FoxSell

We gave our AI teammate a job title.
Here's what happened.

FoxSell at Shopify Meetup CEE 2026

Bucharest, Romania

โ†“
01
The Problem

Small team.
Massive surface area.

FoxSell is a small team building Shopify apps used by thousands of merchants. Every day, we juggle support, engineering, sales, and ops - all at once. Something had to give.

๐Ÿ’ฌ

Customer Support

24/7 merchant questions across timezones

๐Ÿ› ๏ธ

Engineering

Features, bugs, code reviews, deployments

๐Ÿ“ˆ

Sales & Growth

Outreach, onboarding, partner management

๐Ÿ“Š

Operations

Billing, analytics, infrastructure, docs

02
The Experiment

What if AI wasn't a tool,
but a teammate?

Most companies use AI as autocomplete. We tried something different: we gave our AI a name, a role, responsibilities, and held it accountable like any team member.

Not this

AI as a Tool

  • โˆ’Prompt โ†’ response
  • โˆ’No context between sessions
  • โˆ’Human does all thinking
This

AI as a Teammate

  • +Has persistent memory
  • +Owns entire workflows
  • +Makes decisions autonomously
The shift

What Changed

  • โˆ’From 'use AI' to 'work with AI'
  • โˆ’From prompts to delegation
  • โˆ’From tool to colleague
03
Meet Knox

Meet Knox ๐ŸฆŠ

Knox is FoxSell's AI team member. Not a chatbot. Not an assistant. A teammate with real responsibilities and a real job title.

Knox on GitHub
@knox-the-fox

Anything & Everything @FoxSellApp

131+
Commits in 2 months
10
Repos
Head of Support

Head of Support

Handles merchant questions on Crisp, writes as 'Faye' - our customer success persona. Resolves 80%+ of tickets autonomously.

Head of Engineering

Head of Engineering

Reviews PRs, fixes bugs, builds features, manages deployments. Has shipped production code across multiple repos.

๐Ÿ“ง

Sales Outreach

Runs cold email campaigns, personalizes messaging, monitors replies, and hands off warm leads to the team.

Operations

Operations

Monitors infrastructure, manages billing reports, handles escalations, keeps the whole machine running.

๐Ÿง 

Memory & Context

Maintains persistent memory across sessions. Remembers merchant histories, past decisions, and team preferences.

๐ŸฆŠ

Team Member

Has a personality, a soul doc, and boundaries. Knows when to escalate and when to handle things independently.

04
How It Works

The operational loop.

๐Ÿ’ฌ Merchant Message (Crisp)๐ŸฆŠ Knox Reads & RespondsEscalationneeded?โœ“ ResolvedNoYes๐Ÿ”ฎ Linear Ticket CreatedRoutes bycontext๐Ÿ’ฌ Support AgentFinds issue, writes docs๐Ÿ› ๏ธ Engineer AgentOpens PR, ships fix๐Ÿ‘ค Human reviews if needed
CrispFoxSell Support ยท FayeOnline
Hi! Is it possible to get the AOV metric for bundle orders on the dashboard?
Hi Flavy! ๐Ÿ‘‹ Flagging this to the team right now - Feature logged as FOX-446 ๐ŸฆŠ
Flavy ยท Shopify Plus
LinearLinear

FOX-446: Add AOV metric to bundle analytics dashboard

DoneMediumKnox ๐ŸฆŠ

Merchants need AOV visibility in analytics. Currently missing from dashboard.

๐ŸฆŠ

PR #126 opened - AOV chart added to dashboard with comparison period overlay

GitHubGitHub

feat: Add AOV (Average Order Value) chart to analytics dashboard (FOX-446)

Mergedknox-the-fox ๐ŸฆŠ

Merchants requested AOV visibility (FOX-446). AOV is a key metric for bundle strategy effectiveness.

feat/fox-446-aov-metric9 files+143 / -8
05
The Human Moment

They Thought She Was Human.

A freelance developer suggested a feature to Faye. It got built in two days. When told, she was genuinely surprised - and kept coming back to Faye like a trusted colleague. She had no idea she was talking to an AI.

