Meridian Consulting Group — AI Academy Case Study
ACA Teaching Case Study

Meridian Consulting Group

From Assessment to Academy: A Complete Engagement Arc — She Leads AI Academy | AI Consulting Accelerator | Cohort 1
This is a fictionalized case study based on a real client engagement. All names, organizations, and identifying details have been changed. The methodology, outcomes, and teaching patterns are real.

Meridian Consulting Group — AI Academy Case Study

From Assessment to Academy: A Complete Engagement Arc

She Leads AI Academy | AI Consulting Accelerator | Teaching Case Study

NOTE — This is a fictionalized case study based on a real client engagement. All names, organizations, and identifying details have been changed. The methodology, outcomes, and teaching patterns are real.


THE CLIENT

Meridian Consulting Group is a philanthropy consulting firm specializing in campaign counsel, feasibility studies, and grateful patient programs for hospitals and higher education institutions. The firm has three principals and several independent contractors, all working remotely.

Key people

  • Joan — Founder / Visionary. Strategy, innovation, thought leadership. Also serves as a 5-day/week fractional leader at a health system in another state.
  • Laura Chen — President / Integrator. Runs day-to-day operations, conducts discovery calls, manages consultant relationships. Uses EOS framework.
  • Sarah Oakes — VP of Client Services. All proposals, all contracts, engagement setup. The operational backbone. Has nearly quit from overwhelm.
  • Karen Walsh — Consultant (Campaign). Coaches philanthropy executives, trains gift officers, builds tools and workbooks.
  • Tom Bennett — Consultant (Campaign). Campaign planning studies, 80+ interviews per study. Self-described late technology adopter.
  • Rachel Diaz — Consultant (Grateful Patient / Physician Engagement). All clinician training for the firm. Deliverable-based pricing.
  • Mia — Designer. All proposals and reports flow through her for formatting.
  • Cindy Nakamura — Consultant. Board retreats, strategic planning.
  • Rebecca Lane — Team member. Research and content support.

THE CONTRACT — Standard Consulting SOW Structure

Before any work begins, a Statement of Work is signed. This is the standard modular structure.

SOW Template — Consulting Engagement

Prepared for: [Client Name], [Organization], [Address]

Prepared by: Anne Murphy Philanthropy LLC dba She Leads AI, [Address]

Primary Contacts: Anne Murphy (Founder and CEO), Don Wackerly (VP Strategic Partnerships)

Date: [Date]; Effective upon execution by both parties

Service Blocks (modular — pick what the client needs)

Block 1 — AI Impact Assessment ($2,000)

  • Individual 30-minute interviews with selected team members (daily workflows, pain points, efficiency opportunities)
  • Evaluation against broad range of AI applications
  • Identification of safeguards (data privacy, ethics, operational risk)
  • Deliverable: preliminary AI roadmap outlining targeted AI education, governance considerations, and potential AI solutions

Block 2 — Responsible AI Academy ($2,000/participant, minimum 5; $500/person thereafter)

  • Five facilitated training sessions, 60 minutes each, delivered virtually
  • Weekly over five consecutive weeks
  • Office hours and facilitated support sessions as needed

Block 3 — AI Policy Review and Advisory Support ($4,000)

  • Review of existing AI policies
  • Draft and final versions of new AI policy
  • Advisory recommendations informed by assessment, Academy, AI legislation, and industry standards

Block 4 — Custom AI Solutions (variable pricing)

  • Scoped after assessment findings
  • Could include: custom GPTs, knowledge bases, workflow automation, document generation

Standard Contract Clauses

Payment terms — Fixed fees, invoiced per agreed services, due within 15 days of receipt. Pricing valid 30 days.

Term and termination — Begins upon execution, concludes upon completion. Either party may terminate with 30 days written notice; Client invoiced for completed services.

Confidentiality — Both parties maintain confidentiality. Consultant retains rights to proprietary frameworks, materials, and methodologies. Will not retain client data beyond what is necessary.

Responsible AI commitment — Consultant committed to ethical, transparent, secure, equitable AI practices. Client agrees to good-faith responsible AI efforts.

How to adapt this for your client

The SOW is modular. Some clients need all four blocks. Some need only the assessment. The assessment always comes first because it reveals what the other blocks should contain. Block 3 (AI Policy) can be offered at no cost as a bonus to close the deal — Anne did this with one client to sweeten the package.


