← Back to homepage

Practical Guides for AI at work

We share practical, step-by-step guides for using AI in recurring business tasks. These are not client case studies, but blueprints detailing prompt structure, safety rules, and human control.

Practical guide blueprints detailing input structures, processing steps, prompt logic templates, and verification checks.

START HERE

Process first. Then AI.

Before we write the first prompt, we need to understand how the work actually happens. Only then can we define the roles of people, AI and automation.

KEYWORD EXPLAINED

PROMPT = an instruction for AI: what to do, what to use, and what a good result looks like.

  1. 1 Trigger
  2. 2 Decision
  3. 3 Exceptions
  4. 4 Correct outcome

Guide 01

JMHZ and Czech payroll reporting — AI as a control assistant

Focus: Using AI as a preparation and control assistant for payroll processes.

Core concept: A human payroll specialist remains in control of the final output, using AI only to check inputs, format documents, and double-check logic.

Safety: No financial, tax, or legal decisions are made by AI. No raw payroll data or client names are shared without complete anonymization.

Open Guide

Guide 02

From client chat to clear brief

Focus: Turning raw client communications (chat history, transcripts, emails) into a structured brief.

Core concept: Help structure requirements, extract key client objectives, highlight constraints, and flag ambiguous items needing clarification.

Safety: Make sure chat records and customer names are anonymized before pasting them into any public AI tools.

Open Guide

Guide 03

From inquiry to next step

Focus: Initial filtering of inbound client inquiries to draft project scopes and action plans.

Core concept: Extract the prospect's main needs, evaluate them against your business criteria, identify missing details, and draft an initial response outline.

Safety: Keep client details anonymized. Final commercial reviews and outreach must be manually checked and sent by a human.

Open Guide

Need help trying this in your own work?

I can help you adapt a guide to your real process, test it safely on anonymized examples, and set it up together in person or over Zoom.

Need help with implementation Explore Services