Every term. Plain English. No jargon.

McKinsey framework: 70% of value from processes/data/people, 20% from integration, 10% from AI technology. You build the 70% by deploying, not before.

Book context: Ch. 5

An AI system that doesn't just respond to prompts but takes autonomous or semi-autonomous actions - browsing the web, sending emails, updating records.

Book context: Quick Reference Guide; all Plays

Software that performs tasks typically requiring human thinking - analyzing data, drafting content, making decisions based on patterns.

Book context: Quick Reference Guide; used throughout

The connection point that lets two pieces of software talk to each other. When AI tools 'integrate' with your CRM, an API makes it possible.

Book context: Quick Reference Guide; all Plays

A unique credential that identifies your account when making API calls. Like a password for software-to-software communication.

Book context: Play 1 - Step 1; all Plays

Money owed to your firm for services rendered. Play 6 automates follow-up on overdue AR.

Book context: Play 6

Running a process in the background without waiting for it to complete. Used to split long-running n8n workflows to avoid timeouts.

Book context: Appendix C

Software for managing job applications and hiring pipelines. Used as output destination in Play 10.

Book context: Play 10

An alternative data enrichment platform for trigger event monitoring.

Book context: Play 3 - Monitoring

An alternative voice AI platform for building phone agents.

Book context: Play 2 - Voice Track

The total cost to acquire one new customer. Referenced in context of margin trap and ROI calculations.

Book context: Ch. 2

Redirecting time freed by automation toward higher-value work (client relationships, complex engagements, business development).

Book context: Ch. 3; Ch. 8

The process of managing the human side of automation: communication, training, resistance handling, role evolution.

Book context: Ch. 3; Ch. 7

OpenAI's consumer product. Not the same as using the OpenAI API - ChatGPT is a conversational interface; the API is for building systems.

Book context: Quick Reference Guide

An AI model exceptional at writing, synthesis, nuance, and long-form generation. Recommended for client communications and meeting briefs.

Book context: Ch. 6; Appendix B

Anthropic's coding tool. Can generate chatbot interfaces, integration code, and other technical components from natural language descriptions.

Book context: Play 2 - Web Chat

A data enrichment platform that surfaces trigger events (job changes, funding, company updates) for contacts. Connects to n8n.

Book context: Play 3 - Monitoring

A probabilistic measure of how certain the AI is about its output. Used to determine whether output is auto-processed or routed to a human.

Book context: Ch. 5; all Plays

The minimum confidence score required for the AI to take an automated action. Below this, the item routes to the exception queue for human review. Start conservative (80%+); loosen over time as the system proves itself.

Book context: Ch. 5; Play 2

An automated morning email summary of account activity, flagged items, and overdue follow-ups delivered to each team member.

Book context: Play 1 - Step 5

Splitting large documents into smaller sections for indexing in a vector database. Each chunk gets its own embedding for search.

Book context: Play 11

An e-signature platform. Its completion webhook triggers onboarding automation in Play 5.

Book context: Play 5

A legal contract between your firm and an AI vendor that specifies how your data is handled, stored, and protected.

Book context: Security Chapter

A numerical representation of text that captures semantic meaning. Used to power vector search - similar meanings produce similar embeddings.

Book context: Play 11

A secure way to store sensitive information (like API keys) on your server, rather than putting them directly in workflow code.

Book context: Appendix A

Business management software that integrates financials, operations, and HR. Another potential integration point.

Book context: Ch. 6

A dedicated channel (Slack, Teams, or spreadsheet) where low-confidence AI outputs and workflow errors are routed for human review and action.

Book context: Ch. 5; all Plays

A retry strategy where wait time between retries increases exponentially. Used when hitting API rate limits.

Book context: Appendix C

A knowledge graph database optimized for fast context retrieval in chat-based AI.

Book context: Appendix B

An alternative meeting transcription tool.

Book context: Play 1 Prerequisites

The percentage of CRM fields that contain accurate, current data. Play 1 targets 95%+ from a typical starting point of 40-60%.

Book context: Play 1

Customizing a pre-trained AI model with your own data to improve performance on specific tasks. Not required for the Plays in this book.

Book context: General context

A meeting transcription tool that generates call transcripts and summaries. Integrates with n8n for CRM logging.

Book context: Play 1 Prerequisites

An open-source low-code tool for building LLM applications. Mentioned as a modern orchestration option.

Book context: Ch. 2

General Data Protection Regulation. EU data privacy law. Relevant for firms with European clients or operations.

