AI-Powered CS Workflows

I built tools my team actually used every day

Most CSMs talk about using AI. At Cloudflare, I shipped tools adopted across the CS and Sales organizations — saving hours of manual work weekly and raising the quality of every customer touchpoint.

15+
minutes saved per customer meeting
4hrs
saved on month-end reporting
1hr
saved daily answering customer questions
Org-wide
adoption across CS & Sales teams
See the tools in action
📋

Post-Meeting Brief

Mostly automated one-page briefs replacing manually written emails.

Usage Reporting

Painstaking manual month-end data pull turned into a mostly automated output.

🔗

MCP Workflows

Customer answers and crafted responses in minutes, not hours.

🧠

Agent Memory

Persistent context across AI sessions via Cloudflare Agent Memory — no starting from scratch.

Tool 1

The Negotiation Guidance Skill

Paste a difficult customer email into the skill and it applies the full Black Swan methodology — accusation audit, tactical empathy, labels, calibrated questions — then generates a ready-to-send email draft. Below is a realistic end-to-end example: the inbound email, the strategic breakdown, and the actual draft the skill produces. This is the skill the Sr. Director of Global Sales Programs distributed to Cloudflare's Sales org.

1
Inbound — difficult customer email
From: VP of IT, Enterprise Customer  ·  Subject: Renewal concerns
"We've been evaluating our vendor stack and frankly the value just isn't there anymore. We're going to need to have a serious conversation about our renewal. I've already reached out to two of your competitors and the pricing is significantly better. Unless something changes, I don't see a path forward here."
2
Skill output — Black Swan strategy
Accusation Audit
Name what they're thinking first: "You probably feel like we haven't been paying attention to whether this is working for you."
Label
"It sounds like the value you expected when you signed hasn't matched your day-to-day experience."
Calibrated Question
"What would it look like for this partnership to actually be working for your team?" — lets them define success before you offer anything.
What not to do
Don't defend the product, justify price, or pitch. Not yet. The goal of the first response is "that's right" — not a signed renewal.
3
Skill output — ready-to-send draft
Tool 2

Post-Meeting Brief Automation

What used to take 15+ minutes of manual writing after every customer call — a long email with bulleted text that was difficult to digest — became a clean, mostly automated one-page brief customers told me they preferred. Here's the before and after.

Before — Manual Email ~15 min to write
After — Automated Brief <2 min to send
Post-Meeting Brief Sample

The brief doesn't just save time — it's a better experience for the customer. Clean format, clear ownership, easy to act on. Customer approved.

20
Enterprise accounts
15+
minutes saved per meeting
8–16
customer meetings per week
2–4hrs
saved per week
Time saved goes back to focusing on helping customers achieve their goals.
Additional Tools

More workflow automation

Every tool solved a specific problem I was running into every week. The goal was always the same: less time on logistics, more time on customers.

📊

Usage Reporting Automation

Monthly data pulls for usage reporting and QBR prep used to eat 4 hours at month-end and 3 hours before every quarterly review. A colleague and I collaborated to turn that painstaking, manual process into a mostly automated output. That time goes back to customer conversations.

Raw usage data
Automated pull
Formatted report
QBR-ready
7+ hours saved monthly
🔗

MCP-Integrated Research Workflows

Before: hunting across Salesforce, Jira, support tickets, and Google Workspace to research and formulate a response to a customer question — 15 to 45 minutes.

After: one query to a skill connected to all of them via MCP servers. Answers and well-crafted responses in a few minutes. Better customer conversations in near-real time.

Customer question
Single query
4 systems
Crafted response
~1 hour saved daily
🧠

Cloudflare Agent Memory

AI agents are stateless by default — every session starts fresh. When Cloudflare released Agent Memory as a beta product, I got access, configured it on Cloudflare Workers, and integrated it into my workflows. Historical account context, customer preferences, prior conversation threads — all available without re-establishing context every time.

Session starts
Memory retrieved
Full context available
Cloudflare Agent Memory (beta)
🛠️

The Toolchain

None of this required a formal development background. It required curiosity, a clear sense of what problems needed solving, and a willingness to iterate until the tool actually worked the way the workflow needed it to.

Claude Opus 4.x OpenCode Windsurf / Cascade Cloudflare Workers MCP Servers
Organizational Impact

Tools my colleagues actually use

The measure of a useful tool isn't whether you built it. It's whether other people voluntarily use it.

"The negotiation skill changed how our team approaches difficult renewal conversations."
Distributed by the Sr. Director of Global Sales Programs to the Cloudflare sales organization
"Featured in multiple CS team meetings and highlighted in the weekly CS org newsletter with How-To guides."
Cloudflare Global CS Organization
"Invited to join the GCS AI Task Force to help advance AI adoption across the global CS team."
Cloudflare Global Customer Success Leadership
Let's Connect

Would it be a terrible idea to have a talk about what I could build for your CS team?

I'm actively exploring Senior CSM roles in industries such as Cybersecurity, SaaS, and Cloud Infrastructure. If you're building a team that has real problems to solve and you want someone who thinks critically, looks for efficiencies, and is a builder — let's talk.