Ask. Talk to your data.
Any user, plain language, real answers in seconds. No report building, no BI tool, no waiting on the data team.
We are the OSS/BSS with AI built in, not bolted on.
We are making the platform itself AI-native and open, not sprinkling AI around the edges. With mixed application of “AI-native,” what does it honestly mean?
AI-native software is software where the AI is built into the platform's architecture, not wired in from outside.
The intelligence runs on the same data, identity, and permissions as the rest of the product.
Agents are first-class users, with their own identity and audit trail instead of borrowing a person's.
It's the same documented API we build on, not a thinner one handed to customers.
Any user, plain language, real answers in seconds. No report building, no BI tool, no waiting on the data team.
AI helps you build the workflow, and what it builds runs deterministically by default. Drop in an AI block where you want reasoning rather than rules, and a human approval step where someone should sign off before it continues. You decide where each one goes.
Your own agents reach Sonar data through their own access tokens, with activity logged separately from your people's.
All three are in active development, in varying states from internal alpha to closed beta.
Secure. All of it is built to run through a single governed interface, where every agent carries its own identity instead of borrowing a person's. We run the systems our customers depend on, which is why security couldn't be an afterthought.
We're the system of record for ISPs, and the system of action built on top of it. Plenty of vendors have an API. Fewer hand you the same one they build on. There's no private interface we keep for ourselves and a thinner one we give you. Since 2015, Sonar has been open on purpose, so customers keep control of their data and how they run their business. That head start in being open, not a walled garden, is why we can build AI into the platform instead of bolting it on.
Our API has been fully documented since our first generation. Locking customers in was never the plan.
Embedded BI reporting, then Data Connect, and Ask is what comes next. Simply interact with your data in an interface similar to your favorite LLM.
Human in the loop is the default. You decide how much control to hand over, and when.
Every agent reaches Sonar through its own identity, inside the access you've already granted, never past it, and its activity is logged separately from your people's. When Ask doesn't have a verified answer, it says so instead of guessing. You decide how far automation goes in the workflow builder, and we never train on your data.
Every agent acts through its own agent-level access token, not a person's. Every action is attributable and traceable, separate from the human.
An agent only reaches data the human it works for is already allowed to see.
Verified queries get certified answers. When a question doesn't map to one, Sonar says so plainly instead of making something up.
The workflow builder lets you decide how far automation goes. Trust is extended as it's earned.
Deliberate PII handling, no training on customer data, clear storage location, regional residency.
Straight answers on what Ask, Automate, and Extend are, where they stand, and how they handle your data.
AI-native software is software where the AI is built into the platform's architecture, not wired in from outside. In Sonar, that means the intelligence runs on the same data, identity, and permissions as the rest of the product, agents are first-class users with their own identity and audit trail, and it uses the same documented API Sonar builds on.
Sonar is making its OSS/BSS platform for ISPs AI-native and open through three capabilities: Ask, to talk to your data in plain language; Automate, to build supervised workflows; and Extend, to connect your own agents to Sonar data through their own access tokens.
Not yet generally. All three are in active development, in varying states from internal alpha to closed beta.
Every agent reaches Sonar through its own agent-level access token, not a person's, and only reaches data the human it works for is already allowed to see. Its activity is logged separately from your people's, so every action is attributable and traceable.
No. Sonar never trains on customer data, and applies deliberate PII handling, a clear storage location, and regional residency.
Verified queries get certified answers. When a question doesn't map to one, Sonar says so plainly instead of making something up.
Sonar's AI vision continues. New chapters and assets will publish here, and Ask, Automate, and Extend are still being built. Tell us to keep you posted.
Last updated