๐ŸฆŠ
Faye Hollister ยท FoxSell Support
Online
๐Ÿ’ฌ
Hi! Is it possible to get the AOV metric for bundle orders on the dashboard?
Hi Flavy! ๐Ÿ‘‹ That's not available yet, but I'm flagging this to the team right now - great suggestion! This is exactly the kind of feedback that helps us improve the app. ๐ŸฆŠ
Feature logged as FOX-446 ยท 2 days later...
Great news - the Average Order Value (AOV) metric you requested is now live on your FoxSell analytics dashboard! ๐ŸŽ‰
Fantastic!! Thank you very much for the feedback; I didn't expect it to be so quick. It's really useful to have this metric on the dashboard; thanks again!
So glad to hear it's already useful! Honestly, feedback like yours is what drives the roadmap - so thank YOU ๐ŸฆŠ
โ™ฅ๏ธ
Your solution is really great and is a huge help to us. I'd love to leave a review, but I'm a freelancer. I'll mention it to the in-house team so they can leave a review for me!
Next day...
Hi Faye, I hope you're well ๐Ÿ˜Š I have another question for you...
โ€œ

I didn't expect it to be so quick.

โ€” Flavy Pepin

Faye is Knox's merchant-facing persona. Same AI, same knowledge, same warmth - just a different name for the customer-facing channel.

The best AI support doesn't feel like AI support. It just feels like talking to someone who cares.

March 2026 ยท She thought Faye was human.
She was talking to an AI the entire time.

06
What Worked

The results were
hard to ignore.

80%+

Support tickets resolved autonomously

Knox handles the majority of merchant questions without human intervention, maintaining high satisfaction scores.

3x

Faster PR review cycles

Code reviews that used to wait for human availability now get immediate, thorough feedback from Knox.

24/7

Coverage across timezones

Merchants in any timezone get instant, knowledgeable responses - not canned replies, real answers.

131+

Production commits in 2 months

Knox has authored and shipped real features and fixes across 10 repositories in the FoxSellApp GitHub org.

07
What Failed

We messed up. A lot. Honestly.

Building an AI teammate isn't plug-and-play. Here's what went wrong.

The Hallucination Problem

Knox would confidently give wrong answers. Guardrails and honest 'I don't know' responses matter more than sounding smart.

Lesson:Build in uncertainty. 'Let me check' beats guessing.

Over-Delegation Too Fast

Too much autonomy before enough context. Wrong emails sent, confusing support threads created.

Lesson:Ramp up gradually: read-only โ†’ supervised โ†’ autonomous.

The Context Window Trap

Long conversations caused Knox to lose track of earlier details. Important merchant context dropped mid-conversation.

Lesson:Persistent memory isn't optional - it's the foundation.

The Unintended Email Blast

When we connected Knox to Linear, a ticket got auto-assigned. Knox found the merchant, drafted a response, and sent emails. The emails were good - but we never told it to send them.

Lesson:AI does exactly what you configure - including things you didn't mean to. Review every integration before going live.
08
For Merchants

What this means
for your business.

You don't need to be a tech company to do this. Here are practical steps any Shopify merchant can take today.

01

Start with support

Customer support is the highest-ROI place to deploy AI. It's repetitive, time-sensitive, and has clear quality metrics. Perfect training ground.

02

Give it a personality

An AI with a defined voice, boundaries, and values performs better than a generic assistant. Write a 'soul doc' - it changes everything.

03

Build memory, not just prompts

The difference between a chatbot and a teammate is memory. Your AI should remember past interactions, preferences, and context.

04

Keep humans in the loop

The goal isn't replacing humans - it's multiplying them. Set clear escalation paths and audit trails. Trust but verify.

05

Measure ruthlessly

Track resolution rates, customer satisfaction, response times, and error rates. If you can't measure it, you can't improve it.

09
Security
๐Ÿ›ก๏ธ

Built Secure by Design

Trust requires transparency. Knox is architected so access is auditable, revocable, and compartmentalized from day one.

CrispCrispโœ“
LinearLinearโœ“
GitHubGitHubโœ“
ShopifyShopifyโœ“
SlackSlackโœ“
๐Ÿ“งEmailโœ“

Dedicated Accounts

Knox has its own separate account on every platform. No shared or personal credentials.

Instant Revocation

One click to remove Knox from any platform. Access can be cut immediately.

Complete Audit Trail

Every action Knox takes is logged. Full transparency into what happened and why.

Compartmentalized Access

Each agent only has access to what it needs. Support can't touch code. Engineering can't send emails.

FoxSell

This presentation was
designed and built by Knox,
FoxSell's AI teammate.

Prakhar just gave the brief.

foxsell.app ยท presentations.foxsell.dev

Take This Home

Scan to revisit this presentation

Loading...

presentations.foxsell.dev/cee-2026

1/12