PHASE 1 — THE ASSESSMENT ($2,000)

What we did

Six individual interviews, 30 minutes each, all recorded. Interviewed Joan (founder), Laura Chen (president), Sarah Oakes (VP ops), Karen Walsh, Tom Bennett, and Rachel Diaz. Let leadership nominate who to interview, including resistors.

What we found

Assessment results by category

CategoryTime SavingsPriority ItemsBiggest Finding
Stakeholder trust and data handling20-30%3 highNo formal data handling policies. No protocols for AI meeting recorders when consultants share confidential donor information.
Service delivery and operations20-30%2 highInconsistent approaches across consultants. Interview analysis is a major time drain — budgeting 1 hour per interview plus 1 hour to write the summary.
Client and stakeholder communications50-70%2 highTotal customization of every proposal creates bottlenecks. Logjams from never standardizing proposal language.
Internal operations50-80%2 highLast-minute template requests. 2-week turnaround on final reports with heavy designer back-and-forth. Rush job culture eroding goodwill.
Knowledge management30-70%1 highClumsy document management. Template disagreements. 100+ campaign documents with no sharing system.
Future considerations40-60%0 highBusiness development and content strategy as growth opportunities. Bespoke tools to explore later.

Overall average: 35-45% potential time savings across all categories.

Process maps (8-10 boxes per level)

#### Laura Chen — President / Integrator

[Inbound Lead] → [Discovery Call] → [Handwritten Notes]

↓ ↓ ↓

[Notes to Sarah] → [Proposal Drafting] → [Proposal Review]

↓ ↓ ↓

[Pitch] → [Contract/Kick-off] → [Ongoing Management]

Bottlenecks: Proposal iterations (designer + Sarah + consultant + Laura), Joan unresponsive to leadership decisions, core processes never finalized.

#### Sarah Oakes — VP Client Services (single point of failure)

[Lead Arrives] → [Needs Assessment] → [Template Selection]

↓ ↓ ↓

[Proposal Drafting] → [Designer Formatting] → [Review Rounds]

↓ ↓ ↓

[Pitch Support] → [Contract Execution] → [Engagement Kick-off]

Bottlenecks: PDF comment revision cycle, last-minute edits from consultants, ALL proposals funnel through Sarah, cannot edit PDF directly.

#### Tom Bennett — Consultant (Campaign)

[Client Assignment] → [Initial Setup] → [Case Development]

↓ ↓ ↓

[Prospect ID] → [Interviews (80+)] → [Manual Analysis]

↓ ↓ ↓

[Report Writing] → [Campaign Tracking] → [Materials Library]

Bottlenecks: Reading 80 interview responses manually, inconsistent report format, no knowledge sharing system.

The 5 systemic issues

  1. No standardized processes despite EOS adoption. Every consultant does things their own way.
  2. Sarah Oakes is a single point of failure at burnout risk. All proposals and contracts funnel through one person.
  3. Interview analysis is done by hand firm-wide. 80+ hours of manual synthesis per feasibility study.
  4. Joan is unreachable and strategic decisions stall. Two full-time jobs. Weeks pass without responses.
  5. Consultant fear of compensation impact blocks AI adoption. If AI makes them faster, will their hours (and pay) get cut?

Task-level analysis with priorities and time savings

This is the granular breakdown delivered to the client — every task scored by priority and estimated time savings.

Priority 1 (address immediately)

TaskTime SavedRationale
Internal Communications50-70%AI can draft memos, newsletters, announcements with clear, consistent messaging
Data Privacy and Compliance20-40%AI can research regulations, ensure compliance, draft privacy policies
Market Research30-50%AI gathers and summarizes market data, competitor analysis, customer insights
Strategic Planning20-40%AI brainstorms strategies, runs SWOT analysis, predicts outcomes from data
Campaign Implementation20-30%AI schedules campaigns, coordinates tasks, generates performance reports
Content Creation40-60%AI drafts copy, posts, articles, scripts. Image generation adds visual capability
Customer Relations30-50%AI drafts emails, handles inquiries, analyzes feedback. Advanced: sentiment analysis
Long-term Strategic Planning30-50%AI models scenarios, assesses risks, optimizes multi-year strategies
Complex Decision Making10-20%AI weighs options with multiple variables, provides data-driven recommendations
Creative Problem Solving10-30%AI generates ideas, brainstorms solutions, evaluates feasibility
Negotiation and Deal-Making10-20%AI analyzes terms, identifies leverage points, suggests strategies

