ETTE · Case Study

AI Champions: the adoption model ETTE tested on itself first

Between March and June 2026, fifteen ETTE team members ran the AI Champions Program across ticket work, reporting, security analysis, and client communication, entirely in ETTE-governed tools. This page documents how it ran, what changed, and what the model looks like for clients now.

March – June 2026 Governed rollout · ETTE-managed tools and policy

15Team members enrolled

4Months from pilot to closeout

6Documented working sessions

6+Organizational skills deployed

The Model

The AI Champions operating model.

ETTE tested the model on its own service work first, then converted the lessons into a client-ready adoption path.

ETTE's own workflows

Tickets, QBRs, security analysis, compliance docs

AI Champions cohort

Pilot first, full-team rollout, mentor pairing

Shared skills library

Reusable prompts, documentation flows, review habits

Governance lessons

Tool access, AI usage policy, data boundaries, quality control

Client adoption program

The same model adapted to client workflows

The Rollout

A phased rollout built on everyday service work.

The program ran on the team's day-to-day service work, anchored by regular all-team sessions. Team members brought back what worked and what didn't, demonstrated their own work products, and shared learning resources like curated videos and live demonstrations. A small pilot group went first, and the full team joined once the pilot proved out. Useful discoveries became shared skills the whole team now runs by default.

Mentor pairing

Everyone who joined at the full-team kickoff was paired with a peer from the pilot group rather than a manager. Questions about a skill or an odd tool behavior went to someone who had hit the same wall a month earlier.

Skills as shared IP

When someone built a useful pattern, like a QBR template or a ticket documentation flow, it was reviewed and promoted to an organizational skill in the shared library. Everyone on the team gets these skills by default.

Session timeline

Mar 2026

Pilot group · Sessions 1 & 2

Small pilot group: Operations Manager, two Tier 3 engineers, two Tier 2 technicians, one Tier 1 technician

Apr 3

Pilot group · Session 3: wins & feedback

Client security report showcase; decision to expand cohort

Apr 17

Pilot group · Session 4: integrations & mentors

HelpSpot MCP planning; mentor assignments confirmed

Apr 24

Full team · Kickoff: Claude walkthrough

15 participants; tool access granted; AI policy introduced

May 4

Full team · Session 2: skills deep dive

HelpSpot connector live; skill-building workshop

May 27

Full team · Session 3: looping & Claude Code

Recursive goal loops; Claude Code context window demo

Jun 30

Closeout session

Individual callouts; usage guidance; client proposals announced

In Practice

What the work actually looked like.

Every example below comes from session notes or was described directly by the person who did the work. All of it ran inside ETTE's organizational accounts, in tools ETTE manages and governs, under the AI usage policy introduced at kickoff.

Client reporting

QBRs: screenshot deck to polished report

Our Operations Manager redesigned ETTE's quarterly business review format from scratch. The old format relied on screenshots pasted into slides. Using Claude's Cowork mode and a refined skill, he rebuilt the QBR workflow so that raw system data goes in and a formatted, professional client report comes out. The same approach was applied to a client's board-level security presentation, co-produced with a Tier 3 security engineer.

A two-week deliverable produced in minutes

Data analysis

Government client 5-year wireless analysis

ETTE manages wireless infrastructure at five government agency sites under a multi-year contract. Our Program Co-Lead needed a comprehensive five-year analysis to support a contract renewal, a task spanning roughly 300 monthly reports. Claude's SharePoint connector ingested the documents and the contract, and the team produced a draft that would have required thousands of dollars of manual analysis time.

300 reports analyzed; draft produced in one session

Security

Client security data → board-ready presentation

A client's board needed a security briefing on very short notice, and the underlying material was dense: screenshots from email security tools, Microsoft Secure Score exports, and reports from other security platforms. A Tier 3 security engineer consolidated all of it, worked it through Claude to surface the risks a non-technical audience needed to see, and iterated the result into a board-ready presentation.

Board-ready security briefing, turned around same day

Automation

Monthly invoicing script, manual to fully automated

A Tier 3 systems engineer ran a monthly invoicing script that required entering credentials and completing two-factor authentication manually for each Microsoft-only client. Claude walked through creating an app registration with the right permissions and a stored secret. The script now runs without any manual intervention.

Recurring monthly manual task eliminated entirely

Infrastructure

AWS LightSail deployment and code audit

A Tier 2 infrastructure engineer's first LightSail deployment took two days of manual SSL configuration, PHP setup, and Amazon-specific quirks. The next one, with Claude reading the documentation live and walking through each step, was done in a couple of hours. He also started a Claude Code audit of a long-running legacy codebase.

