Skip to content
Page 5 of 7

Challenge 4: Go Live

Recap:

In Challenge 3, your project context file changed everything. AI knew your project from the first word of every conversation, and you used that speed to build ambitious features: integrating new data, adding domain-specific logic, and whatever your team chose to push toward. Your app went from displaying real data to being genuinely useful.

Then in Lesson 4, you zoomed out. You learned what deployment means: moving your app from your private Environment to a live URL that anyone can visit. You saw that your Save & Sync habit has been building toward this moment all along, that the deployment pipeline is already set up, and that going live is one prompt away. You ran the pre-flight check and made a game plan for this final sprint.

You also learned some important truths about AI-built software: the two-week cliff (things break when they interact), the validation gap (AI says it's done before you've verified it works), and why authoritative data sources matter more than AI-generated explanations. Those aren't reasons to hold back. They're reasons to ship thoughtfully.

This is it. Your final sprint. Time to ship.

The Challenge

Deploy your Federal Vendor Intelligence Tool to a live URL. Then polish, fix, and add that final touch. By the end of this challenge, your tool should be live: a real application, accessible to anyone with a browser, built by a team that had never written code before today.

This is the challenge where your workshop project becomes a real product. Deploy first, then iterate.

What to Build

Items are listed in priority order. If time is tight, focus on the items near the top first.

  • Your Federal Vendor Intelligence Tool is deployed to a live URL: tell your AI coding assistant to deploy the project and get back a working URL that anyone can visit from any device
  • Fix any issues from the pre-flight check: if something was broken or outdated when you checked in Lesson 4, fix it now
  • Vendor comparison: a side-by-side view of two or more vendors, comparing award amounts, top agencies, NAICS codes, contract vehicles, and business types. If your team already built a two-vendor comparison, expand it: Can a user compare three or four vendors at once? Can they sort or filter the comparison by a specific dimension?
  • Acceptance criteria audit: go back to your acceptance criteria from Challenges 1, 2, and 3. Verify every feature works correctly on the live URL. Fix anything that does not fully pass. Your live tool should meet every baseline capability you have built so far, not just the ones from this challenge.
  • Your live version reflects your latest work: save and sync so the live URL shows everything you have built, not an earlier version

These are options for teams that finish the baseline capabilities.

  • Interactive AI advisor: add an "Ask about this vendor" input where a user types a natural language question and gets an AI-generated answer grounded in the vendor's actual data. Think about the questions you would ask during vendor research: "Is this vendor a good candidate for a small business set-aside under NAICS 541330?" or "What agencies has this vendor never worked with?" or "Compare this vendor's risk profile to Boeing's." The data is in the tool. Let users query it conversationally.
  • Executive one-pager: a print-formatted, AI-generated single-page summary designed for a decision-maker who has 60 seconds. One paragraph recommendation, go/no-go, key numbers, top risk. If your team built a due diligence briefing in Challenge 3, this is the complement: that briefing is comprehensive, this is decisive.
  • Data quality audit: AI scans the vendor's records across all sources and flags gaps, inconsistencies, or staleness. Expired SAM registration, NAICS codes that appear in award history but not in the SAM profile, missing exclusion records, business type mismatches. Surface a confidence indicator for each data point so the user knows what to trust and what to investigate further.
  • Revisit earlier stretch goals: look back at the stretch goals from Challenges 1, 2, and 3. Are there features your team skipped or started but did not finish? This is your chance to go back and build them. The AI-powered intelligence features from Challenge 3 (vendor risk score, due diligence briefing, competitive landscape intelligence, trend analysis with narrative) are some of the most ambitious and rewarding options if your team has not tackled them yet. Contract vehicle breakdowns, subsidiary information, spending trends, and active opportunities are also worth revisiting.
  • Mobile-friendly: test your live URL on a phone and make sure it looks good on a small screen
  • About page: explain what the tool is, where the data comes from, what questions it helps answer, how to interpret any AI-generated assessments, and who built it

Final Growth Check-in

There is a final growth check-in on the next page. Make sure to navigate there and fill it out before the final reflection begins.

Tips

  • Deploy first, improve second. Do not try to perfect everything before you deploy. Get the URL, confirm it works, then use the remaining time to polish. You can redeploy as many times as you want. Save and sync your work, and the pipeline updates the live version automatically.
  • Use the game plan you made in Lesson 4. Your team already discussed what to tackle first. Stick to the plan, or adjust it now that you are in the sprint.
  • Test the live URL, not just the Live Preview. Your Live Preview and the deployed version should look the same, but always check the real URL. Open it in a new browser window or on a teammate's phone. Check the things that matter most: Does the vendor profile show real data? Do the pages and navigation work? Does the exclusion check display correctly? These are your acceptance criteria from earlier challenges; they should still pass on the live version.
  • If something goes wrong, do not panic. If deployment fails, ask your AI assistant: "The deployment didn't work. Can you investigate why and fix it?" Your AI assistant can see what the pipeline flagged and make the necessary changes. If the live URL looks different from your Live Preview, save and sync again to trigger a fresh pipeline run. If a feature that worked in your Live Preview is broken on the live URL, save and sync to make sure your latest code is pushed.
  • Save and sync after each change. This is the same habit you have been building. Make a change, verify it in the Live Preview, then save and sync. The pipeline checks your code and updates the live version automatically. If the pipeline flags a problem, your AI assistant can investigate and fix it.

Go build. That is the brief. Spend the rest of this session block working on your challenge with your team. Your Facilitator will let you know when it is time for the Reflection.