Build an AI Transformation Roadmap
Plan a 90-day AI strategy for a B2B affiliate marketing company.
Introduction
30 Second Summary
Every company wants to use AI, but most don't know where to start. Without a clear plan, AI projects stall, waste budget, or solve the wrong problems. The difference between success and failure is a structured roadmap that identifies the highest-value opportunities first.
In this project, you will build a professional 90-day AI transformation roadmap for a fictional B2B affiliate marketing company. You will assess current operations, score AI opportunities using a weighted matrix, and design a phased implementation plan.
What You'll Build
A complete AI transformation strategy for "PartnersFirst Media," a 25-person B2B affiliate marketing company, using Gemini (via your Google AI Pro plan) and Claude to generate realistic fictional company documents and produce strategic deliverables.
By the end of this project, you'll have:
- A fictional company profile and operations assessment documenting current workflows across Google Workspace, Slack, Notion, and HubSpot.
- A scored AI Use Case Prioritization Matrix ranking 10 AI agent opportunities by business value, data readiness, feasibility, and time-to-value.
- A 3-phase implementation roadmap (Foundation, Pilot, Scale) mapped to a 90-day timeline with governance and change management protocols.
- Secret Mission: Create a "Week 1 Quick Wins" playbook with 3 immediately-deployable AI automations using the company's existing tools.
Are there any prerequisites?
You need a Google AI Pro plan (which gives you access to Gemini in Google Docs) and a free Claude account at claude.ai. You will also sign up for free accounts on Notion and HubSpot CRM during the first step. No coding or prior AI strategy experience is required.
Before We Start
Before We Start...
Set Up Your Workspace
Before you can build a professional AI transformation roadmap, you need a workspace that mirrors how real strategy consultants operate. Your deliverables will live across multiple tools, each serving a specific purpose in the planning process.
In this step, you will set up three browser-based tools. Google Docs is where your final roadmap deliverable will live (powered by Gemini), Notion is where you will build your AI opportunity scoring matrix, and HubSpot is the CRM you will reference when documenting the fictional company's workflows.
In this step, get ready to:
- Create your roadmap document in Google Docs and enable Gemini.
- Set up a free Notion account and create your project page.
- Create a free HubSpot CRM account and explore the interface.
Create your Google Doc and enable Gemini
Your final roadmap deliverable will be a Google Doc. With your Google AI Pro plan, you have access to Gemini directly inside Docs, which will help you draft and refine content throughout the project.
- Open your browser and navigate to docs.google.com.
- Sign in with your Google account that has the AI Pro plan.
- Click the + (blank document) option to create a new document.
- Click on Untitled document at the top left and type PartnersFirst Media - AI Transformation Roadmap as the title.
- At the top right of the document, click Ask Gemini to open the Gemini side panel.
The Gemini side panel should now appear on the right side of your document. You will use this panel throughout the project to generate and refine content.
What is Gemini in Google Docs?
Gemini in Google Docs is an AI writing assistant built into the side panel. It can draft content, summarize text, reference other files from your Google Drive, and pull information from the web.
Your Google AI Pro plan gives you expanded usage limits and access to Gemini's latest model. This makes it ideal for generating detailed strategy documents.
Don't see Ask Gemini?
Make sure you are signed into the Google account that has your AI Pro subscription. If you are signed into a different account, click your profile icon in the top right corner of the browser and switch accounts.
help me find the Gemini button in Google Docs
Set up your Notion workspace
Notion's free plan includes databases with formula properties. You will use this to build a scored prioritization matrix that auto-calculates weighted scores for each AI opportunity.
- Open a new browser tab and navigate to notion.so.
- Click Get Notion free (or Sign up) and create an account using your email or Google login.
- Complete the onboarding prompts that Notion shows you.
- Once you reach your Notion workspace, click + New page in the left sidebar.
- Type AI Transformation Project as the page title and press Enter.
You should now see a blank page titled AI Transformation Project in your Notion sidebar.
Why Notion for the scoring matrix?
Notion databases support formula properties that automatically calculate values from other fields. This means your prioritization matrix will dynamically rank AI opportunities as you adjust scores, without manual recalculation.
The free plan gives you unlimited pages and blocks, plus full database functionality including formulas, relations, and multiple views.
Stuck on Notion's onboarding?
Notion's onboarding flow may ask about your team size and use case. Choose any options that feel relevant. These selections only affect which templates Notion suggests. They do not limit your workspace features.
help me get through Notion's onboarding setup
Create your HubSpot CRM account
The fictional company in this project uses HubSpot's free CRM for contact management, deal pipelines, and email tracking. By exploring the interface yourself, you will write more realistic assessments of the company's current workflows.
- Open a new browser tab and navigate to hubspot.com/products/crm.
- Click Get started free.
- Sign up using your email address or Google login. No credit card is required.
- Complete the brief onboarding questions HubSpot asks about your role and goals.
- Once you reach the HubSpot dashboard, take a moment to explore the main navigation.
- Click Contacts in the top menu to see the contact management area.
- Click Sales and then Deals to see the deal pipeline view.
You should now see HubSpot's visual pipeline board with sample deal stages. You do not need to add any data yet.
What is HubSpot's Free CRM?
HubSpot's Free CRM is a browser-based customer relationship management tool. It includes contact management, deal pipelines, email tracking, and a basic reporting dashboard.
The free tier has no expiration date and supports up to 2 user seats with 1,000 contacts. For this project, you are using it as a reference to understand what the fictional company works with daily.
Having trouble signing up?
HubSpot supports multiple sign-up methods including email, Google login, and Apple ID. If one method is not working, try another. Make sure you select the free option and not a paid plan trial.
help me create my free HubSpot CRM account
Your workspace is ready. Next up, you will use Gemini to create a detailed fictional company profile for PartnersFirst Media, complete with operational documents that will inform your AI transformation assessment.
Create the Fictional Company Profile
Your workspace is ready. Now it is time to populate it with the fictional company you will analyze throughout this project.
Every AI transformation starts with understanding the current state of a business. In this step, you will use Gemini and Claude to generate a realistic company profile and operational documents for "PartnersFirst Media." These fictional documents serve the same purpose as real discovery interviews and process audits in a consulting engagement.
In this step, get ready to:
- Generate a detailed company profile for PartnersFirst Media using Gemini in Google Docs.