Book context: Security Chapter

Google's AI model. Fast and inexpensive, especially the Flash version. Best for high-volume, lower-stakes tasks and web research.

Book context: Ch. 6; Appendix B

AI that creates new content (text, images, code) rather than simply analyzing existing data. LLMs are a type of generative AI.

Book context: Implied throughout

A revenue intelligence platform with call recording and transcription capabilities.

Book context: Play 1 Prerequisites

OpenAI's coding tool (ChatGPT). Can generate code for chatbot widgets and integrations.

Book context: Play 2 - Web Chat

Elon Musk's AI model from X (formerly Twitter). Best for personality in messaging and real-time news via X platform integration.

Book context: Ch. 6; Appendix B

When an AI generates specific claims - client names, case outcomes, dollar amounts - that are not in the source material. A known failure mode requiring human review.

Book context: Play 4; Play 9; Appendix C

Health Insurance Portability and Accountability Act. Governs health data privacy. Relevant for healthcare-adjacent consulting firms.

Book context: Security Chapter

Requiring human approval before an AI agent takes a specific action. Used for sensitive tasks (financial, client-facing) or when AI confidence is low.

Book context: Quick Reference Guide; Ch. 5

A standard way to send data to or request data from a web service. Used in n8n to connect to any CRM or tool via its API.

Book context: Play 1 - Step 2

Traditional phone menu systems ('Press 1 for sales...'). What modern voice AI like Retell replaces.

Book context: Play 2

A standard data format used by APIs. Looks like organized key-value pairs. AI models return data in JSON format for easy processing.

Book context: Play 1 - Step 2; all Plays

A database that understands relationships between entities (people, companies, deals) - not just content similarity. More advanced than vector search alone.

Book context: Play 11 Optional

A framework for building AI applications. More developer-oriented than n8n. Mentioned as a modern orchestration option.

Book context: Ch. 2

The time delay between input and response. In voice AI, latency under 600-800ms is critical for natural conversation.

Book context: Play 2 - Voice

The engine behind AI writing and reasoning tools. Trained on massive text data, it predicts the most useful response to a given input. Examples: GPT, Claude, Gemini.

Book context: Quick Reference Guide; Ch. 6

A hosted workflow automation alternative. Easier onboarding than n8n but per-task pricing and less flexibility.

Book context: Appendix B

A knowledge graph database optimized for real-time streaming analytics.

Book context: Appendix B

A system design where each component (orchestration, AI, CRM, communication) is independent and swappable without rebuilding.

Book context: Ch. 6

The simplest working version of a system. Build the MVP first, then iterate. Referenced in Play 11 for vector-only before adding knowledge graphs.

Book context: Play 11

An open-source workflow automation platform that connects your tools and AI models. The orchestration layer for all Plays in this book.

Book context: Quick Reference Guide; Ch. 6; all Plays

The branch of AI focused on understanding and generating human language. LLMs are the most advanced form of NLP.

Book context: General context

The market-leading knowledge graph database. Open-source. Used for understanding how contacts relate to opportunities, companies, etc.

Book context: Play 11; Appendix B

A web server tool used to set up a custom URL and SSL for your n8n instance. Can be configured via OpenCode.

Book context: Play 1 - Step 1

A single step in a workflow. Each node does one thing: trigger, transform data, call an API, run AI, send a message. Nodes connect to form workflows.

Book context: Play 1; all Plays

An authorization standard that lets you securely connect tools (like n8n to Gmail) without sharing your password. You 'authorize' access through a login flow.

Book context: Play 1 - Step 2; n8n setup

Software whose code is publicly available. Typically free, more flexible, and no vendor lock-in.

Book context: Quick Reference Guide; Ch. 6

The company behind GPT models and ChatGPT. Their enterprise API is used for extraction, classification, and structured output in this book.

Book context: Quick Ref; Ch. 6

A free tool that uses AI to configure servers and install software via natural language commands. Shortcut for n8n setup.

Book context: Play 1 - Step 1

Work the firm does because it's always done it that way, not because there's no alternative. Typically 20-30% of total labor cost.

Book context: Ch. 2

An alternative meeting transcription tool.

Book context: Play 1 Prerequisites

An alternative e-signature and document automation platform.

Book context: Play 5

Data that can identify a specific individual: name, SSN, email, phone number. Must be handled carefully in AI workflows.

Book context: Security Chapter

A managed vector database for enterprise-level AI search. Zero infrastructure management but higher cost.

