Aurora Assistant — Run banking, travel, calls and the smart home from one voice or chat assistant.
An autonomous AI assistant for home and business. Banking, travel, smart home, phone calls and email are handled by voice or chat, in one assistant instead of a dozen apps.
The problem
Turning a spoken request into a completed task across very different services, quickly enough to feel like a conversation, while keeping account actions such as transfers safe and confirmed.
What we built
A real-time voice and chat front end over an agent that routes each request to the right capability (banking, travel search and booking, smart-home control, automated calls and email), confirming actions before it carries them out.
What it does today
Ask by voice or chat and the task gets done — a flight found and booked, a balance checked, lights switched off — with a confirmation before anything that moves money or makes a booking.
— Under the bonnet
What makes a product like this hard to build
- 01
Real-time voice
A conversation only feels natural if replies start within about a second. Speech recognition, the model and speech synthesis all have to stream, and interruptions have to work.
- 02
Acting, not just answering
Booking a flight or switching off lights means calling third-party services with the right details. The agent has to choose the right tool, fill it correctly and recover when a service fails.
- 03
Confirmation before anything that matters
Anything that moves money or makes a booking needs an explicit confirmation step, so the assistant never acts on a misheard word.
- 04
Many integrations, one assistant
Banking, travel, smart home, calls and email each have their own APIs, authentication and edge cases, and they all have to feel like one product.
— Estimated build cost
What a build like this costs
£330,000–£480,000
Voice and chat AI assistant with integrations · indicative range for a comparable build
Sized from the codebase we built: 428k lines of production code, 21k lines of tests, 348 database tables, 2395 API routes.
A product like Aurora Assistant is a bespoke build, not a template: most of the cost is in the hard parts above, plus testing against real data before anyone relies on it. Where your version lands in the range depends on the factors listed here. You get a fixed price once scope is agreed — usually after a short discovery sprint. Hosting and AI model usage are billed at cost on your own accounts.
Get an estimate for yoursWhat moves the price
- Number and type of integrations (each has its own API and sign-in)
- Voice latency targets
- Security around payments and account access
- Phone calling as well as web and app
- Ongoing model and voice costs per user
Built with
- React
- TypeScript
- Node.js
- Express
- OpenAI
- Anthropic Claude
- Playwright
- Python
Designed, built and run by our Manchester team. Want something similar? See our AI development services, AI automation and pricing.