- Create 3 fictional operational documents using Claude.
- Assemble a Current State Brief with executive context in your Google Doc.
Generate the company profile with Gemini
A strong AI transformation roadmap starts with a detailed company profile. This gives you the context to identify where AI fits and where it does not.
- In your Google Doc, click the Ask Gemini sparkle icon in the top-right corner to open the Gemini side panel.
- Paste the following prompt into the text box at the bottom of the Gemini side panel:
Generate a detailed company profile for "PartnersFirst Media," a fictional B2B affiliate marketing company. Include the following details:
- Company size: 25 employees
- Business focus: Manages 200+ affiliate partnerships across SaaS and fintech verticals
- Tool stack: Google Workspace (Business Standard), Slack (Pro plan), Notion (Business plan), HubSpot (Free CRM)
- Departments: Sales, Marketing, Operations, Partner Management, Finance
- Revenue model: Commission-based with performance bonuses
- Key metrics to include: total partner count, monthly active affiliates, average deal size
Write this as a formal company overview section suitable for a strategic planning document. Use professional language and include realistic numbers for each metric.
What does this prompt do?
This prompt asks Gemini to generate a realistic company profile that mirrors what you would find in a real consulting engagement's discovery phase. The specific details (25 employees, 200+ partnerships, named departments) give Gemini enough constraints to produce consistent, believable output.
The tool stack (Google Workspace, Slack, Notion, HubSpot) will become important later when you map AI opportunities to specific platforms.
- Press Enter to send the prompt.
- Review Gemini's output. You should see a structured company profile with sections for each category you specified (company size, departments, revenue model, metrics).
- Click Insert to add Gemini's response directly into your Google Doc.
Regenerate if needed
If Gemini's output feels too generic or is missing any of the required categories, click the regenerate button and try again. AI outputs vary between runs, so your profile will look slightly different from other learners' profiles.
Gemini side panel not responding?
Make sure you are signed into your Google account with the AI Pro plan. The Gemini side panel requires an eligible subscription. If the sparkle icon is not visible, try refreshing the page.
help me troubleshoot the Gemini side panel in Google Docs
Generate fictional operational documents with Claude
A company profile tells you what a business looks like on paper. Operational documents reveal how work actually gets done day to day. You will now generate three documents that represent the messy reality of PartnersFirst Media's workflows.
- Open a new browser tab and go to claude.ai.
- Sign up for a free Claude account (or sign in if you already have one).
- Start a new conversation and paste the following prompt:
Generate 3 fictional operational documents for "PartnersFirst Media," a 25-person B2B affiliate marketing company managing 200+ affiliate partnerships across SaaS and fintech verticals. The company uses Slack (Pro plan), Notion (Business plan), and HubSpot (Free CRM).
Create these 3 documents:
1. A weekly team standup transcript from a Slack #partnerships channel. Include 4-5 team members discussing partner updates, blockers, and wins from the past week. Use realistic names and specific partner/deal references.
2. A partner onboarding checklist from Notion. Include 10-12 checklist items covering the steps to bring a new affiliate partner from signed contract to active status. Show which team is responsible for each item.
3. A deal pipeline summary from HubSpot's free CRM. Include 8-10 deals at various pipeline stages (New Lead, Qualified, Contract Sent, Onboarding, Active) with partner company names, deal values, and brief status notes.
Make each document realistic and detailed enough to serve as a basis for operational analysis. Format each document with a clear heading.
Why use Claude for these documents?
Claude excels at generating longer-form structured content with consistent detail across multiple sections. Using both Gemini and Claude in the same project mirrors how AI-native teams work in practice. Different tools have different strengths.
Gemini integrates directly into your Google Doc for in-context editing. Claude produces detailed standalone content that you can review and selectively import.
- Press Enter to send the prompt to Claude.
- Review Claude's output. You should see three clearly separated documents: a Slack standup transcript with named team members, a Notion onboarding checklist with 10-12 items, and a HubSpot pipeline summary with 8-10 deals.
- Select all of Claude's response and copy it to your clipboard.
Claude's output looks incomplete?
If Claude stops mid-response, type "please continue" in the chat to get the remaining content. Free-tier responses occasionally hit length limits on very detailed prompts.
help me get a complete response from Claude
Assemble the Current State Brief
You now have two pieces: a company profile (from Gemini) and three operational documents (from Claude). Combining them into a single "Current State Brief" section creates the foundation your entire roadmap will build on.
- Switch back to your Google Doc.
- Scroll below the company profile that Gemini already inserted.
- Type the heading Current State Brief and format it as Heading 2 using the toolbar dropdown.
- Paste the three operational documents you copied from Claude beneath this heading.
Your Google Doc now has the raw material. The final piece is an executive context paragraph that frames why this company is pursuing AI transformation.
- Below the three operational documents, type a subheading called Executive Context and format it as Heading 3.
- Open the Gemini side panel again and paste the following prompt to generate the executive context paragraph:
Write a one-paragraph executive context statement for PartnersFirst Media's AI transformation initiative. The core driver is: the company needs to scale from managing 200 affiliate partners to 500 affiliate partners without proportional headcount growth. The paragraph should explain why AI transformation is a strategic priority now, referencing operational strain, competitive pressure from AI-native competitors, and the opportunity cost of not acting. Keep it to 4-5 sentences in formal business language suitable for an executive audience.
What does this paragraph accomplish?
Executive context answers the question every stakeholder asks first: "Why are we doing this now?" The scaling challenge (200 to 500 partners without proportional headcount) gives the roadmap a concrete business justification rather than a vague "AI is important" framing.
This single paragraph will appear at the top of your final roadmap document and anchors every recommendation that follows.
- Press Enter to send the prompt.
- Click Insert to add the executive context paragraph into your document beneath the Executive Context heading.
Your Google Doc should now contain three sections: the company profile at the top, the three operational documents in the middle, and the executive context paragraph at the bottom.
Document feels disorganized?
Make sure each section has a clear heading formatted using Google Docs heading styles (Heading 2 or Heading 3). This makes the document navigable and ensures later sections of your roadmap will be easy to find.
help me organize my AI transformation roadmap document in Google Docs
You now have a complete picture of PartnersFirst Media's current state. Next, you will analyze these documents to identify operational bottlenecks and score them for AI readiness.
Build the Operations Assessment Framework
Your fictional company profile and operational documents are ready. You now have a realistic snapshot of how PartnersFirst Media runs day-to-day.