Priority 2 (next wave)

TaskTime SavedRationale
Crisis Management20-30%AI drafts statements, responses, communication plans
Budgeting50-70%AI generates templates, allocates resources, tracks spending
Sales Forecasting40-60%AI analyzes historical data, trends, seasonality
Performance Analysis and Reporting60-80%AI collects, analyzes campaign data, generates comprehensive reports
Email Marketing50-70%AI writes copy, personalizes subject lines, segments audiences
Event Planning20-40%AI helps with ideation, promotional materials, follow-up communications
Marketing Automation40-60%AI creates automated workflows for nurturing, email, social
Lead Generation and Nurturing30-50%AI writes lead magnets, creates drip campaigns, scores leads
Product Launch30-50%AI develops strategies, generates promotional materials, writes descriptions
Adapting to Emerging Trends20-40%AI identifies and analyzes trends, predicts impact, recommends strategies
Resource Allocation Optimization20-40%AI develops models considering budget, demographics, and historical performance

Priority 3 (foundational)

TaskTime SavedRationale
Website Optimization20-40%AI suggests improvements to copy, navigation, CTAs
Competitive Analysis40-60%AI gathers and summarizes competitor products, pricing, strategies
SEO Optimization20-40%AI suggests keywords, optimizes content, analyzes traffic
Partnerships and Influencer Marketing20-30%AI drafts outreach, tracks performance
Predictive Analytics and Modeling40-60%AI refines statistical models, uncovers patterns

Priority 4-5 (lower urgency)

TaskTime SavedRationale
Social Media Management40-60%AI schedules posts, drafts content, analyzes performance
A/B Testing30-50%AI designs tests, analyzes results, recommends winners
Public Relations30-50%AI drafts press releases, pitches, talking points
Brand Management20-40%AI maintains voice and consistency, monitors mentions
Attribution Modeling20-40%AI assesses marketing touchpoint impact, optimizes budget allocation

E0 — AI cannot help

TaskTime SavedWhy
Team Management0%Requires interpersonal communication and leadership

Three-tier solution structure (per category)

  • Low-hanging fruit — AI skills/prompt training, knowledge base with content and templates
  • Done-with-you — Custom GPTs for proposals, meeting follow-up, prospect research
  • Higher investment — Document generation automation, standardized design system, campaign analytics

Implementation timeline

  • Immediate (0-3 months): Consultant training, feasibility interview analysis workflow, proposal template with AI customization, knowledge base development
  • Short-term (3-6 months): Core processes supported by AI, client deliverable enhancement, AI-powered final reports
  • Medium-term (6-12 months): Advanced campaign analytics, business development enhancement, content marketing strategy, custom tools

THE ROADMAP — What the Client Sees as a One-Pager

After the assessment, the client receives a one-page service overview. This is the sales tool, not the full report. It frames four service offerings that flow from the assessment findings.

AI Adoption Roadmap: Delivering on the Promise of AI While Protecting Stakeholder Trust

  1. AI Readiness Assessment — Evaluates technology infrastructure, workforce and organizational culture, leadership, strategy, change management capabilities, and vendor/partnership ecosystem
  1. AI Council — Composed of key leaders and staff; provides recommendations, reviews use cases, supports AI adoption. Includes forming the council, defining mission, stakeholder communication, agenda setting, meeting facilitation, outcome reporting
  1. AI Policies — Guides AI policies to align with strategic goals and ethical standards; adaptable to evolving AI landscapes; compliant with regulations; robust governance framework
  1. AI Academy — Tailored training programs with use-case-based education for rapid AI adoption; foundation in ethical AI practices

The Roadmap positions the assessment as the entry point and shows the client the full journey. It answers the question "what happens after the assessment" before they ask it.


