Value Matrix — what to automate first
Score every use case on time saved per week × confidence the AI does it well. Start at the top of this list.
Quick Wins — High Value, Low Effort
Do first -- these build momentum and prove ROI fast
- Email Drafting & Summarisation — AI drafts replies, summarises long threads, and extracts action items from your inbox. _Tools:_ M365 Copilot, Gemini, ChatGPT. _Impact:_ ~1 hour/day saved per user. _Effort:_ Deploy existing tool. _ROI:_ 13 hrs/week saved per marketer (CoSchedule 2025).
- Meeting Summaries & Action Items — Automatic transcription, notes, follow-ups, and action item tracking from meetings. _Tools:_ M365 Copilot (Teams), Gemini (Meet). _Impact:_ Eliminates manual note-taking. _Effort:_ Configuration only. _ROI:_ 30 min saved per meeting (McKinsey 2024).
- Document First-Draft Generation — Generate memos, reports, proposals, and briefs from a short prompt or outline. _Tools:_ M365 Copilot (Word), Claude, ChatGPT. _Impact:_ Drafting from 8 hours to 15 min. _Effort:_ Prompt training. _ROI:_ 87% time savings on document writing (Anthropic 2025).
- Code Autocompletion & Generation — Real-time code suggestions, explanations, test generation, and bug detection in your IDE. _Tools:_ GitHub Copilot, Claude Code. _Impact:_ 30-55% faster coding. _Effort:_ License + IDE plugin. _ROI:_ 20-45% increased dev productivity (McKinsey 2024).
- Knowledge Base Q&A — Ask questions about your company docs, policies, and procedures in natural language. _Tools:_ Perplexity Enterprise, Copilot Studio. _Impact:_ 50% reduction in research time. _Effort:_ Upload documents. _ROI:_ 50% reduction in research time (Industry avg.).
- Competitive Research — Real-time sourced intelligence on competitors, market trends, and industry news with citations. _Tools:_ Perplexity. _Impact:_ Hours of research in minutes. _Effort:_ Subscription only. _ROI:_ Hours compressed to minutes (OpenAI 2025).
- Slide Deck Creation from Outlines — Generate presentation slides from bullet points, documents, or verbal briefs. _Tools:_ M365 Copilot (PowerPoint), Gemini (Slides). _Impact:_ Hours to minutes. _Effort:_ Existing M365/Google license. _ROI:_ 25.1% faster task completion (Harvard/BCG 2024).
- Policy / FAQ Chatbot — Tier-1 support bot answering common employee or customer questions from your knowledge base. _Tools:_ Copilot Studio, ChatGPT. _Impact:_ 40-60% ticket deflection. _Effort:_ Low-code setup. _ROI:_ 30% avg. support cost reduction (Freshworks 2025).
- Status Report Generation — Automated weekly/monthly reports compiled from raw data, notes, and project updates. _Tools:_ M365 Copilot, ChatGPT, Claude. _Impact:_ Consistent reports in minutes. _Effort:_ Template setup. _ROI:_ 40%+ higher quality output (Harvard/BCG 2024).
- Contract & Document Review — AI reads long documents, highlights risks, extracts key terms, and summarises obligations. _Tools:_ Claude, Perplexity. _Impact:_ Hours of review in minutes. _Effort:_ Upload + prompt. _ROI:_ Handles 500+ page documents (Anthropic 2025).
Major Projects — High Value, High Effort
Plan and resource properly -- transformative but need commitment
- Enterprise Agentic Workflow Automation — Multi-step AI agents that handle entire processes: procurement, approvals, onboarding pipelines. _Tools:_ Copilot Studio, Custom agents. _Impact:_ Process transformation -- minutes vs. days. _Effort:_ Integration + governance. _ROI:_ 307% ROI in under 18 months (HyperSense 2025).
- AI-Powered Customer Service Transformation — 24/7 multilingual omnichannel support with WhatsApp, chat, email, and phone agents. _Tools:_ Copilot Studio, ElevenLabs, Trillet.ai. _Impact:_ 24/7 coverage, 60% cost reduction. _Effort:_ Multi-system integration. _ROI:_ Up to 68% cost reduction in 3 months (NexGen Cloud 2025).
- AI Security Operations Centre — AI-powered threat detection, incident response, and security monitoring across your stack. _Tools:_ Security Copilot, Custom agents. _Impact:_ Incident response in minutes. _Effort:_ Infrastructure + training. _ROI:_ IT ops: 52% automation ROI (HyperSense 2025).
- Autonomous Code Review & Deployment — AI reviews PRs, generates tests, checks for vulnerabilities, and manages CI/CD pipelines. _Tools:_ GitHub Copilot, Claude Code. _Impact:_ Faster releases, fewer bugs. _Effort:_ DevOps pipeline redesign. _ROI:_ Reusable code: 30-50% faster dev (McKinsey 2024).
