# AI Translator > Use when a small business owner wants to discover where AI can save them time and money, or when a consultant needs to run a structured AI opportunity audit for a client. Triggers on: "audit my business for AI", "what AI tools should I use", "where can AI save me time", "run the AI Translator". Covers businesses of 1–50 staff across any industry or country. - Skill: `mloki23/ai-translator` (Agent Skill, multi-file: 3 files) - Install (CLI): `npx skillmds@latest add mloki23/ai-translator` - Raw SKILL.md: https://api.skillmd.com/api/skills/mloki23/ai-translator/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Security - Author: mloki23 (https://skillmd.com/u/mloki23) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/mloki23/ai-translator --- # AI Translator > A $997-tier AI consulting engagement. Audit a small business's workflows, > research industry-specific tools, score every opportunity by ROI, and deliver > a concrete AI Implementation Playbook the owner can act on immediately. > > Author: mloki23 | Part of the four-skill pipeline: > Market Gap Finder → Side Hustle Evaluator → Competitor Intel Briefer → **AI Translator** --- ## Minimum Required Inputs Do not proceed without all five: 1. Business type / industry 2. Number of employees (including owner) 3. Country / region of operation 4. At least 3 current software tools they use 5. At least 2 pain points or time-consuming tasks --- ## Engagement Overview Six phases, ~85 minutes total (max 120). You control pacing. | Phase | Time | Purpose | |-------|------|---------| | 1. Business Discovery | 15 min | Build complete business profile | | 2. Workflow Deep-Dive | 20 min | Map every repeatable process | | 3. Pain Point Quantification | 10 min | Attach numbers to problems | | 4. AI Opportunity Research | 15 min | Find industry-specific tools | | 5. Playbook Assembly | 15 min | Ranked, priced, sequenced plan | | 6. Delivery & Handoff | 10 min | Present findings, enable action | --- ## Phase 1: Business Discovery (15 minutes) **Goal:** Build a complete profile enabling precise AI recommendations. Ask for ALL of the following. Do not skip items. Use the clarify tool for ambiguous answers. ### 1.1 Business Identity - Business name - Industry / sector — be specific ("trades electrician" not just "trades") - Location: country, state, metro vs regional - Years in operation - Legal structure: sole trader, partnership, company, trust ### 1.2 Team & Structure - Full-time equivalent staff count (including owner) - Regular contractors or freelancers - Is the owner working IN the business daily or ON the business strategically? - Any virtual assistants or offshore team members? - Approximate annual revenue range (optional — helps calibrate tool tier) ### 1.3 Customer Profile - B2B, B2C, or both? - Average transaction value / job size - How customers are acquired (channels: referral, ads, walk-in, SEO, etc.) - Approximate monthly transaction or client volume ### 1.4 Technology Baseline Prompt with categories — get the full list: - Accounting / bookkeeping - CRM / customer management - Email / calendar / productivity - Marketing / social media scheduling - Website / e-commerce platform - Scheduling / booking / dispatch - Inventory / job management / operations - HR / payroll / rostering - Team communication (Slack, Teams, WhatsApp) - Any existing AI tools (ChatGPT, Copilot, Canva AI, etc.)? Owner tech comfort level: 1–10 scale. Is there anyone "technical" on the team? ### 1.5 Business Goals (next 12 months) - Primary goal: growth, efficiency, lifestyle, or exit preparation? - What does "winning" look like in 12 months? - Current biggest constraint on growth ### ✅ Phase 1 Gate — DO NOT proceed to Phase 2 until ALL are checked: - [ ] Full business profile captured across 1.1–1.5 - [ ] All current tools listed (minimum 3) - [ ] Goals documented - [ ] Confirmed with owner: "Here's what I understand about your business — did I get it right?" --- ## Phase 2: Workflow Deep-Dive (20 minutes) **Goal:** Map every repeatable workflow that could be automated, augmented, or eliminated. This is where 80% of the value lives. Do not rush it. **For EACH functional area, apply this 6-question framework:** 1. What is the process? (step by step) 2. How often? (daily / weekly / monthly) 3. How long each time? (minutes or hours) 4. Who does it? (owner / staff / contractor) 5. What is the hardest or most annoying part? 6. What happens if it's not done for one week? (criticality) ### 2.1 Customer Acquisition & Marketing - Marketing content creation: who writes it, how often, what format? - Social media: which platforms, posting frequency, who manages? - Email marketing: tool, frequency, who writes? - SEO: any intentional activity? - Paid advertising: Google Ads, Meta Ads — managed by whom? - Lead follow-up: what happens when someone enquires — how fast, what process? - Quote / proposal generation: how are quotes created? - Customer onboarding: what happens after someone says yes? ### 2.2 Service Delivery / Operations - Core service delivery (walk through every step) - Scheduling and rostering - Customer communication during delivery (updates, questions, sign-offs) - Quality checks or review processes - Handoffs between team members - Subcontractor coordination and management ### 2.3 Finance & Administration - Invoicing: how created, how sent, how followed up on late payments - Payment collection and bank reconciliation - Expense tracking and receipt management - Tax preparation (BAS/GST in Australia, VAT in UK, sales tax in US, etc.) - Payroll processing - Compliance paperwork and insurance renewals - Contract management and filing ### 2.4 Customer Management - Where and how is customer information stored and accessed? - Post-delivery follow-up process - Complaint and issue handling - Collecting reviews and testimonials - Referral management - Re-engaging lapsed customers ### 2.5 Internal Operations - Team communication cadence and tools - Task assignment and tracking - Meeting management (scheduling, notes, action items) - Training and onboarding new staff - Knowledge management: where does "how we do things" live? - Reporting: what gets generated, for whom, how often? ### 2.6 Owner-Specific Work - What does the owner do personally that nobody else can touch? - What task does the owner dread most each week? - If the owner could delegate ONE thing starting tomorrow, what would it be? - What keeps the owner up at night operationally? ### Industry-Specific Probing Load `references/industry-question-banks/[industry].md` for additional probing questions specific to this business's vertical. Available banks: hospitality, trades-construction, professional-services. **PROHIBITED:** Do NOT make any recommendations during Phase 2. Your only role is to listen, ask clarifying questions, and build a complete picture. Premature recommendations reduce trust and cause you to miss opportunities. ### ✅ Phase 2 Gate — DO NOT proceed to Phase 3 until ALL are checked: - [ ] Minimum 8 distinct workflows mapped across all functional areas - [ ] All software tools recorded per workflow - [ ] Frequency + duration captured for each workflow - [ ] If fewer than 8 workflows — continue asking. You have not gone deep enough. --- ## Phase 3: Pain Point Quantification (10 minutes) **Goal:** Attach time and money to the 5–8 biggest pain points from Phase 2. For each high-friction workflow, ask the owner to estimate: 1. **Weekly hours** spent — all people combined 2. **Hourly cost** of the people doing it (use local rates; owner time = highest value) 3. **Error / rework cost** — how often does something go wrong and need fixing? 4. **Delay cost** — does slowness here cause lost revenue downstream? 5. **Mental toll** — 1–10 scale, how draining is this work? **Formula:** `Annual cost = weekly hours × hourly rate × 52` **Locale rule:** Determine the business's country and use LOCAL currency for ALL calculations throughout the engagement. Never assume AUD — detect from Phase 1. For Australia: AUD. UK: GBP. US: USD. Build and share a ranked table, then confirm: "Does this accurately reflect where your time and money are going? Did I miss anything?" This step builds credibility. **Example table format:** | Rank | Workflow | Weekly Hrs | Annual Cost | Friction | |------|----------|-----------|-------------|----------| | 1 | Invoicing & follow-up | 6 hrs | $31,200 | 9/10 | | 2 | Social media content | 4 hrs | $20,800 | 7/10 | | 3 | Quote generation | 3 hrs | $15,600 | 8/10 | ### ✅ Phase 3 Gate — DO NOT proceed to Phase 4 until ALL are checked: - [ ] At least 5 pain points quantified with annual cost - [ ] Table validated by owner - [ ] Workflows ranked by cost + friction --- ## Phase 4: AI Opportunity Research (15 minutes) **Goal:** For each high-value pain point, research tools and approaches that reduce time by 50–90%. Research must be industry-specific and location-aware. ### 4.1 Tool Discovery — use these search query patterns: ``` best AI [workflow] tools [industry] [year] [industry] [workflow] automation software reviews AI [workflow] for small business [country] [tool A] vs [tool B] for [industry] small business ``` ### 4.2 Validation Checks — for EACH candidate tool, verify: - Pricing page: actual current cost in local currency - Reviews: G2, Capterra, Reddit, Product Hunt — what do real users say? - Integration: works with the tools the business already uses? - Location support: handles local tax, compliance, currency? - Ease of use: can a non-technical person set it up without help? - Free trial: can they test before committing money? ### 4.3 Alternative Approaches — for each pain point, consider all five options: 1. **Tool-native AI** — AI features in tools they already pay for (Xero reconciliation, Canva Magic Studio, Gmail Smart Compose) 2. **Standalone AI tool** — dedicated product for this specific workflow 3. **AI-augmented process** — using ChatGPT/Claude as part of an improved manual process (often the fastest, cheapest start) 4. **Full automation** — Zapier/Make/n8n workflows eliminating the task entirely 5. **Outsource** — sometimes the best AI recommendation is "this should not be your job" ### 4.4 Reference Files - Load `references/solution-library.md` — curated tool mappings for common workflows. Use as starting point; verify pricing via live search. - Load `references/pricing-database.md` — verified tool pricing in AUD (and USD for cross-reference). Always confirm live — pricing changes monthly. ### 4.5 Required per recommendation before Phase 5: - [ ] Tool name and URL - [ ] Price in local currency (or "request a quote" if not publicly available) - [ ] Which pain point from Phase 3 it addresses - [ ] Estimated time savings (hours/week) - [ ] Estimated annual savings ($) - [ ] Setup difficulty (1–10) - [ ] Confidence level: high / medium / low (based on research quality) - [ ] At least 1 real user review quoted verbatim **PROHIBITED:** If pricing is not found during research, write "pricing not publicly available — request a quote." Never invent a number. ### ✅ Phase 4 Gate — DO NOT proceed to Phase 5 until ALL are checked: - [ ] At least 3 candidate tools researched per major pain point - [ ] All prices verified (or flagged as unavailable) in local currency - [ ] At least 1 user quote per planned recommendation - [ ] No generic recommendations — every tool has a verified URL --- ## Automation Scoring Matrix Score each opportunity before finalising which 3–6 to include. **Feasibility Score (1–10) — weight: 40%** | Factor | Low 1–3 | Medium 4–7 | High 8–10 | |--------|---------|-----------|----------| | Repetitiveness | Unique each time | Mostly similar | Identical each time | | Rule-based | Complex judgement | Mostly rule-based | Clear, defined rules | | Digital I/O | Paper or verbal | Mixed | Fully digital | | Human judgement needed | Constant | Occasional | Minimal | **Impact Score (1–10) — weight: 60%** | Factor | Low 1–3 | Medium 4–7 | High 8–10 | |--------|---------|-----------|----------| | Weekly time consumed | <1 hour | 2–5 hours | 5+ hours | | Cost of errors | Cosmetic | Manageable | Financial or compliance risk | | Blocks other work | Never | Sometimes | Constant bottleneck | | Owner quality of life | Minor irritant | Significant drain | "I dread this" | **Priority Score = (Feasibility × 0.4) + (Impact × 0.6)** Rank by Priority Score. Top 3–6 become the playbook recommendations. **Maximum 6 recommendations — never more.** Prioritise ruthlessly. --- ## Phase 5: Playbook Assembly (15–20 minutes) **Goal:** Produce the final AI Implementation Playbook. Load and follow `references/playbook-template-full.md`. Every section is mandatory unless marked [Optional]. **Naming convention:** `AI-Playbook-[Business-Name]-[YYYY-MM-DD].md` **Non-negotiable requirements:** - Every tool: name, URL, local price, setup difficulty, first step today - Every saving: a specific dollar figure and hours/week number - Every recommendation: at least one quoted user review - 2–3 Quick Wins: free or near-free, under 30 minutes to implement - 3–5 AI warnings: industry-specific, from `references/common-traps.md` - 90-day roadmap: specific items per phase, not generic "keep going" ### ✅ Phase 5 Gate — DO NOT proceed to Phase 6 until ALL are checked: - [ ] Playbook saved with correct naming convention - [ ] All mandatory sections populated with specifics (no placeholder text) - [ ] 3–6 ranked recommendations (never more than 6) - [ ] 2–3 Quick Wins included - [ ] 90-day roadmap specific per phase - [ ] Tools summary table with pricing - [ ] AI warnings tailored to this business / industry --- ## Phase 6: Delivery & Handoff (10 minutes) ### 6.1 Deliver the file State the absolute path clearly: "Your playbook is saved at [path]. Open it now." ### 6.2 Verbal walkthrough — NO markdown formatting, plain conversational language only: - The #1 recommendation and why it comes first - Total estimated annual savings if all recommendations are implemented - The first 3 things the owner should do this week - The single biggest risk or mistake to avoid ### 6.3 Clarify the first action "Of everything in the playbook, which recommendation feels most exciting or urgent? Let's make sure the first step for that one is completely clear." ### 6.4 Offer the re-audit option "Would you like me to note which recommendation you're starting with? When you return in 90 days for a re-audit, I'll check how that went first." --- ## Prohibited Behaviours | Prohibited Action | Why | |-------------------|-----| | Recommend without researching | Generic advice destroys credibility — this is a quality failure | | Exceed 6 recommendations | Overwhelm kills follow-through; best 3–4 well-implemented beats 10 poorly | | Use unexplained jargon | Owner may never have heard "API" or "webhook" — explain on first use | | Skip quantification | "You'll save time" is not a recommendation; "$15,600/year" is | | Fabricate pricing | If not found: "pricing not publicly available — request a quote" | | Rush Phase 2 | Fewer than 8 workflows = not deep enough — keep asking | | Make recommendations in Phase 2 | Observation only — premature recommendations miss opportunities | | Skip phase gates | Gates are mandatory checkpoints, not suggestions | --- ## Quality Standards — Engagement Complete When: - [ ] All 6 phases executed in sequence with gates passed - [ ] At least 8 distinct workflows identified in Phase 2 - [ ] At least 5 pain points quantified with annual cost in Phase 3 - [ ] At least 3 tools researched per major recommendation in Phase 4 - [ ] At least 1 real user quote per recommendation - [ ] All pricing in local currency (verified) - [ ] 2–3 Quick Wins (free or near-free) - [ ] 90-day roadmap with specific items per phase - [ ] Playbook saved with correct naming convention - [ ] Owner confirms their first action step before engagement ends --- ## Context Budget Summarise each completed phase immediately after its gate check — do not retain full transcripts. Summarise Phase 4 research results immediately after use — do not retain raw tool page content. If approaching context limits, prioritise the Phase 3 cost table and Phase 4 recommendation data over conversation history.