Inspiration

Every company I've observed drowns in data — spreadsheets, support tickets, Slack messages, CRM records — yet the people who need insights the most (managers, ops teams, non-technical staff) can't access any of it without waiting days for a data analyst. That waiting game kills decisions.

I asked myself: what if anyone could just ask their company data a question, like texting a colleague who knows everything?

That's OneAsk.

What it does

OneAsk is an internal enterprise knowledge hub where employees type questions in plain natural language and get instant, structured answers pulled from multiple company data sources — with auto-generated charts and source attribution.

Example:

"What were our top 3 customer complaints last month?"

OneAsk scans connected sources (spreadsheets, support tickets, Slack logs), synthesizes the answer, shows a visualization, and tells you exactly where the data came from — in under 10 seconds.

No SQL. No data team. No waiting.

How we built it

Built entirely using MeDo — I described the entire application in natural language across multiple iterations. I defined the UI structure, dummy data schema, interaction flows, and design system through conversational prompts. MeDo handled all the code generation, component building, and deployment.

The iteration process was: describe → generate → test → refine. No manual coding involved at any point.

Challenges we ran into

The biggest challenge was making the app feel real despite using simulated data. I had to carefully craft the dummy data and interaction flows so judges could immediately see the real-world value without needing an actual database connection.

Also, getting the dark-mode enterprise aesthetic right took several MeDo iteration rounds — enterprise apps need to feel trustworthy, not playful.

Accomplishments that we're proud of

  • Built a fully functional, demo-ready enterprise SaaS prototype with zero manual coding
  • Created a product that solves a genuine, universal business problem
  • Designed an experience where any judge can type a question and understand the value in under 10 seconds

What we learned

AI-powered no-code tools like MeDo fundamentally change what "building" means. The bottleneck is no longer coding — it's thinking clearly about the problem. The better you understand the problem, the better the output MeDo generates.

What's next for OneAsk

  • Real data source integrations (Google Sheets API, Zendesk, Slack API)
  • Role-based access control (executives vs. ops team see different data)
  • Scheduled insight digests sent to email every morning
  • Natural language alerts: "Notify me when support tickets exceed 200/month"

Built With

  • cssdarkmode
  • javascript
  • medo
  • natural-language-ui
  • react
  • recharts
  • simulated-rest-api
Share this project:

Updates

Submission history