AI Voice Call Screening PoC need Web Development
Contact person: AI Voice Call Screening PoC
Phone:Show
Email:Show
Location: Härnösand, Sweden
Budget: Recommended by industry experts
Time to start: As soon as possible
Project description:
"I want to prove out an all-IP call-handling flow that lets an AI pick up every inbound call, have a short natural-language conversation (in swedish or english), decide what the caller wants, and—if the call is worth my time—bridge it straight through to my mobile as a normal phone call. No handset apps, no special dial-ins, just a single number that “guards” me.
Here is the flow I need assembled:
• A DID (phonnumber) recieves the call
* If the caller-number is whitelisted, connect to my mobile.
• If no whitelist, an NLU-driven dialog (Dialogflow CX, Lex, Azure Bot, or similar—your choice) greets the caller, lets them speak freely, and determines whether the intent is social, seller, health-care or acute.
• Based on that intent, the system either connects 1) the caller to my mobile automatically, 2) connects the caller to another mobile, 3) gives the option to use a pin to be connetcted, or 4) lets the user end the call.
• Every interaction is logged with time-stamps, audio, transcript and the accept/reject decision, visible in a lightweight web admin page where I can also change the destination numbers or tweak category rules.
The result I expect is a working proof of concept that runs end-to-end: I dial the test number, speak to the AI, and—if classified as “social” or “health-care” for example—my personal phone rings without me pressing any extra keys. Calls flagged “seller” should be politely terminated or gicven option to be connected to other person.
Calls from hidden numbers should be met with a polite information that the caller does not want to be contacted from hidden numbers, and information about how to call without hiding the number.
Deliverables
• Deployed PoC reachable from a public swedish test number
• Source code and any configuration or build scripts
• Simple web admin/dashboard plus audit log storage
• Step-by-step setup notes so I can reproduce the environment
Feel free to leverage Twilio Programmable Voice, Amazon Connect, SignalWire, FreeSWITCH, Asterisk—whatever lets you move quickest while keeping everything IP-based. My main requirement is that the dialog relies on true natural language understanding, not just DTMF menus or keyword spotting.
If you have relevant telecom experience and can stand up the demo quickly, let’s talk timelines and milestones." (client-provided description)
Matched companies (4)

JanakiBhuvi Tech Labs Private Limited

Conchakra Technologies Pvt Ltd

Appsdiary Technologies