Book context: Play 11; Appendix B

A method where n8n checks a source (like email) at regular intervals (e.g., every 5 minutes) for new data. Alternative to webhooks.

Book context: Play 1 - Step 2

The instruction or input you give to an AI model. Consists of a system message (role/rules) and a user message (specific task/data).

Book context: Play 1 - Step 2; all Plays

Software for managing projects, resources, and billing in services firms. An integration point for the Plays.

Book context: Ch. 6

A flexible open-source vector database for teams with some technical aptitude.

Book context: Appendix B

Deploying AI infrastructure so clients (and sometimes staff) don't know it's there. The firm appears more responsive and organized without visible AI.

Book context: Ch. 4

Restrictions on how many API calls you can make per minute/hour. AI model APIs have rate limits. Workflows need retry logic to handle this.

Book context: Appendix C

The most common type of API. Uses standard web addresses (URLs) to send and receive data. Every modern CRM has one.

Book context: Play 1 - Step 2

A voice AI platform for building phone agents that handle inbound calls, qualify callers, and book meetings. Primary voice recommendation.

Book context: Play 2 - Voice Track

A standard format for subscribing to website updates. Used in n8n to monitor company news and press releases for trigger events.

Book context: Play 3

Running software on your own server (that you control) instead of using a vendor's cloud. Recommended for n8n in this book for data control and cost.

Book context: Play 1 - Step 1; Ch. 6

Search based on meaning rather than exact keyword matching. Powered by embeddings and vector databases.

Book context: Play 11

AI's ability to classify the emotional tone of text (positive, neutral, negative / needs attention). Used in email logging and daily digests.

Book context: Play 1 - Step 2

A security compliance standard that verifies a vendor handles customer data securely. Table-stakes requirement for AI vendors.

Book context: Security Chapter

The time between a lead inquiry and first contact. The single highest-leverage variable in converting inbound to meetings.

Book context: Play 2

Encryption that protects data in transit (the 'https' in web addresses). Required for secure n8n access.

Book context: Play 1 - Step 1

An open-source database platform with built-in vector search capabilities. Recommended for most firms in Play 11.

Book context: Play 11; Appendix B

An alternative voice AI platform for building phone agents.

Book context: Play 2 - Voice Track

The instruction given to an AI model that defines its role, rules, and output format. Stays consistent across interactions.

Book context: Play 1 prompts; all Plays

The point where a services firm can no longer grow without adding headcount - because human-judgment work hasn't been separated from automatable work.

Book context: Ch. 1

The structural problem where revenue growth in professional services leads to proportional headcount growth, keeping margins flat.

Book context: Ch. 2

The basic unit of text that AI models process. Roughly 1 token = 0.75 words. Relevant for API pricing and context limits.

Book context: General; pricing context

The massive text datasets used to build an LLM. Enterprise AI APIs explicitly do not train on your data - this is a key vendor question.

Book context: Security Chapter

The event that starts a workflow. Can be time-based (schedule), event-based (new email, webhook), or manual.

Book context: Ch. 4; all Plays

A change in a prospect's situation that creates a legitimate reason to re-engage: leadership change, funding, hiring, contract renewal, regulatory shift.

Book context: Play 3

A communications platform for sending SMS messages programmatically. Used in Play 8 for partner escalation notifications.

Book context: Play 8

A popular Linux operating system used for running servers. n8n self-hosting requires an Ubuntu server.

Book context: Play 1 - Step 1

The specific data or question passed to the AI for a given task. In n8n, this is often the dynamic input (e.g., email content).

Book context: Play 1 prompts; all Plays

A database designed to store and search embeddings (numerical representations of text). Powers semantic search for Knowledge Base Q&A.

Book context: Play 11; Appendix B

A defined list of high-value clients (top 20 accounts) who receive emergency response treatment. Must be explicitly defined in the CRM.

Book context: Play 8

A remote computer you rent to run software like n8n. Providers: DigitalOcean, GCP, Azure. Costs $12-18/month.

Book context: Play 1 - Step 1

A way for one application to send real-time data to another when an event happens. Example: DocuSign sends a signal to n8n the moment a contract is signed.

Book context: Play 2; Play 5

A repository of past successful proposals organized by scope type, client profile, case summaries, and methodology. Feeds the RFP draft generator.

Book context: Play 4

A connected sequence of nodes in n8n that automates a process from trigger to output. Each Play contains one or more workflows.

Book context: All Plays

A popular workflow automation tool. Fine for simple automations but expensive at scale, less customizable than n8n.

Book context: Ch. 6; Appendix B