The next step in any AI transformation strategy is understanding where the pain actually lives. Before recommending AI solutions, consultants document current workflows, identify bottlenecks, and quantify the time cost of manual processes. That assessment becomes the evidence base for everything that follows.
In this step, get ready to:
- Document current usage patterns, pain points, and data flows across 4 tool categories.
- Generate a Process Mapping Table covering 8 key workflows with scoring columns.
- Synthesize a Key Findings summary with a quantified business case for AI transformation.
Create the Operations Assessment section using Gemini
An operations assessment examines each tool in a company's stack and asks: what works, what hurts, and where does time disappear? For PartnersFirst Media, you will assess four tools that represent their entire operational backbone.
- Switch back to your PartnersFirst Media - AI Transformation Roadmap Google Doc.
- Place your cursor below the Current State Brief section.
- Type a new heading: Operations Assessment.
- Open the Gemini side panel by clicking the Gemini icon (sparkle shape) on the right side of your Google Doc.
- Paste the following prompt into the Gemini side panel:
Help me create an Operations Assessment for PartnersFirst Media, a 25-person B2B affiliate marketing company. For each of these 4 tools, write a subsection covering: (1) Current Usage Patterns, (2) Pain Points/Bottlenecks, (3) Manual Processes That Consume Disproportionate Time, and (4) Data Flows Between Tools.
The 4 tools to assess:
1. Google Workspace (Business Standard) - used for email, docs, sheets, calendar, video calls
2. Slack (Pro plan) - used for team communication, partner channels, standup updates
3. Notion (Business plan) - used for partner onboarding, content briefs, project tracking
4. HubSpot (Free CRM) - used for contact management, deal pipeline, email tracking
Context: The company manages 200+ affiliate partnerships across SaaS and fintech verticals. They want to scale to 500 partners without proportional headcount growth. Focus on bottlenecks that emerge specifically from affiliate partnership management at scale.
Format each tool section with clear subheadings. Be specific about the types of manual work (e.g., 'manually copying commission data from partner portals into Google Sheets' rather than 'data entry').
What does this prompt do?
This prompt asks Gemini to produce a structured assessment of four tools. The key design choices in the prompt are:
- It specifies four assessment dimensions per tool (usage, pain points, manual processes, data flows) so the output is consistently structured.
- It provides company context (200+ partners, SaaS/fintech verticals) so Gemini can generate industry-relevant bottlenecks.
- It requests specificity over generality. Vague assessments are useless for prioritizing AI interventions.
- Review Gemini's output in the side panel.
- Click Insert to add the generated content into your document below the Operations Assessment heading.
- Read through each tool section and confirm it includes all four dimensions (Current Usage Patterns, Pain Points, Manual Processes, Data Flows).
Refine if needed
If Gemini's output is too generic (e.g., "data entry is slow" without naming what data), follow up with a prompt like: "Make the pain points more specific to affiliate marketing. Name the exact manual tasks a partner manager would do, like reconciling commission discrepancies across 15 different affiliate network portals."
Don't see the Gemini side panel?
The Gemini icon appears on the right side of your Google Doc when you are signed into a Google AI Pro account. If you do not see it, confirm you are using the same Google account that has the AI Pro subscription.
Help me find the Gemini side panel in Google Docs with my Google AI Pro plan.
Generate the Process Mapping Table using Claude
An operations assessment identifies problems qualitatively. A process mapping table quantifies them. By attaching numbers to each workflow (hours per week, manual step count, error rate), you create the raw data needed to score AI opportunities later.
- Switch to Claude at claude.ai in your browser.
- Paste the following prompt into a new conversation:
Create a Process Mapping Table for PartnersFirst Media, a 25-person B2B affiliate marketing company managing 200+ affiliate partnerships across SaaS and fintech verticals. They use Google Workspace, Slack (Pro), Notion (Business), and HubSpot (Free CRM).
The table must cover these 8 workflows:
1. Partner Recruitment & Outreach
2. Content Creation & Approval
3. Link/Offer Management
4. Partner Communications
5. Performance Reporting
6. Commission Tracking
7. Compliance & Brand Safety
8. Competitive Intelligence
For each workflow, provide these columns:
- Process Name
- Primary Tool (which of the 4 tools is mainly used)
- Hours/Week (realistic estimate for a 25-person team managing 200+ partners)
- Manual Steps Count (number of distinct manual actions per cycle)
- Error Rate (Low / Med / High)
- AI Readiness Score (1-5, where 5 = highly automatable with current AI)
Format as a markdown table. Make the Hours/Week estimates realistic and specific — for example, Performance Reporting for 200+ partners likely takes 12-15 hours/week because someone is manually pulling data from multiple affiliate network dashboards.
After the table, add a one-line note explaining what AI Readiness Score means: 1 = requires significant custom development, 5 = can be addressed with existing AI tools and APIs with minimal configuration.
Why use Claude for this?
Claude excels at generating well-structured tabular content with internally consistent numbers. The prompt asks for realistic estimates across 8 workflows. Claude will consider the relationships between workflows (e.g., if Commission Tracking is high-error, Performance Reporting inherits that uncertainty) and produce coherent numbers.
This mirrors real-world AI strategy work where teams use different AI tools for different strengths.
- Review Claude's markdown table output.
- Copy the entire table from Claude's response.
- Return to your Google Doc and create a new heading below the Operations Assessment section: Process Mapping Table.
- Paste the table into your Google Doc. Google Docs will convert the markdown table into a formatted table automatically.
What should I see?
Your table should show 8 rows (one per workflow) with 6 columns. The Hours/Week column should total somewhere between 40 and 60 hours, representing significant manual effort across the team. Performance Reporting and Commission Tracking typically score highest in hours because they involve pulling data from multiple external partner portals.
Table not formatting correctly in Google Docs?
If the markdown table pastes as plain text with pipe characters, try pasting it into Gemini's side panel with the instruction: "Convert this markdown table into a Google Docs table and insert it." Alternatively, select the text and use Format > Table to manually create the structure.
Help me get my markdown table into Google Docs as a formatted table.
Write the Key Findings summary
Raw data is not a business case. The Key Findings section translates your process mapping numbers into a narrative that executives can act on. You will identify the top bottlenecks and attach a recoverable hours estimate to make the case for AI investment.
