Daily recap for June 4, 2026

Today we published 11 new step-by-step tutorials for small-business AI, covering legal intake, clinic scheduling, expense reporting, and more. Each one comes with a downloadable zip and a walkthrough you can follow to get it running.

New tutorials

Intake automation agent for solo immigration attorney

Automates client intake for a solo immigration attorney. A prospective client fills out a web form; the AI screens the inquiry, cross-references past cases for conflicts, classifies the legal need, and responds with empathy — all while scrubbing PII and adding compliance disclaimers. Intake time drops from 30 minutes to 5.

Read the tutorial → · Download the code (zip)

Built with agnostic LLM, lead-intake, 139 tests, 99.11% coverage.

Deposition prep summarizer for plaintiff litigation paralegal

A deposition transcript summarizer for plaintiff litigation paralegals. Feed it a deposition PDF — it extracts and chunks the text, stores it in memory, plans the context window to handle long transcripts, and produces a concise, citable summary using AI. The tool tracks token costs and supports evaluation and replay. What took hours now takes minutes.

Read the tutorial → · Download the code (zip)

Built with agnostic LLM, document-pipeline, 76 tests, 96.93% coverage.

Google Gemini voice agent for clinic appointment scheduling

A voice AI receptionist for medical and dental clinics. It answers phone calls via Twilio, transcribes speech with Deepgram, uses Google Gemini to understand and respond, and speaks back through Cartesia TTS. It connects to the clinic’s EHR to check availability, book appointments, and send SMS confirmations and reminders. Built to handle after-hours and peak-time call volumes, reducing the load on front-desk staff.

Read the tutorial → · Download the code (zip)

Built with Google Gemini, voice-agent, 155 tests, 97.70% coverage.

Mistral AI document pipeline for Xero expense report processing

A document pipeline that automates expense reporting. Drop in a receipt PDF or expense sheet — it extracts the data, uses Mistral to parse line items into JSON, and repairs malformed output automatically with a six-strategy repair engine. It enforces daily spend budgets and pushes validated expenses into Xero as invoices. No more manual data entry.

Read the tutorial → · Download the code (zip)

Built with Mistral, document-pipeline, 60 tests, 100% coverage.

Google Gemini runbook automation for PagerDuty SMB incidents

An incident runbook generator for small DevOps teams using PagerDuty. Run a CLI command — it pulls service and incident data from PagerDuty, uses Google Gemini to write recovery steps, and outputs a clean Markdown runbook. It skips duplicates if a similar runbook already exists. Schedule it to keep runbooks fresh as your services change.

Read the tutorial → · Download the code (zip)

Built with Google Gemini, reliability-ops, 73 tests, 95.90% coverage.

OpenAI lead intake for BigCommerce small business sales

A lead intake system for BigCommerce. When a contact form comes in, it classifies the inquiry’s intent and confidence. High-confidence sales questions get an instant AI reply, support requests route to a specialist, and low-confidence cases go to a human triage queue. It also sends email confirmations via SendGrid, notifies your team on Slack, and creates HubSpot tickets — all tracked with Langfuse.

Read the tutorial → · Download the code (zip)

Built with OpenAI, lead-intake, 79 tests, 96.74% coverage.

Cohere agent mesh for multi-channel customer support triage

A multi-channel support triage using Cohere. It ingests tickets from email, chat, and social media, classifies intent, and routes them to the right specialist agent — returns, billing, technical — based on confidence scores. Low-confidence tickets get human escalation. Per-channel cost tracking gives you visibility into where your AI spend goes.

Read the tutorial → · Download the code (zip)

Built with Cohere, agent-mesh, 99 tests, 95.04% coverage.

OpenAI guardrail layer for SMB customer chat safety

A security layer for OpenAI chatbots. It wraps the OpenAI SDK and adds configurable guardrails that detect prompt injection, redact PII using Microsoft Presidio, and filter toxic content. The result is a drop-in replacement for your existing client that screens both input and output without changing your application code.

Read the tutorial → · Download the code (zip)

Built with OpenAI, security-guardrails, 55 tests, 99.13% coverage.

OpenAI voice agent for Aircall small business support

A voice AI agent for Aircall phone numbers. It answers calls, handles FAQs (hours, pricing, services), and escalates to a human agent when it can’t resolve the issue. Built to reduce hold times and catch after-hours calls, keeping your business available around the clock.

Read the tutorial → · Download the code (zip)

Built with OpenAI, voice-agent, 70 tests, 97.40% coverage.

Anthropic document pipeline for Klaviyo SMB email campaigns

An email campaign generator for Klaviyo. Pull in your product catalog and customer segments; it uses Anthropic Claude to draft personalized email content, repairs the output to match your templates, enforces a daily budget to control costs, and pushes the campaign to Klaviyo for scheduling. No more hours spent copy-pasting.

Read the tutorial → · Download the code (zip)

Built with Anthropic, document-pipeline, 92 tests, 99.55% coverage.

Anthropic LLM observability for SMB AI operations

Observability instrumentation for Anthropic API usage. Add it to your app and get real-time dashboards of token spend, latency, error rates, and per-tenant budget enforcement. Exports to Langfuse or Phoenix with zero configuration. Designed for small businesses running AI-powered features who need to keep costs in check.

Read the tutorial → · Download the code (zip)

Built with Anthropic, observability, 85 tests, 100% coverage.

Browse the full catalog at reaatech.com/products.