you can expand their authority from recommendations to coordination and execution within explicit mandates.

How do we define and approve the business meaning?

Tell HelloTwin what you want to measure, govern or achieve in natural language. HelloTwin suggests the required terms, metrics, rules or goals, including definitions, formulas and the relevant data mappings.

An authorised owner reviews, edits and approves each suggestion before it becomes part of the governed business meaning. You control who can view, create, change or approve definitions by setting and managing access rights.

Every approved change is versioned and applied consistently, so you can see who approved what and reproduce previous answers.

How can we verify an answer, and what if the source data is wrong?

Every answer shows the relevant definition and version, calculation, source evidence and point in time attached. You can see which records were included or excluded and which governed meaning was applied.

HelloTwin does not turn incorrect source data into correct data. It makes missing, conflicting or low-quality source data visible instead of allowing AI to silently infer around it.

How is our data protected?

Application connections are read-only, each customer’s data is isolated and EU storage is the default.

See Legal & Privacy for current information about GDPR, subprocessors and security controls.

Why not connect AI directly to our data lakehouse through MCP?

You can. MCP allows AI to query whatever has already been modelled in your data lakehouse. But it does not create or continuously maintain the operational context AI also needs: shared identity, relationships, historical state, company-specific terms, metrics, segments, rules, goals and the point in time that applies.

All of this can be built in a data lakehouse. But in a fast-changing organisation, every new product, process, segment, policy or definition becomes another data-modelling project and an ongoing maintenance obligation.

HelloTwin itself runs on Snowflake. It uses that scalable data foundation to provide a prebuilt operating model and make operational context governable and adaptable, without requiring every company to build and maintain the entire model from scratch.

Can’t AI simply read our rules from Confluence or Notion?

It can read them, but reading a document is not the same as enforcing it. AI must still find the correct page and version, interpret the prose, map it to the right fields and records, and translate it into the filters, mappings and calculations applied to the actual records.

HelloTwin turns approved meaning into versioned, executable logic that is applied and traced in every answer. Confluence or Notion can remain the documentation layer.

The Private Beta is open.

Start managing outcomes, not agents. Be among the first to run a living Semantic Digital Twin of your business: seven Digital Authorities, definition-verified and traceable answers, and connections to all your apps. No data team required.

Apply for the private beta →