- In your Google Doc, create a new heading below the Process Mapping Table: Key Findings.
- Open the Gemini side panel and paste the following prompt:
Based on the Operations Assessment and Process Mapping Table above in this document, write a "Key Findings" summary with these components:
1. Top 3 Operational Bottlenecks: Identify the 3 workflows that consume the most hours relative to their value, have the highest error rates, or create the most downstream problems. For each bottleneck, name the specific workflow, the primary pain, and the weekly hours consumed.
2. Estimated Weekly Hours Recoverable: Calculate how many hours per week could be recovered if AI agents handled the automatable portions of these bottlenecks. Be specific — for example: "42 hours/week of manual reporting could be reduced to 4 hours with AI agents handling data aggregation, leaving humans for interpretation and partner communication only."
3. Business Impact Statement: One paragraph framing these findings as the quantified business case. Connect the time savings to the company's goal of scaling from 200 to 500 partners without proportional headcount growth.
Keep the tone professional and evidence-based. Reference specific numbers from the Process Mapping Table.
What does this prompt do?
This prompt uses a technique called document-grounded generation. By referencing "the Operations Assessment and Process Mapping Table above in this document," you are asking Gemini to base its output on the specific content already in your Google Doc rather than generating generic advice.
The three components (bottlenecks, hours recoverable, business impact) follow a standard consulting framework: identify the problem, quantify it, then frame the opportunity.
- Click Insert to add the Key Findings content below your heading.
- Review the output and confirm it names 3 specific bottlenecks with hours attached.
- Check that the total recoverable hours is realistic (aim for 40+ hours/week based on the process mapping numbers).
Adjust the numbers if needed
If Gemini's estimate lands below 40 hours/week recoverable, review your Process Mapping Table. The highest-hour workflows (Performance Reporting, Commission Tracking, Partner Recruitment) are typically 70-80% automatable. Edit the Key Findings to reflect at least 42 hours/week recoverable, which is the threshold that makes a compelling executive case for a 25-person company.
Gemini not referencing your document content?
If Gemini produces generic findings that do not reference your specific table numbers, try highlighting the Process Mapping Table in your document before opening the side panel. Then modify the prompt to start with: "Look at the highlighted Process Mapping Table in this document and use those specific numbers to write..."
Help me get Gemini to reference my existing document content in Google Docs.
Your operations assessment is complete. You now have a structured, evidence-based view of where PartnersFirst Media's time goes and where AI can reclaim it. Next up, you will turn these findings into a scored prioritization matrix that ranks specific AI agent opportunities.
Score and Rank 10 AI Agent Opportunities
Your operations assessment identified 42+ hours per week of manual work ripe for automation. Now the question becomes: which AI agents should the company build first?
Not all AI opportunities are equal. Some deliver massive business value but need data that does not exist yet. Others are technically simple but save minimal time. In this step, you will build a structured scoring matrix in Notion that ranks 10 AI agent opportunities by four criteria, giving PartnersFirst Media a data-driven deployment order.
In this step, get ready to:
- Build a scored database in Notion with a weighted formula that auto-calculates priority.
- Use Gemini to brainstorm and score 10 AI agent opportunities for B2B affiliate marketing.
- Assign priority tiers and copy the ranked matrix into your Google Doc roadmap.
Create the AI Agent Opportunities database in Notion
A prioritization matrix needs structure. You will create a Notion database with specific properties for each scoring criterion, plus a formula that automatically calculates a weighted total score.
- Switch back to your Notion workspace in the browser.
- Open your AI Transformation Project page from the sidebar.
- Click inside the page body and type /database to open the database menu.
- Select New empty database from the menu that appears.
- Click the database title at the top and rename it to AI Agent Opportunities.
What is a Notion database?
A Notion database is a structured collection where each row is its own page with configurable properties (columns). You can sort, filter, and run formulas across all entries. The free plan supports unlimited databases with formula properties.
Now you need to add the scoring properties. The default first column is already a Title property. You will rename it and add six more columns.
- Click the first column header (it says Name by default) and rename it to Agent Name. This stays as the Title type.
- Click the + button to the right of the last column to add a new property.
- Name it Target Tool and set the type to Select.
- Add four options to the Select property: Google Workspace, Slack, Notion, and HubSpot.
- Add another property called Business Value with type Number.
- Add a property called Data Readiness with type Number.
- Add a property called Technical Feasibility with type Number.
- Add a property called Time to Value with type Number.
Next, create the formula property that auto-calculates a weighted score for each agent.
- Click + to add one more property.
- Name it Weighted Score and set the type to Formula.
- Click Edit formula to open the formula editor.
- Paste the following formula into the editor:
prop("Business Value") * 5 + prop("Data Readiness") * 5 + prop("Technical Feasibility") * 4 + prop("Time to Value") * 4
What does this formula do?
The formula multiplies each criterion by a weight, then sums the results. Business Value and Data Readiness each get a weight of 5 (highest priority). Technical Feasibility and Time to Value each get a weight of 4. This means an agent that scores 5 across all criteria gets a maximum Weighted Score of 90.
Weighting business value and data readiness highest matches standard AI prioritization methodology. An agent with high business impact and clean data is easier to deploy and justify than one that needs months of data preparation.
- Click outside the formula editor to save it. You should see 0 in the Weighted Score column for any empty rows (since all number values default to empty).
Formula showing an error?
The formula editor is case-sensitive. Make sure your property names in the formula match your column headers exactly. For example, Business Value must have a capital B and capital V.
Also check that each number property is spelled identically to what appears in your formula. If you renamed a property after creating it, update the formula to match.
help me fix my Notion formula error
Brainstorm and score 10 AI agent opportunities with Gemini
With the database structure ready, you need to fill it with realistic AI agent opportunities scored against the fictional company's context. You will use Gemini to generate and score the 10 agents.
- Return to your Google Doc and click the Gemini icon in the side panel.
- Prompt Gemini with the following context and request:
Based on the PartnersFirst Media company profile and operations assessment in this document, score the following 10 AI agent opportunities on a scale of 1-5 for each criterion: Business Value, Data Readiness, Technical Feasibility, and Time to Value.
