Customer Support & Service, and AI & Automation

I get things done, and I make the people around me better.

A decade leading global customer support, and the AI and automation I build on top of it.

10 yrs
leading people
6
systems shipped
~90%
CSAT sustained
Mike Box

You are not hiring a rehearsed pitch. You are hiring a person: driven, honest, harder on himself than any boss. I hear a problem and my first thought is how do I fix this now.

The half people skip

Before I built a single system, I led the people who ran the work.

A decade in customer support and service taught me the thing no tool teaches. Operations are people first. Get that right, and the technology finally has something worth amplifying.

Built and led a global team

A 32-person support org across the US, EU/UK, China, and the Philippines. Hiring, structure, coverage, and culture across four time zones and eight-plus languages.

Customer experience as a discipline

I ran CX as the thing the operation is built around, not a queue to clear. ~90% CSAT held steady through a full automation rollout, because service quality was the constraint I protected.

Coaching, not oversight

I get fired up when someone on the team levels up because of something I built. I lead by building trust, not carrying a big stick. The measure is whether the people around me grow.

Designed my own succession

When the highest-leverage work moved to building, I hired and trained a successor for the people-management side and handed it off cleanly. The team was stronger for it.

10 yrs
support & service leadership
32
people led, 4 regions
~90%
CSAT through automation
8+
languages supported
The half that scales

Then I taught myself to build the systems that scale it.

None of these were assigned. I saw the gap, learned what I needed, and shipped working systems that proved their value in production before they ever had a budget line.

Problem

Support capacity gated response speed and quality. We wanted to own the customer-facing voice end to end, with deeper knowledge-base integration and per-response quality scoring, instead of moving at a vendor's pace.

Approach

Built an autonomous support AI from scratch on the Claude API, wired into the existing ticketing system. Domain-specific knowledge layer, multi-step reasoning, quality scoring on every response. Shipped without formal headcount approval.

Outcome

Touches every inbound ticket and fully resolves 40 to 50% autonomously (the ceiling is data-privacy policy, not capability). Nets roughly 8 to 12 FTE of created capacity at ~90% CSAT, about $500K a year. Successor CARA 3.0 is in development.

Claude APIZendesk APINode.jsCustom KB
Problem

About 35K RMA records sat in a spreadsheet with no real-time view for engineering, product, or QC. Quality spikes surfaced weeks late, in monthly review.

Approach

Built a dashboard with incremental sync (135x faster than full pulls), multi-device trend overlays, and severity-scored spike alerts.

Outcome

Quality data is now shared across engineering, product, and QC. Per-device drill-downs surface SKU patterns monthly reports miss, and spike alerts catch issues weeks earlier.

ExpressSQLiteBitable APIChart.jsOpen the live demo
Problem

Fraudulent lookalike sites were impersonating the brand on similar domains. Nothing detected them; issues surfaced only when customers complained.

Approach

Python scanners for DNS-twist analysis and Certificate Transparency logs, plus a search crawler, feeding a dashboard with a documented takedown workflow.

Outcome

Catches lookalike domains in CT logs within hours of registration. A reactive, complaint-driven problem became proactive defense.

PythonNext.jsCT logsDNS-twist
Problem

A single inbound mailbox served 22 distinct audiences across 8-plus languages on a broken pipeline that mis-routed daily.

Approach

Built a classifier on Claude Haiku with a multi-rule routing table and overrides (support topics always go to global support, regardless of language). A dashboard turns human corrections into few-shot training.

Outcome

The inbox triages itself with a self-improving feedback loop, and the support-topic policy holds across all 22 categories.

Claude HaikuFeishu Mail APINode.jsPM2
Problem

Zendesk got support to a working baseline. As we matured, we wanted a platform that fit our workflows exactly: custom routing, deeper CARA integration, and the freedom to ship UI on our own schedule.

Approach

Built a complete platform from scratch, 48 routes, 19 workers, 61 database migrations, a three-panel CARA simulator, Zendesk article import, and the full ticket lifecycle.

Outcome

A working, production-ready platform, held on purpose pending an organizational chat-strategy call. Activatable the moment alignment lands. Knowing when to wait is part of shipping.

Node / ExpressPostgreSQLReactTailwind
Problem

AI tools only work if the team beside them does. Without a documented framework and real training, even strong systems gather dust.

Approach

Co-authored an enterprise-wide AI Playbook with executive leadership, stood up an internal Claude interface, and ran intro training for ~20 staff with a full enablement kit.

