Inspiration
I worked for over 10 years as a field technician on heavy construction machines in Italy. Every single day I experienced the same frustration:
- A machine breaks down → I call the repair office from the nearest fixed phone
- End of shift → I call again to report production
- When I'm sick → I call the secretary
90% of the time that information was never tracked. Parts arrived late, production numbers were wrong, absences were not properly justified.
The problem was never the workers it was the tool.
There is, however, one device every worker has used for decades and knows perfectly: the telephone.
That's when the idea hit me: stop forcing workers to learn Jira teach Jira to understand workers.
What it does
Pronto AI Jira turns any ordinary phone call into perfectly structured Jira Service Management tickets in seconds.
The worker simply dials the number configured in JSM, speaks naturally, and hangs up.
Example call:
"Pronto, hydraulic leak on the telehandler, job site Milan, urgent. I'm sick tomorrow."
→ The app automatically creates the correct ticket (maintenance, spare parts, absence, production report, etc.) with the right project, request type, summary, and description.
3–5 seconds after hanging up, the worker receives a confirmation SMS:
✅ Your request has been received!
• Absence request HR-884 created
We'll take care of it. Thank you!
No more "Did they get it?" stress the worker instantly knows everything is tracked.
How we built it
100% on the Atlassian platform:
| Technology | Purpose |
|---|---|
| Atlassian Forge | Fully serverless runtime |
| Forge UI Kit | Native admin interface |
| jira:issueActivity | for call metadata |
| Forge Webtrigger | webhook |
| Forge Queue | long processing |
| Forge Storage (KVS) | |
| Jira Service Management REST API | Ticket creation |
| Twilio | Telephony + instant SMS confirmation |
| OpenAI Whisper + GPT 4o | Will be replaced later with Forge LLMs API |
Everything runs inside Atlassian’s cloud. Zero servers to manage today and fully ready to switch to Forge LLMs when the API is released.
Challenges we ran into
- The 25-second synchronous limit → Solved with Forge Queue for async processing -Twilio webhooks → Solved by using a Forge Webtrigger
Accomplishments that we're proud of
- Built a fully serverless AI-powered telephony system entirely on Atlassian Forge
- AI-powered request type selection
- Demonstrated how AI can make Jira dramatically more accessible to non-technical teams
- Long processing
- Call recording attachment Audio file automatically attached to the ticket for reference
- Instant feedback loop SMS confirmation closes the anxiety gap for workers
What we learned
Workers don't resist digital transformation they resist tools that ignore how they actually work.
Pronto AI Jira is the bridge that finally brings Jira to the 95% of industrial workers who have never used it — and gives them immediate peace of mind with a simple confirmation message.
What's next for Pronto AI Jira
- Outbound calls — Enable the system to call workers back with ticket updates using text-to-speech
- Rovo Agent integration — Connect with Rovo for smarter routing, triage, and automated responses
- Telephony provider abstraction — Support additional providers like Telnyx and Vonage alongside Twilio
- Analytics dashboard — Visualize call volume, common request types, processing time, and ticket outcomes
- Improved admin experience — More intuitive configuration options and clearer UI for non-technical managers
- Expanded field support — Dynamically capture additional structured fields during call processing based on user-defined configuration
- Offline SMS fallback — Allow ticket creation via SMS when calling is not possible
- Caller ID intelligence — Automatically identify the worker and pre-fill requester information
One phone call → perfect ticket → SMS confirmation.
No computer required.
Log in or sign up for Devpost to join the conversation.