THE RESPONSIBLE AI PLAYBOOK — What the Client Keeps

The Playbook is the lasting artifact. After the Academy ends, this is what the team refers back to. For Meridian Consulting Group, the Playbook covered:

  1. Foundation setup — ChatGPT installation, security protocols ("I will not put confidential information into AI"), interface navigation, subscription tiers
  2. Core conversation skills — Prompting structure (greet, set context, give specifics, request format, ask for clarifying questions), advanced techniques (role assignment, bribery method, expertise assertion, brutally honest feedback), re-prompting
  3. Multimodal capabilities — Photo upload, cross-platform integration, file uploads
  4. Document creation workflows — Complete proposal writing process, proposal evaluation, template creation
  5. Meeting management systems — Fathom setup, recording consent protocol (specific language provided), transcript analysis, contact report generation
  6. Presentation and content creation — Gamma for presentations, NotebookLM for audio
  7. Advanced automation — Knowledge base development (Projects), Custom GPT development, personal writing style training
  8. Quality control and ethics — Expert-in-the-loop principle, avoiding "AI slop" (105 overused phrases), AI limitations and bias awareness, privacy levels, organizational ethics

The Playbook is titled "Using AI for Improved Outcomes in Consulting Projects (and life stuff, too!)." It is based on the actual Academy sessions, customized with the client's specific use cases and workflows.


ADAPTING THE FRAMEWORK — The Key Changes Method

When you take this framework to a different client in a different industry, you do not rebuild from scratch. You adapt. Here is the actual adaptation document used to move this framework from a philanthropy consulting firm to a property management company.

The 10 swaps

#What changesFrom (philanthropy consulting)To (property management)
1Opening/welcome"Laura and Joan and crew""Steve and the [Company] team"
2Industry experience"Nonprofit and higher education advancement""Small business operations and real estate management"
3AI expertise framingAI for nonprofitsAI for "property management and real estate businesses"
4Use case examplesGrant writing, feasibility studies, donor relationsLease agreement generation, maintenance request triage, tenant screening, market analysis, compliance monitoring, owner communication
5Target audience"Small businesses inundated with opportunities""Property management companies managing multiple properties and stakeholders"
6Key outcomesDonor trust, campaign efficiencyFaster maintenance response, rental compliance, resident retention, vendor coordination
7Academy session topicsConsultancy ethics, donor dataProperty management business ethics, protecting owners/residents/vendors
8Governance framingProtecting donor trustProtecting property owners, residents, and vendors' interests
9Title/subtitle"Responsible AI for consultancies""Responsible AI for Property Management"
10PricingNo changesSame structure ($2,000 assessment, $2,000/person Academy, $4,000 governance)

Strategic considerations for the adaptation

The property management client had specific attributes that made the framework fit:

  • 47+ years in business, technology-forward (modern software, online portals)
  • Oregon landlord-tenant law complexity (AI for compliance monitoring)
  • Multi-stakeholder management (owners, residents, vendors, staff)
  • Already data-driven (KPIs, benchmarking)
  • Scale challenges (multiple properties across three cities)

Additional use cases added for property management

  • Document management (leases, inspections, maintenance records, owner communications)
  • Predictive maintenance
  • Resident retention analysis
  • Financial optimization
  • Staff training via AI knowledge base
  • Communication automation

The teaching point for the cohort: The framework is the same. The 6 assessment dimensions are the same. The SOW structure is the same. The Academy curriculum structure is the same. You swap the industry language and use cases. That is it.


PHASE 2 — THE ACADEMY (5 sessions)

Session 1 — Getting Started (96 minutes)

Session arc

Anne opened with a hands-on exercise designed to eliminate intimidation immediately. Rather than starting with theory, she had participants download ChatGPT on their phones, photograph their pantries, and generate dinner recipes — establishing multimodal AI use within the first 15 minutes. She escalated the exercise into repurposing (recipe → dinner invitation), introduced tone and audience customization, then framed the five-session curriculum. The middle portion addressed the emotional and ethical landscape of AI adoption. The session closed with a tour of the ChatGPT ecosystem and a live demo of the Fathom meeting recorder workflow.

Teaching pattern

  1. Administrative setup while people trickled in (LinkedIn, app download, pledge)
  2. Hands-on exercise FIRST, before any lecture
  3. Escalation within the exercise (photo → recipe → invitation → tone shifts)
  4. Named what they just learned after they experienced it
  5. Ethical/emotional framing in the middle (bias, identity threat, data privacy)
  6. Tour of the full ecosystem (Projects, Custom GPTs, memory)
  7. Live demo of Fathom + transcript workflow to close

Key moments

Rebecca Lane's pregnancy prompt breakthrough. She told ChatGPT she was 34 weeks pregnant and needed protein. It explained why certain meals mattered, gave three options, and recommended option two with reasoning. The group saw personal context unlock personalized, useful output.