Outcome

The adoption framework is live across the company and Level 2 power-user training is queued. The pattern repeats: build the system, train the people, measure.

ClaudeOpenClawGammaFeishu Docs

Click any system to open the full case study. Want more than the summary? Ask my AI anything about how these were built.

What I have shipped this year.

Reverse-chronological. Roughly the last twelve months of work that landed in production or made real progress.

May 2026
Lark Mail RouterLive

Replaced a broken pipeline with a Claude Haiku classifier that routes one mailbox to 22 audiences and learns from human corrections.

May 2026
QC DashboardLive

Real-time hardware quality intelligence over a 35K-record dataset, with severity-scored spike alerts.

May 2026
COROSclaw rolloutShipped

Stood up an internal Claude interface and ran hands-on training for ~20 staff with a full enablement kit.

Apr 2026
Feishu / Lark MCPLive

Dual-server MCP architecture connecting Claude directly to internal Bitable, calendar, tasks, and docs.

Apr 2026
SENTINELLive

Brand-protection monitor watching Certificate Transparency logs for fraudulent lookalike domains.

Mar 2026
VOICELive

Feedback-intelligence platform over 12K+ support tickets, surfacing themes and sentiment in real time.

Mar 2026
SCOUTLive

Continuous competitor monitoring, quiet daily output for sales, marketing, and product.

Feb 2026
RELAYOn hold

From-scratch Zendesk replacement, built and validated, held on purpose pending a chat-strategy call.

Q4 2025
CARA 3.0In development

Next-gen agentic architecture that brings the AI stack fully in-house. Active development.

2025
CARAIn production

The original autonomous support AI, and the anchor of everything above.

Anyone can pick a side. My work lives in the and.

Knowing systems is one thing. Knowing systems and people, so they actually work together, is the whole job. Three places it shows up.

AI that frees people, it does not replace them

CARA takes the repetitive 40 to 50% so the team spends its time on the hard, human half: judgment, relationships, the edge cases. The capacity it creates is not headcount removed, it is headcount freed to do work that actually needs a person. We held ~90% CSAT straight through the rollout.

Ship the system, then enable the team

A tool nobody adopts is a hobby. So the systems ship with the other half. I co-authored an AI Playbook with leadership and ran hands-on training. The build and the adoption both get my time, on purpose. That is why the systems actually stick.

Left the team stronger to go build

When building became the higher-leverage work, I did not just walk. I hired and trained a successor for the people side, handed off cleanly, and moved into a full-time AI and Automation role. The measure of the work is whether the people around me are better for it, not whether I am indispensable.

Four things I actually believe.

People first, then the tech

Operations are people before they are systems. Get the team, the coverage, and the culture right, and the technology finally has something worth amplifying. Skip it, and you have just automated a mess.

Prove it before you pitch it

Working systems beat decks. CARA, SENTINEL, the QC dashboard all ran in production before they had a budget. The fastest way to fund something is to make it quietly indispensable first.

AI is the suit, not the hero

Iron Man works because Tony Stark is inside it. AI amplifies what good people already do, and exposes what weak process was hiding. Without the human judgment inside, the technology is just code.

I am a person, not a machine

I can excel and lead. I can also burn out and have a bad day. Those are human things. I do my best work where people get that, and where I can be the one they count on to get it done.

Two disciplines, one person.

I started at Apple support in 2014, advanced through Special Projects, and served as Team Lead Admin on the AMR Commitments Team. Nine years there taught me to treat customer experience not as a department, but as the discipline you build everything else around.

At COROS I came on to lead Customer Support in October 2023, building and running a 32-person global team across four regions. That is the half of my background people tend to skip past, and it is the half that makes the rest work.

It is one thing to know a system. It is another to know systems and people, and get them working together. That overlap is where I live.

By mid-2025 the highest-leverage thing I could do had shifted from managing people to building the systems that scale them. So I designed my way into it. I hired and trained a successor for the people side and moved into a full-time AI and Automation role. CARA, SENTINEL, RELAY, the QC dashboard: none were assigned, all got built and proven before they had a budget.

I have never been the guy who interviews easy or shows off well. What you get instead is honest, driven, and steady. Find the real problem, build the thing that solves it, bring the people with you, measure everything, move on.

Looking for someone who does both?

I want roles where operational leadership and hands-on AI building are not two different people. Customer experience, support operations, or automation with real stakes. Also open to select consulting through Infinitus Digital.

Send me an email
LocationWashington State, remote
StatusOpen to conversations