- Enterprise Knowledge Graph — Build a queryable knowledge graph from all company documentation, emails, and chat history. _Tools:_ Perplexity Enterprise, Custom RAG. _Impact:_ Institutional knowledge preservation. _Effort:_ Taxonomy + data mapping.
- AI-Powered Financial Forecasting — Predictive models for revenue, cash flow, variance analysis, and scenario planning. _Tools:_ M365 Copilot for Finance, Claude. _Impact:_ Faster close, better accuracy. _Effort:_ Data warehouse + modelling. _ROI:_ 80% faster financial analysis (Anthropic 2025).
Fill-ins — Low Value, Low Effort
Do when capacity allows -- nice extras, not priorities
- Calendar Scheduling Optimisation — AI suggests optimal meeting times, identifies conflicts, and manages scheduling. _Tools:_ M365 Copilot, Gemini. _Impact:_ Minor time savings. _Effort:_ Already bundled.
- Grammar & Tone Checking — Polish communications with grammar correction, tone adjustment, and style consistency. _Tools:_ Gemini, ChatGPT, Claude. _Impact:_ Polished communications. _Effort:_ Built-in.
- Social Media Post Drafting — First drafts of social media posts for review and refinement. _Tools:_ ChatGPT, Gemini. _Impact:_ First drafts for review. _Effort:_ Prompt training.
- Data Visualisation Narration — AI explains existing charts, generates narratives from data, and adds context to dashboards. _Tools:_ M365 Copilot (Excel). _Impact:_ Explanation of existing charts. _Effort:_ Configuration.
Avoid — Low Value, High Effort
Skip these -- high effort with little return
- Building Custom LLMs from Scratch — Training your own large language model instead of using existing commercial models. _Tools:_ n/a. _Impact:_ Massive cost, existing models are superior. _Effort:_ Months of work, millions in compute.
- AI for Low-Volume Specialised Decisions — Using AI for rare, highly contextual decisions better served by domain experts. _Tools:_ n/a. _Impact:_ Insufficient data, high customisation cost. _Effort:_ Extensive custom work.
- Replacing Proven Deterministic Systems — Swapping working rule-based systems with probabilistic AI where certainty exists. _Tools:_ n/a. _Impact:_ Regression risk -- introducing uncertainty where none existed. _Effort:_ Replatforming cost.
AI workspaces compared
Microsoft 365 Copilot (Microsoft)
- Best for: M365-embedded productivity across Word, Excel, PowerPoint, Outlook, Teams
- Pricing: $30/user/month add-on (Copilot Chat free for all M365)
- Context window: Enterprise data graph
- Stack fit: microsoft
- Strengths: Native in Word, Excel, PowerPoint, Outlook, Teams; Enterprise data security -- stays in your tenant; Meeting summaries and action items in Teams
GitHub Copilot (Microsoft)
- Best for: Code autocompletion, generation, and explanation in IDEs
- Pricing: $10/user/month (Individual) | $19/user/month (Business)
- Context window: Full repository context
- Stack fit: any
- Strengths: Real-time code suggestions in VS Code, JetBrains, Neovim; Multi-file context awareness; Test generation and bug detection
Copilot Studio (Microsoft)
- Best for: Low-code custom AI agents for customer service, IT, and internal processes
- Pricing: $200/month per agent (includes 25K messages)
- Context window: Enterprise connectors
- Stack fit: microsoft
- Strengths: No-code/low-code agent builder; Native WhatsApp deployment; Computer use for legacy apps (preview)
Claude (Anthropic)
- Best for: Coding, complex analysis, long document processing, regulated industries
- Pricing: Free tier | Pro A$34/month (US$20) | Team US$25/user/month (min 5)
- Context window: 200K tokens (up to 1M)
- Stack fit: any
- Strengths: Lowest hallucination rates — 32% enterprise market share (2026); 77.2% SWE-bench coding score (industry-leading); Handles 500+ page documents
ChatGPT (OpenAI)
- Best for: Creative content, brainstorming, general-purpose AI assistance
- Pricing: Free | Plus $20/month | Team $25/user/month
- Context window: 400K tokens (GPT-5.1)
- Stack fit: any
- Strengths: Strongest at brainstorming and creative writing; Largest plugin/GPT marketplace; DALL-E image generation built in
Perplexity (Perplexity AI)
- Best for: Real-time research, fact-checking, competitive intelligence with citations
- Pricing: Free | Pro $20/month | Enterprise custom
- Context window: Live web access
- Stack fit: any
- Strengths: Every answer includes verifiable citations; Deep Research mode for comprehensive synthesis; Always current -- accesses live web
Gemini (Google)
- Best for: Google Workspace-native AI across Gmail, Docs, Sheets, Meet, Drive
- Pricing: Included in Google Workspace Business/Enterprise (no add-on since Jan 2025)
- Context window: 1M tokens (Gemini 2.0)
- Stack fit: google
- Strengths: Native in Gmail, Docs, Sheets, Slides, Meet; Included at no extra cost in Workspace plans; Automatic meeting notes and summaries in Meet
Microsoft Copilot ecosystem
- Microsoft 365 Copilot ($30/user/month add-on) — AI embedded in Word, Excel, PowerPoint, Outlook, and Teams. Uses GPT models for reasoning and writing within your M365 data.