The agents are:
1. Partner Outreach Drafter (Google Workspace) - drafts personalized partner outreach emails
2. Commission Dispute Resolver (HubSpot + Sheets) - analyzes dispute claims against commission data
3. Content Brief Generator (Notion) - creates content briefs from campaign parameters
4. Weekly Performance Reporter (Notion + Sheets) - compiles weekly affiliate performance summaries
5. Partner Onboarding Assistant (Notion) - guides new partners through onboarding steps
6. Campaign ROI Forecaster (Sheets) - predicts campaign returns based on historical data
7. Slack Standup Summarizer (Slack) - summarizes daily standup threads into action items
8. Lead Qualification Scorer (HubSpot) - scores inbound partnership leads by fit criteria
9. Contract Renewal Alerter (Google Calendar + Docs) - flags upcoming renewals and drafts outreach
10. Affiliate Fraud Detector (Sheets + HubSpot) - identifies suspicious click/conversion patterns
For each agent, provide scores and a one-sentence justification per criterion. Consider: Does PartnersFirst Media already have the data? Is the integration technically simple with their current tools? How quickly would the team see ROI?
Format as a table with columns: Agent Name, Target Tool, Business Value (1-5), Data Readiness (1-5), Technical Feasibility (1-5), Time to Value (1-5).
Why use Gemini in the side panel?
Gemini in Google Docs can read the content of your document for context. Because your company profile and operations assessment are already in this doc, Gemini can reference them when scoring each agent. This produces scores grounded in the fictional company's actual data availability and tool stack.
- Review the scores Gemini generates. Adjust any scores that seem off based on your understanding of the company. For example, if Gemini rates Data Readiness too high for an agent that relies on data the company does not currently track, lower it.
- Switch back to your AI Agent Opportunities database in Notion.
- Click New page to add the first entry.
- Enter Partner Outreach Drafter as the Agent Name.
- Set Target Tool to Google Workspace.
- Fill in the four number scores from Gemini's output (Business Value, Data Readiness, Technical Feasibility, Time to Value).
- Repeat for all 10 agents, entering each name, target tool, and four scores.
Here are the 10 agents and their target tools for reference as you enter them:
- Partner Outreach Drafter Google Workspace.
- Commission Dispute Resolver HubSpot.
- Content Brief Generator Notion.
- Weekly Performance Reporter Notion.
- Partner Onboarding Assistant Notion.
- Campaign ROI Forecaster Sheets.
- Slack Standup Summarizer Slack.
- Lead Qualification Scorer HubSpot.
- Contract Renewal Alerter Google Workspace.
- Affiliate Fraud Detector Sheets/HubSpot.
As you enter each row, the Weighted Score column should auto-calculate based on your formula.
Speed up data entry
In Notion's table view, you can click directly into cells to enter values without opening each page. Tab between cells to move quickly across columns. Enter all 10 agents in table view rather than opening each page individually.
Sort the database and assign priority tiers
With all 10 agents scored, you need to rank them by their Weighted Score and group them into deployment phases. The highest-scoring agents go into Phase 1 for immediate deployment.
- Click Sort at the top of your database (or find it in the View options menu).
- Add a sort on Weighted Score in Descending order.
- Your agents should now be ordered from highest to lowest score.
Now add a Priority Tier property to group agents into deployment phases.
- Click + to add a new property called Priority Tier with type Select.
- Add three options: Tier 1, Tier 2, and Tier 3.
- Assign Tier 1 to the top 3 agents (these deploy in Phase 1 of the roadmap).
- Assign Tier 2 to the next 4 agents (Phase 2 deployment).
- Assign Tier 3 to the bottom 3 agents (Phase 3 deployment).
Why this tier split?
The 3/4/3 split matches a common deployment cadence. Phase 1 tackles quick wins to build momentum and prove value to stakeholders. Phase 2 handles the bulk of medium-complexity agents once governance and infrastructure are proven. Phase 3 addresses agents that need more data preparation or carry higher risk.
Finally, copy your ranked matrix into the Google Doc so the roadmap is self-contained.
- In Notion, select all rows in your sorted database table.
- Copy the selection (Cmd+C on macOS or Ctrl+C on Windows).
- Switch to your Google Doc.
- Create a new section heading called AI Agent Prioritization Matrix below your Key Findings section.
- Paste the table (Cmd+V on macOS or Ctrl+V on Windows). Google Docs will convert it into a formatted table.
Alternative to copy-paste
If the copy-paste does not preserve formatting well, take a screenshot of your sorted Notion database instead. Insert the image into your Google Doc using Insert > Image > Upload from computer. Add a caption noting this is the AI Agent Opportunities matrix sorted by Weighted Score.
Your Google Doc now contains the scored and ranked prioritization matrix alongside your operations assessment.
You now have a data-driven deployment order for 10 AI agents. Next up, you will turn this ranked list into a concrete 90-day implementation roadmap with phases, milestones, and governance checkpoints.
Design the 3-Phase Implementation Roadmap
Your prioritization matrix tells you which AI agents to deploy. Now you need to answer when and how.
A phased implementation plan prevents the chaos of deploying everything at once. You will structure 90 days into three phases: Foundation, Pilot, and Scale.
In this step, get ready to:
- Create a 3-phase, 90-day implementation roadmap with objectives, activities, metrics, and governance checkpoints.
- Draft one-paragraph Agent Specs for all 10 AI agents, organized by deployment phase.
- Add a Governance & Change Management subsection with policies, risk classification, and a RACI chart.
Create the 90-day phased roadmap
A phased approach means each deployment wave builds on the infrastructure and learnings of the previous one. Phase 1 lays governance and deploys quick wins. Phase 2 expands with measurement. Phase 3 scales and formalizes.
- Switch back to your Google Doc ("PartnersFirst Media - AI Transformation Roadmap").
- Scroll to the end of your document, below the AI Agent Opportunities table you added in the previous step.
- Add a new heading: 90-Day Implementation Roadmap.
- Open the Gemini side panel by clicking the Gemini icon.