2-day deployment compressed to a few hours

"Multiple customers have now commented: 'Wow, you guys really know about this stuff — that's really impressive what you're doing.' That's a differentiator for us. We want to make sure we're pulling away from the pack of other MSPs."

— Program Lead, Closeout Session · June 30, 2026

The Results

What changed by the time it closed.

Ticket and note quality

Program leadership observed a clear improvement in ticket documentation across the team. Notes became more descriptive and more useful to colleagues, and client communications got more professional.

Client satisfaction

Client survey results improved during the program period. In direct client conversations, feedback had been "very, very happy." Multiple clients specifically commented on ETTE's AI knowledge as a differentiator.

Organizational skills library

Six or more organizational skills were deployed to all team members by closeout, covering client communication, ticket documentation, QBR generation, onboarding, and performance reporting. Each skill is version-controlled and available by default.

Mentorship structure

Everyone who joined at the full-team kickoff was paired with a mentor from the pilot group. Pairings were skill-matched where possible.

Continued learning

Training continued past closeout. Team members, including program leadership and several engineers, are pursuing further structured AI learning as part of how ETTE keeps its practice current.

The People

AI Champions across roles.

Descriptions reflect documented contributions from session notes and program closeout observations.

Program Lead

Management · Operations

Designed and ran the program. Built the organizational skills library in GitHub, set up the mentorship pairings, and led all six sessions. Used Claude to compress recurring compliance documentation work that would normally take months into a matter of weeks.

"We're already proposing engagements like this to potential customers — taking the lessons, good and bad, into future client programs."

Program Co-Lead

Management · Client Services

Co-led the program and handled client-facing AI strategy. Used Claude to analyze 300 monthly wireless reports across five government agency sites for a five-year analysis. Connected financial and payroll systems through Claude connectors.

"Let's utilize it while we have it — the subsidy era is probably coming to an end."

Operations Manager

Operations

One of the program's most prolific contributors. Rebuilt ETTE's QBR format, replacing screenshot-heavy slide decks with data-rich client reports. Uses Claude daily for research, threat analysis, and report generation, after starting as a regular ChatGPT user.

"I cannot be more grateful about it. I was using a lot of ChatGPT but Claude is way better."

Tier 3 Security Engineer

Security

Turns dense technical data into client-friendly deliverables. Identified an 8–10% Secure Score improvement achievable without disrupting users. Reviews every AI-assisted deliverable before it goes to a client.

"Always review everything — your work has to look consistent and professional, not just copy-pasted from AI."

Tier 2 Infrastructure Engineer

Development / Infrastructure

Went deepest on the technical side. Cut an AWS LightSail deployment from two frustrating days to a few hours. Started a Claude Code audit of a long-running legacy client codebase. Wired up HelpSpot's API for private ticket documentation.

"I'm using it to document tickets more than anything lately — and it can respond directly to the ticket privately via the API. That's great."

Tier 2 Technician

Support

Cut user guide creation from hours to roughly five minutes. Used the HelpSpot connector to find historical tickets from a vague memory cue. Drafted sensitive client emails in several tone variations and picked the best fit.

"It was more of five minutes for me — what it would take way longer than that. I saved a lot of time with that."

Tier 3 Systems Engineer

Systems Administration

Used Claude to set up Microsoft Graph API app registration for ETTE's monthly invoicing scripts, which used to require manual credential entry and a two-factor prompt on every run. Recognized at closeout for standout documentation quality.

"I am loving it right now. I don't need to do that manually anymore."

Tier 1 Technician

Security Support

Runs suspect email headers and URLs through Claude for a rapid first-pass phishing assessment. Also uses it to draft clearer responses to non-technical users on security tickets.

"I copy the headers and ask Claude to do some research — also some URLs, suspected web pages, just to be sure they're safe or not."

Tier 1 Technician

Infrastructure

Joined at the April 24 kickoff. Planned and executed a migration in a single day that would previously have taken several days. Leans heavily on screenshot uploads for visual troubleshooting.

"It just became my second best friend — I love the part that I can upload a lot of pictures to troubleshoot."

What ETTE Now Offers

AI Champions cohort engagements for client organizations.

ETTE now runs the same phased-cohort approach inside client organizations: tool selection and access, an AI usage policy, mentor pairing, structured sessions, and an organizational skills library built around the client's actual workflows. These engagements are part of ETTE's AI Strategy & Enablement practice.

Clients in legal, government contracting, financial services, and nonprofit work have already expressed interest. ETTE ran the program on its own team before offering it to anyone else, and its sessions, usage policy, and skills library came out of that experience.

Talk With an Advisor Explore AI Advisory Services