- GitHub Copilot ($10-19/user/month) — AI code completion, generation, and explanation in your IDE. Multi-file awareness and memories that learn your coding standards.
- Copilot Studio ($200/month per agent) — Low-code platform for building custom AI agents. Deploy to WhatsApp, websites, Teams, and internal tools.
- Security Copilot (Included with M365 E5) — AI-powered security operations across Defender, Entra, Intune, and Purview. 77+ agents available.
- Copilot for Sales (Included with Dynamics 365 Sales) — Sales-specific AI that connects to Salesforce and Dynamics 365 for pipeline management and outreach.
- Copilot for Finance (Included with M365 Copilot + Finance module) — Finance-specific workflows in Excel with variance analysis, macroeconomic summarisation, and external data.
Department → tool mapping
| Department | Primary tool | Secondary tool | Key workflow | |---|---|---|---| | Engineering | Claude / GitHub Copilot | ChatGPT | Code generation, review, documentation | | Marketing | ChatGPT | Gemini / Copilot | Content creation, campaign ideation | | Legal / Compliance | Claude | Perplexity | Contract analysis, regulatory research | | Sales | M365 Copilot | Perplexity | Email drafting, CRM updates, competitive research | | Finance | M365 Copilot for Finance | Claude | Financial modelling, variance analysis | | Research / Strategy | Perplexity | Claude | Market research, due diligence | | IT / Operations | Copilot Studio | Security Copilot | Helpdesk automation, security ops | | HR | M365 Copilot | ChatGPT | Onboarding automation, policy Q&A | | Customer Service | Copilot Studio | Trillet.ai / ElevenLabs | Chatbots, ticket routing, voice AI | | Executive Leadership | M365 Copilot | Perplexity | Meeting summaries, strategic briefings |
Rollout phases
Phase 1: Discovery & Assessment
Identify high-impact use cases, assess data readiness, select tools
Activities:
- Audit current workflows for automation opportunities
- Map pain points to AI tool capabilities
- Complete the Value/Effort matrix with your team
- Assess data sensitivity and compliance requirements
- Evaluate current stack (Google / Microsoft / other)
Deliverables:
- Prioritised use case list with Quick Wins identified
- Data readiness assessment
- Tool selection recommendations
- Budget estimate for pilot phase
KPIs:
- Number of use cases identified
- Stakeholder alignment score
- Data readiness score
Phase 2: Pilot & Prove
Deploy Quick Wins with 2-3 champion teams, measure KPIs, gather feedback
Activities:
- Select 2-3 champion teams across different departments
- Deploy top 3 Quick Wins from the matrix
- Set up measurement baselines (time spent, error rates)
- Provide hands-on training for champion teams
- Weekly check-ins and feedback collection
Deliverables:
- Pilot results with measured time savings
- User satisfaction scores from champions
- Refined training materials
- Business case for scale phase
KPIs:
- Hours saved per user per week
- User adoption rate (%)
- Task accuracy improvement (%)
- User satisfaction (NPS)
Phase 3: Scale & Expand
Expand to additional departments, refine training, begin Major Projects
Activities:
- Roll out proven Quick Wins organisation-wide
- Begin planning top Major Projects from the matrix
- Establish AI governance policies and guidelines
- Create department-specific playbooks
- Set up ongoing training program
Deliverables:
- Organisation-wide deployment plan
- AI governance policy document
- Department playbooks
- Major Project roadmap with resourcing
KPIs:
- Organisation-wide adoption rate (%)
- Total hours saved per month
- Number of departments using AI daily
- ROI vs. tool subscription costs
Phase 4: Optimise & Transform
Full deployment, continuous improvement, advanced agentic use cases
Activities:
- Execute Major Projects from the Value/Effort matrix
- Implement agentic workflows for complex processes
- Establish AI centre of excellence or practice
- Redesign workflows around AI capabilities (the 80/20 rule)
- Continuous measurement and optimisation
Deliverables:
- Transformed workflows with measured impact
- AI centre of excellence charter
- Quarterly AI impact report
- Next-generation use case pipeline
KPIs:
- Revenue/cost impact from Major Projects
- Process cycle time reduction (%)
- Employee AI proficiency scores
- Innovation pipeline -- new use cases per quarter