- Paste the following prompt into the Gemini side panel to generate your 3-phase roadmap structure:
Create a 90-Day Implementation Roadmap for PartnersFirst Media's AI transformation. Structure it into three phases:
Phase 1 "Foundation" (Days 1-30):
- Objectives: Establish governance scaffolding, complete data audit across all 4 tools, deploy 3 Tier-1 quick-win agents
- Key Activities: Define AI usage policies, audit data flows between Google Workspace/Slack/Notion/HubSpot, configure and launch the top 3 scored agents from our prioritization matrix
- Success Metrics: Governance policy signed off, data audit complete, 3 agents live with baseline measurements
- Governance Checkpoint: Week 4 review with department leads
Phase 2 "Pilot" (Days 31-60):
- Objectives: Deploy Tier-2 agents in production with measurement, connect cross-tool integrations, run first ROI measurement cycle
- Key Activities: Launch 4 Tier-2 agents, establish data pipelines between tools, measure time savings vs baseline, gather user feedback
- Success Metrics: 4 additional agents deployed, first ROI report generated, user adoption rate above 70%
- Governance Checkpoint: Week 8 ROI review with executive sponsor
Phase 3 "Scale" (Days 61-90):
- Objectives: Deploy Tier-3 agents, executive P&L review, document operating procedures, plan Q2 expansion
- Key Activities: Launch remaining 3 agents, compile full ROI analysis, create SOPs for each agent, present P&L impact to leadership, draft Q2 roadmap
- Success Metrics: All 10 agents operational, documented 40+ hours/week recovered, executive sign-off on Q2 plan
- Governance Checkpoint: Day 90 executive steering committee presentation
For each phase, write 2-3 paragraphs elaborating on the activities and how they connect to PartnersFirst Media's goal of scaling from 200 to 500 affiliate partners.
Why three phases instead of deploying all at once?
The 90-day cycle approach uses repeating phases of Assess, Prioritize, Execute, Measure, and Adapt. Early cycles focus on establishing visibility and foundations before scaling.
Deploying all 10 agents simultaneously would overwhelm the team with change management. Phasing lets each wave prove value before the next one adds complexity.
- Review Gemini's output in the side panel.
- Click Insert to add the generated roadmap content into your document below the "90-Day Implementation Roadmap" heading.
- Verify that each phase has four clearly labeled sections: Objectives, Key Activities, Success Metrics, and Governance Checkpoint.
Refine the output if needed
If Gemini's output is too generic or misses the affiliate marketing context, ask a follow-up in the side panel: "Make this more specific to B2B affiliate marketing operations. Reference partner onboarding, commission tracking, and content approval workflows."
Gemini not generating useful output?
If the side panel returns a short or off-topic response, try breaking the prompt into smaller pieces. Ask for Phase 1 first, then Phase 2, then Phase 3 in separate prompts. Help me get better results from Gemini in Google Docs for a multi-phase roadmap document.
Draft Agent Specs for all 10 AI agents
Each agent needs a specification that a technical team could use to begin building. The spec defines what the agent does, what it needs, who owns it, and what happens if it fails.
- In the Gemini side panel, paste the following prompt to generate Agent Specs for your Tier-1 agents (Phase 1):
For the following 3 Tier-1 AI agents (to be deployed in Phase 1, Days 1-30), write a one-paragraph Agent Spec for each. Each spec must cover: (1) which platform feature it leverages, (2) what data it needs access to, (3) who owns it (role at PartnersFirst Media), (4) the human-in-the-loop requirement, and (5) the rollback plan if it produces poor outputs.
Tier-1 Agents (top 3 from our prioritization matrix):
1. Partner Outreach Drafter (Google Workspace) - drafts personalized outreach emails for affiliate recruitment
2. [Second highest-scored agent from your matrix]
3. [Third highest-scored agent from your matrix]
Context: PartnersFirst Media is a 25-person B2B affiliate marketing company. Roles include Head of Ops, Marketing Lead, Partner Manager, and Finance Lead. The company uses Google Workspace Business Standard, Slack Pro, Notion Business, and HubSpot Free CRM.
What goes into an Agent Spec?
The five components of each spec map to real implementation decisions. "Platform feature leveraged" identifies the API or integration point. "Data access needed" defines permissions. "Owner role" assigns accountability. "Human-in-the-loop" sets the review threshold. "Rollback plan" ensures you can revert safely.
- Click Insert to add the Tier-1 Agent Specs into your document beneath Phase 1.
- Now repeat the process for Tier-2 agents. Paste this prompt into the Gemini side panel:
Now write one-paragraph Agent Specs for the 4 Tier-2 agents (to be deployed in Phase 2, Days 31-60). Same format: platform feature leveraged, data access needed, owner role, human-in-the-loop requirement, and rollback plan.
Tier-2 Agents (next 4 from our prioritization matrix):
4. [4th scored agent from your matrix]
5. [5th scored agent from your matrix]
6. [6th scored agent from your matrix]
7. [7th scored agent from your matrix]
Same company context: 25-person B2B affiliate marketing company with roles including Head of Ops, Marketing Lead, Partner Manager, and Finance Lead.
- Click Insert to add the Tier-2 specs beneath Phase 2.
- Repeat once more for the 3 Tier-3 agents (Phase 3). Paste this prompt:
Finally, write one-paragraph Agent Specs for the 3 Tier-3 agents (to be deployed in Phase 3, Days 61-90). Same format: platform feature leveraged, data access needed, owner role, human-in-the-loop requirement, and rollback plan.
Tier-3 Agents (bottom 3 from our prioritization matrix):
8. [8th scored agent from your matrix]
9. [9th scored agent from your matrix]
10. [10th scored agent from your matrix - Affiliate Fraud Detector (Sheets + HubSpot)]
Same company context. Note that Tier-3 agents are lower priority due to higher complexity or lower immediate ROI, so rollback plans should be especially detailed.
- Click Insert to add the Tier-3 specs beneath Phase 3.
- Scroll through all 10 Agent Specs. Confirm each one covers all five components: platform feature, data access, owner, human-in-the-loop, and rollback.
Replace bracket placeholders with your actual agents
The prompts above use placeholders like "[Second highest-scored agent]". Replace these with the actual agent names and tools from your Notion database before pasting. Your specific scores may have produced a slightly different ranking than another learner's.
Specs missing one of the five components?
If Gemini omits the rollback plan or human-in-the-loop requirement, ask a follow-up: "For each agent spec above, add the missing rollback plan and human-in-the-loop requirement. Be specific about who reviews outputs and what happens if quality drops." Help me get complete Agent Specs from Gemini that include all five required components.
Add governance and change management
Deploying AI agents without governance is how organizations end up in the 80% failure rate. This subsection defines the guardrails that keep your roadmap safe and accountable.
- Switch to Claude at claude.ai.
