A practical case study from a digital project manager on scoping, building, and launching a real AI-powered chatbot MVP no engineers needed."
tags: ai, project-management, startups, customer-experience, nocode
series: "From Chaos to Clarity: My Journey into Digital PM in the Age of AI"
If you're a digital PM or startup operator being told to "bring in AI" without a proper roadmap, you're not alone.
In this post, I walk through how I turned a vague CX goal into a functioning, measurable AI customer experience project without engineers, custom code, or fancy frameworks.
🧠 No engineers
🧰 No custom code
📉 Just a real business problem, scoped clearly, solved with the right tools
- The Problem Was Obvious (But Nobody Scoped It)
We reviewed customer support tickets. One issue dominated:
“Where’s my order?” made up 30% of our volume.
Leadership wanted to “use AI to improve CX.” AI isn’t a strategy it’s a tool.
So I reframed it as:
Reduce repetitive “Where’s my order?” support requests by 50% using AI automation
That’s how solid projects start: with a clear pain point and KPI.
- MVP Plan: What We Did (and Didn’t Do)
We didn’t hire engineers or build complex systems.
We used:
- ✅ Chatbase to build a GPT powered chatbot
- ✅ Google Sheets as a simple order status database
- ✅ Zapier to connect the chatbot to the sheets with no code
Time from planning to live test? Just 10 days.
- Results After 2 Weeks
Tested on 20% of site visitors:
Metric | Result |
---|---|
Queries handled by AI | 58% |
Customer satisfaction | ↑ from 3.5 → 4.4 |
Ticket volume | ↓ 42% |
Escalations | ↓ 47% |
These quick, real results built support for a full rollout.
- Lessons for Other PMs If you're leading AI adoption, focus on: Start with repetition, not ambition then tackle the most common queries Pick tools that fit your stack and ignore buzz, choose utility Scope fast, iterate fast,the MVPs that work beat perfect plans. Measure impact AI only matters if it saves time or improves experience
Let’s Discuss
Have you done your own AI MVP?
What hurdles did you face?
Share your experience or your toughest challenge below 👇
Coming next in the series: Part 3 “No Engineers Needed: AI Tools That Actually Work for Small Teams”
Connect with me:
🔗 LinkedIn – Sonnia Okoye
🐙 GitHub – @sonniaokoye
Top comments (0)