Laura Chen's street corn tacos. She photographed her fridge and got a creative meal suggestion. Her reaction ("That sounds really good") was genuine delight, not polite interest.

The "I'm not special" preemptive move. Anne opened her bio with "There is nothing special about my background or my education or my aptitude for me to be good at AI. This is all self-taught." This removed the expertise barrier before anyone could feel it.

Frameworks taught

  • The Marathon Analogy — You learn AI by using AI. No adjacent preparation substitutes.
  • Information → Insights → Knowledge — Clients don't need information. They need knowledge they can act on.
  • Expert in the Loop — When you're in your zone of genius, you are the quality check.
  • Context as the Lever — More context = faster path to usable output.
  • The Red Line Framework — Each person identifies their own boundaries for what they would never tell AI.

Resistance handled

  • Rebecca Lane's privacy question: acknowledged, promised future deep dive, gave enough to satisfy now
  • Technical confusion (wrong app, sync issues): normalized, provided workarounds, kept momentum
  • Identity threat: named directly ("folks are kind of having an identity crisis"), then pivoted to empowerment

Session 2 — Proposal Writing and Data Privacy

Session arc

Opened with AI image bias and ethical storytelling, then moved into a live proposal-writing exercise. Participants transformed a deliberately mediocre proposal into a polished deliverable using the role/task/context prompting structure. Anne demonstrated re-prompting, memory logging, and the proper noun anonymization rule.

Key moments

Tom Bennett's critic pivot. Instead of following the exercise as prescribed, Tom asked ChatGPT to evaluate the proposal for weaknesses rather than rewrite it. His readback of the AI's critique was a breakthrough — the group saw AI as an analytical partner, not just a content generator.

Rachel Diaz's eagerness. "I can't wait. I really want to put in one of the proposals and be like, what is wrong with this?" She was already bridging from exercise to real workflow.

The bribery technique. Anne taught "I know you have this. You are brilliant. My boss really wants me to get this right. I'll give you $100 if you do this right." She was visibly uncomfortable ("I hate that this works") but shared it because it is effective.

Frameworks taught

  • The Proper Noun Rule — Search-and-replace all PII with fictional names before uploading anything. Non-negotiable.
  • Three-Phase Proposal Pipeline — Upload rough draft → prompt for polished version → apply template formatting
  • AI as Critic vs. Generator — Ask AI to evaluate, not just produce (Tom Bennett's contribution)
  • Memory Logging — Tell ChatGPT to "log" recurring preferences so they compound over time

Resistance handled

  • Data privacy fears: validated, showed concrete anonymization protocol
  • Karen Walsh obsessing about not using email: let the between-session processing happen naturally
  • Technical confusion with ChatGPT interface: normalized, treated as expected

Session 3 — Peer Teaching and Advanced Workflows

Session arc

Anne invited Cindy Nakamura to walk the group through her board retreat project — a full consulting deliverable built end-to-end in ChatGPT over four hours. This peer teaching moment was the session's centerpiece. It led organically into data privacy practices, competitive research capabilities, and the trust/verification problem.

Key moments

Cindy Nakamura's board retreat walkthrough. She designed a complete board retreat for a foundation with three hospitals: design thinking exercises, hospital-specific scenarios, facilitator script with timing, handouts, supply checklist, post-retreat debrief from whiteboard photos, and a follow-up email. Four hours total. The group saw the full arc of what is possible.

Rachel Diaz on earned confidence. Her friend asked "Don't you feel bad not thinking of it yourself?" Rachel replied: "I've been thinking of it myself for 20 years. I'm allowed to get a little help at this point in my career."

Karen Walsh's trust problem. She said "I don't believe anything that I read on it." Anne validated it, reframed verification as a professional standard, and scheduled a targeted office hour for Karen and Sarah Oakes together.

Rachel Diaz's role-play dodge. Had to deliver a training on calling patients as a fundraiser. Was "dying inside" at role-play. Fed her existing resource into ChatGPT and got back scripts, questions, and transition phrases so thorough that the role-play was unnecessary.