- Paste the following prompt to generate your governance framework:
Create a "Governance & Change Management" section for an AI transformation roadmap at PartnersFirst Media (25-person B2B affiliate marketing company). Include three components:
1. AI Usage Policy Outline:
- Acceptable use guidelines (what employees can and cannot use AI agents for)
- Data handling rules (what company/partner data can be processed by AI, what is off-limits)
- Output review requirements (when AI outputs must be reviewed by a human before being sent externally)
2. Risk Classification System:
- Low Risk: Reversible actions, internal-only outputs (e.g., internal summaries, draft content for team review)
- Medium Risk: Partner-facing outputs that require human review before sending (e.g., outreach emails, partner communications)
- High Risk: Financial or compliance impact (e.g., commission calculations, contract terms, fraud detection flags)
For each level, specify the review process and approval authority.
3. RACI Chart:
Create a RACI (Responsible, Accountable, Consulted, Informed) chart for AI agent deployment activities. Roles: Head of Ops, Marketing Lead, Partner Manager, AI Champion.
Activities to cover: Agent deployment approval, Data access permissions, Output quality monitoring, Incident response (agent produces harmful output), Policy updates, Training & onboarding, ROI reporting.
Format this professionally as if it would appear in a strategy document presented to a leadership team.
Why use Claude for governance content?
Claude excels at generating longer-form structured content like policy documents and frameworks. Gemini is stronger for in-context editing within Google Docs. Using both mirrors how AI-native teams actually work: different tools for different strengths.
- Review Claude's output. Verify it contains all three components: AI Usage Policy outline, Risk Classification (Low/Medium/High), and the RACI chart.
- Copy the entire governance section from Claude's response.
- Return to your Google Doc.
- Paste the governance content below your Phase 3 Agent Specs, under a new heading: Governance & Change Management.
What is a RACI chart?
RACI stands for Responsible (does the work), Accountable (signs off), Consulted (provides input before), and Informed (told after). It prevents ambiguity about who owns what when multiple roles are involved in AI deployment decisions.
- Verify that the RACI chart includes all four roles: Head of Ops, Marketing Lead, Partner Manager, and AI Champion.
- Confirm the risk classification system has three distinct levels (Low, Medium, High) with specific examples relevant to affiliate marketing.
RACI chart formatting looks messy after pasting?
If the table loses its structure when pasted into Google Docs, highlight the RACI content and use Insert > Table to create a proper table, then paste the values into cells. Alternatively, ask Gemini in the side panel: "Format the RACI chart below as a clean table." Help me format a RACI chart properly in Google Docs.
Your roadmap now has the structure a real implementation team needs: phased timelines, detailed agent specs, and the governance guardrails to keep everything accountable. Next up, you will write the executive summary that ties all of this together into a one-page leadership presentation.
Write the Executive Summary
Your roadmap now has the full technical blueprint: phased implementation, agent specs, governance guardrails, and a RACI chart. But none of that matters if leadership never reads past page one.
In this step, you will write the executive summary that sits at the top of your document and translates everything into business-stakeholder language. You will also add a Cost & Resource Requirements section so decision-makers can see the financial picture at a glance.
In this step, get ready to:
- Draft a one-page executive summary using Claude that synthesizes your full roadmap into a leadership-ready narrative.
- Add a Cost & Resource Requirements callout box with a recommended next step for executive sign-off.
Draft the one-page executive summary with Claude
An executive summary is the single most-read page of any strategy document. It needs to answer four questions in under a minute of reading time: why now, what did we find, what is the opportunity, and what is the plan.
Why use Claude for this section?
Claude excels at synthesizing long documents into structured, concise prose. You have already built the raw material across your Google Doc. Feeding Claude a targeted prompt with the key data points produces a polished executive summary faster than drafting from scratch.
- Switch back to Claude at claude.ai.
- Start a new conversation and paste the following prompt:
Write a one-page executive summary (approximately 400-500 words) for an AI transformation roadmap document. The summary is for "PartnersFirst Media," a 25-person B2B affiliate marketing company managing 200+ affiliate partnerships across SaaS and fintech verticals.
Synthesize these four sections:
1. BUSINESS CASE (Why Now): The company needs to scale from 200 to 500 affiliate partners without proportional headcount growth. Current manual processes cannot sustain this growth.
2. ASSESSMENT FINDINGS: An operations assessment across Google Workspace, Slack, Notion, and HubSpot revealed 3 top bottlenecks:
- Partner performance reporting consumes 12+ hours/week of manual spreadsheet work
- Partner onboarding takes 5 business days due to sequential manual steps across 4 tools
- Commission tracking and dispute resolution involves 8+ hours/week of cross-referencing data between HubSpot and Google Sheets
Total: 42+ hours/week of recoverable manual effort identified across 8 core workflows.
3. OPPORTUNITY: 10 AI agents were identified and scored using a weighted prioritization matrix (Business Value, Data Readiness, Technical Feasibility, Time to Value). Projected savings: 40+ hours/week once all agents are deployed. Top 3 agents: Partner Outreach Drafter, Commission Dispute Resolver, Weekly Performance Reporter.
4. THE PLAN: A 3-phase, 90-day implementation:
- Phase 1 "Foundation" (Days 1-30): Governance scaffolding, data audit, deploy 3 Tier-1 quick-win agents
- Phase 2 "Pilot" (Days 31-60): Deploy Tier-2 agents, cross-tool integrations, first ROI measurement
- Phase 3 "Scale" (Days 61-90): Deploy Tier-3 agents, executive P&L review, document operating procedures
Tone: Professional, concise, action-oriented. Use bullet points for findings and the plan. Open with a compelling one-sentence hook about the scaling challenge. Close with a clear call to action for leadership.
What does this prompt cover?
The prompt feeds Claude the four pillars of your roadmap: the business case for transformation, the top 3 bottlenecks from your assessment, the 10-agent opportunity with projected time savings, and the 3-phase plan.
The data points (42+ hours/week, 10 agents, 90 days) come directly from the work you built in Steps 3 through 5. This is why you did the analysis first. The executive summary is a synthesis layer, not new research.
- Review Claude's output. Check that it covers all four sections and stays under one page in length.
- If the output feels too long or too generic, ask Claude a follow-up: "Make this more concise and add specific numbers from the data I provided."
- Copy the final executive summary text from Claude.
- Return to your Google Doc titled "PartnersFirst Media - AI Transformation Roadmap."
