Internal Survey · March 2026

AI at Naura

How our team interacts with AI tools day-to-day — 25 responses, honest answers, and a few laughs.

Powered by NOVA — Naura's Observational Voice of Analytics. Here to keep it honest (and occasionally roast your responses).

25

Responses

96%

Use AI daily

96%

Use AI for coding

48%

Save 20%+ time

⏱️

How much time with AI?

Hours per day the team spends interacting with AI tools.

NOVA: 80% of the team is in the 1-3 hour sweet spot — not casual dabbling, but a genuine daily workflow. The one person who doesn't interact directly? They generate AI images for UI mockups. Technically in the ecosystem. 🎭

🤖

The Provider Landscape

Which providers does the team use? Multi-select — 50 total selections.

NOVA: OpenAI leads with 20 mentions, but Gemini (12) and Anthropic (11) are close behind. Someone's running Ollama locally 🏆. No loyalty, just results.

⚡

Where it shines

Starting points for tasks + where AI adds the most value.

Where Do You Start?

AI's Sweet Spot

🐛

Debugging

11

🔄

Refactoring

11

🔍

Code Review

10

🧪

Unit Tests

10

🏗️

From Scratch

9

📄

Docs

2

NOVA: Debugging and Refactoring tied at 11 — AI is the team's rubber duck that talks back. Official docs still win as the starting point (36%), which shows healthy skepticism. But docs as a task AI helps with? Only 2 votes. Y'all are sleeping on the one thing AI would happily do for free while you drink coffee. ☕

✅

Adoption Patterns

Beyond coding — how deep does the AI rabbit hole go?

PR Reviews

Documentation

Brainstorming

NOVA: 20% for PR reviews (early but promising), 64% generating docs (biggest win outside code), and 56% brainstorming — AI is becoming a thinking partner, not just a coding tool.

⏳

Time Saved & Quality

How much time AI saves — and how often you need to fix its work.

Time Saved

NOVA: 40% save 0-20%, 36% save 20-50%, and 12% hit the 50-70% zone. That last group has either mastered prompting or mastered delegation. Either way, they're ahead. 😏

Fix Frequency

NOVA: Rating 3 ("sometimes") dominates with 10 people — a perfect bell curve. AI gets it right enough to trust, wrong enough to keep you sharp. The real deal. 🎯

🎭

When AI Gets Confident

The most "confident" AI errors caught recently.

"CLI commands that are non-existing and then pointing out that I might have done stuff wrong. Only accepted it was wrong when I pasted the official documentation."

🤡 The AI gaslit this person. "No, YOU ran the wrong command." Peak behavior.

"Changed working code of setting correct test status to Pass/Fail to code that always set tests as Pass."

🚨 "All tests passing!" — the tests, lying. The AI didn't fix the tests, it fixed your conscience.

"On dev: insisting on deprecated frameworks. Outside dev: suggesting places on the wrong neighbourhood for trip plans."

🗺️ At least the deprecated framework won't send you to the wrong neighborhood on vacation.

"Multiple, syntax errors, fundamental code errors, not following design patterns told in the prompt. The time saved writing code is eaten by debugging and refining prompts."

🔥 They're not angry, they're experienced. The voice of someone who's been through the wars.

🏆 Standout Response

The Promptinator

"I haven't seen a confident error in a long time... I guide and orchestrate it. AI results depend entirely on the quality of the context and the skills/rules you provide."

This person has transcended the "ask and hope" paradigm. They don't prompt — they direct. When you give AI the right context, clear rules, and a well-defined plan, it stops guessing and starts delivering. The future isn't AI replacing developers — it's developers who master AI replacing those who don't.

🏁

Key Takeaways

What the data tells us about AI adoption at Naura.

🌊

AI is mainstream

96% of the team uses AI daily. It's no longer experimental — it's part of the workflow.

🧠

Quality of input matters

The biggest differentiator isn't which tool you use — it's how well you guide it.

⚖️

Trust but verify

Nobody rated fix frequency at 5. The team uses AI as a tool, not an oracle.

📈

Room to grow

PR reviews (20%), docs as a task (2 votes) — there's untapped potential.

AI adoption at Naura is strong and getting stronger. The team gets it — AI is a multiplier, not a replacement. Keep refining those prompts, keep verifying the output, and keep sharing what works. The best AI strategy is the one that's already happening.

Thank You

NOVA signing off. Until next time. 🌙