Frameworks taught

  • "Start with the end in mind" workflow — Cindy's success came from knowing the full deliverable set before opening ChatGPT
  • "Badgering ChatGPT" — Anne's term for iterating past generic answers. Push harder when you have domain expertise.
  • AI intuition development — The more you use it, the more you recognize what else it could do
  • The sycophancy problem — AI is dangerously affirming. It won't push back on bad ideas.

Resistance handled

  • Karen Walsh's blanket distrust: validated, reframed as professional rigor
  • Sarah Oakes' confidence gap: acknowledged, paired with Karen for targeted support
  • Amy Lester's proprietary data concern: showed anonymization live, drew nuanced line between information and PII

PHASE 3 — WHAT THE ENGAGEMENT PRODUCED

Direct deliverables

  • AI Readiness Assessment (15-slide deck with findings, time savings, priorities, three-tier solutions)
  • 5-session Responsible AI Academy (customized to philanthropy consulting)
  • Responsible AI Playbook (comprehensive guide to using AI in consulting practice)
  • Office hours and ongoing support

What changed at Meridian Consulting Group

  • Cindy Nakamura built a full board retreat in ChatGPT (4 hours vs. days)
  • Rachel Diaz eliminated role-play from her training delivery using AI-generated scripts
  • Tom Bennett discovered AI-as-critic for proposal quality
  • Laura Chen identified the proposal assembly workflow as the highest-value AI use case
  • The team adopted Fathom for meeting recording with proper consent protocols
  • A knowledge base (ChatGPT Projects) was established for shared organizational AI use

The financial arc

$2,000 assessment → $10,000+ Academy + Playbook + Office Hours + ongoing relationship

The assessment revealed what they needed. The Academy delivered it. The Playbook made it stick.


PHASE 4 — CLIENT CULTIVATION (ongoing revenue after the Academy)

The engagement does not end when the Academy ends. Here is the menu of ongoing revenue streams. Each is a natural extension of the assessment and Academy work.

The menu

OfferingWhenRevenue
Office hours (monthly drop-in)Months 2-6Included in SOW, then $500/session
30-day Playbook check-inMonth 1Free (maintains relationship)
AI Council formation facilitationMonths 4-5$2,000-4,000
Quarterly policy review retainerOngoing$1,000-2,000/quarter
Custom GPT maintenanceOngoing$2,000/year
Annual re-assessmentMonth 12$2,000-4,000
Phase 2 Academy (advanced)Month 12+$2,000/person
Referral pipelineMonth 6+"Who else in your network is wrestling with this?"

The financial arc for Meridian Consulting Group

PhaseRevenueTimeline
Assessment$2,000Month 0
Academy (5 sessions, 5 participants)$10,000Months 1-2
Policy framework (waived as bonus)$0 (would be $4,000)Month 2
AI Council facilitation$3,000Months 4-5
Custom GPT maintenance$2,000/yearOngoing
Annual re-assessment$3,000Month 12
Phase 2 Academy$10,000Month 12-14
Total 18-month relationship$28,000+

The $2,000 assessment opened a $28,000+ relationship.

[NOTE: Full cultivation strategy and the Executive Roundtable demand generation plan are in Module 7 materials — see `aca-module7-cultivation-and-roundtable.md`]


TEACHING NOTES — For ACA Cohort Use

What makes this case study useful for the cohort

  1. It shows the full arc from assessment to delivery to outcomes
  2. The client is a consulting firm — similar to what many cohort members will serve
  3. Real resistance patterns are documented — not everyone was enthusiastic
  4. The teaching methodology is visible — Anne's session structure, exercise design, and resistance handling are all here
  5. The numbers are real — 35-45% average time savings, $2,000 entry, $10,000+ relationship

Questions for the cohort

  • Which of the 5 systemic issues would you address first?
  • How would you adapt this assessment for YOUR client's industry?
  • What would your Academy curriculum look like for a different sector?
  • Tom Bennett used AI as a critic. How could you build that into your assessment deliverable?
  • Cindy Nakamura had a breakthrough between sessions. How do you create conditions for that?

She Leads AI Academy | AI Consulting Accelerator | Cohort 1

Teaching Case Study — Fictionalized from real engagement