- Scroll to the very top of the document, above your company profile section.
- Add a new heading: Executive Summary.
- Paste Claude's executive summary text below the heading.
Your document now opens with the leadership-facing narrative. Anyone reading the roadmap sees the business case, key findings, and the plan before diving into the details.
Position the summary as page one
In real transformation projects, the executive summary is always the first thing in the document. Decision-makers may never read past this page. Everything below it serves as supporting evidence for the claims made here.
Add the Cost & Resource Requirements section
Leadership always asks "what does this cost?" before approving a plan. A dedicated cost section removes ambiguity and shows you have thought through the financial implications.
- In your Google Doc, scroll to the end of your executive summary section.
- Add a subheading: Cost & Resource Requirements.
- Format this section as a callout box or bordered text block in Google Docs (select the text, then click Format > Borders and shading to add a visible border).
Inside the callout box, add three categories of costs:
- Under Existing Tools ($0 incremental), list: Google Workspace (Business Standard, already licensed), Slack (Pro plan, already licensed), HubSpot (Free CRM, no upgrade required).
- Under AI Platform Costs, list: Gemini (covered by existing Google AI Pro plan at $19.99/month per user, already active), Notion AI Custom Agents (requires upgrade to Business plan, credits billed as add-on).
- Under Team Time Investment, add estimated hours per week by role for each phase. For example: Head of Ops (8 hrs/week in Phase 1, 4 hrs/week in Phases 2-3), AI Champion (10 hrs/week across all phases), Marketing Lead (3 hrs/week in Phase 2), Partner Manager (3 hrs/week in Phase 2).
Why include team time as a cost?
Dollars spent on tools are only part of the investment. Leadership cares equally about how much employee bandwidth the project will consume. Showing hours by role by phase demonstrates that you have planned for realistic capacity allocation, not just tool procurement.
- Below the cost categories, add a final line with the heading Recommended Next Step.
- Write one sentence calling for executive sponsor sign-off and a Week 1 kickoff meeting. For example: "We recommend VP-level sponsor approval by [date] to begin the Week 1 kickoff with the core team."
What is a Recommended Next Step?
Every strategy document should close with a clear, single action the reader is expected to take. This removes ambiguity about who does what after reading. In transformation projects, that action is almost always "approve the plan and schedule kickoff."
- Review your complete executive summary section, including the Cost & Resource Requirements box and the Recommended Next Step.
- Confirm that the callout box is visually distinct from the rest of the document (bordered or shaded).
Your Google Doc is now a complete, portfolio-ready AI transformation roadmap. It opens with a leadership summary and closes with a clear call to action.
Executive summary feels too long or too vague?
If Claude produced more than one page, ask it: "Cut this to 400 words maximum. Keep all numbers." If the summary feels generic, check that you included the specific data points (42+ hours/week, 10 agents, 3 phases, 90 days) in your original prompt.
You can also use Gemini in Google Docs to refine the pasted text. Click the Gemini icon in the side panel and ask it to tighten the language or add specificity.
Still not looking right? Help me refine my executive summary so it stays under one page with specific metrics.
Secret mission
Create a "Week 1 Quick Wins" Playbook
Your 90-day roadmap is strategic and thorough. But leadership always asks the same question: "What can we do RIGHT NOW?" Design 3 immediately-deployable AI automations that PartnersFirst Media could implement on Day 1 using the tools they already have. Each Quick Win includes a copy-paste-ready prompt the team can use immediately.
Clean Up Your Resources
Clean Up Your Resources
Decide whether to keep your resources running, pause them to come back later, or delete them entirely. This project runs entirely on free tools and your existing Google AI Pro plan, so there are no ongoing costs to worry about.
Resources you used:
- Google Doc: "PartnersFirst Media - AI Transformation Roadmap" (stored in Google Drive).
- Notion page: "AI Transformation Project" containing the AI Agent Opportunities database.
- HubSpot free CRM account (used for interface reference, no charges or expiration).
Keep everything running
No action needed. Choose this if you want to use your roadmap as a portfolio piece or reusable template for real companies.
- Your Google Doc stays in Google Drive at no extra cost beyond your existing Google AI Pro plan.
- Your Notion page and database remain on the free plan with no charges.
- Your HubSpot free CRM account has no expiration and no fees. It stays active indefinitely.
- Consider sharing your Google Doc with collaborators or potential employers to demonstrate AI strategy skills.
Pause - I'll come back to this later
This project is entirely document-based with no running processes. There is nothing to shut down or pause.
- Your Google Doc, Notion database, and HubSpot account all remain available at no cost whenever you return.
- If you want to reduce clutter, move your Google Doc into a dedicated folder in Google Drive.
- In Notion, you can archive the "AI Transformation Project" page by clicking the three-dot menu on the page and selecting archive. This hides it from your sidebar without deleting it.
Delete - I don't want to use this again
Remove all project resources if you no longer need them. Since everything is free, there is no financial reason to delete. Only do this if you want a clean slate.
- Delete your Google Doc by opening Google Drive, right-clicking "PartnersFirst Media - AI Transformation Roadmap," and selecting **Move to trash**.
- Delete your Notion page by opening "AI Transformation Project," clicking the three-dot menu at the top right, and selecting **Move to Trash**. This removes the page and the AI Agent Opportunities database inside it.
- Your HubSpot free CRM account has no charges and no expiration. If you want to delete it, navigate to your account settings in HubSpot and look for the account deletion option under account management.
That's a Wrap!
That's a Wrap!
You just built the kind of AI transformation roadmap that professional consultants deliver to real companies. This is a portfolio-ready strategic deliverable that demonstrates your ability to assess operations, prioritize AI use cases, and design phased implementation plans for a B2B organization.
You've learned how to:
- Assess 8 core operational workflows for AI readiness using a structured framework covering pain points, manual process counts, error rates, and data flows across multiple tools.
- Design and score 10 AI agent opportunities using a weighted prioritization matrix with criteria for Business Value, Data Readiness, Technical Feasibility, and Time to Value.
- Build a phased 90-day implementation roadmap with governance scaffolding, change management protocols, risk classification, and rollback plans for each deployed agent.
- Create a "Week 1 Quick Wins" playbook with 3 immediately-deployable AI automations including copy-paste-ready prompts and measurable success metrics.
Ready to quiz yourself?