Your Business Already Has an AI Strategy
The process documents, email templates and decision rules your team already wrote describe work software could do. The hard part of automating it, writing down how the business works, is done.
Every process document, email template and decision rule your team wrote describes work software could do. The hard part of automating it, writing down how your business works, is already done.
AI adoption is usually framed as a technology project: pick a platform, hire a data scientist, build a model. Start instead from what you already have: process documents, email templates, spreadsheets of decision rules. You wrote them so people would do the work the same way every time, and consistency is what software does well. They are a tested description of how your business runs, and the technology is the wiring.
Documents That Are Already Instructions
- Process documents become workflow automations
- Email templates become instructions for drafting replies
- Decision matrices become rules for sorting incoming work
- FAQ pages become answers a system gives from your own documents
- Onboarding checklists become automated sequences
None was written with AI in mind. Each was written so anyone on the team could pick up a task and do it the same way, which is the brief an automated system works to: follow the instructions, produce consistent output, do not improvise.
From Template to Instructions
Dear [Client], Thank you for reaching out about [topic]. I understand you're interested in [specific need]. Based on your situation, I'd recommend [recommendation]. The next step would be [action]. Please let me know if you have any questions. Best regards, [Name]
You are a client services representative. When a client emails about a topic, acknowledge their specific interest, recommend the most relevant service based on their situation, and suggest a clear next step. Match the tone: professional but warm. Keep responses under 150 words.
One is written for a person and the other for a model, and the structure is the same. Your template already sets the tone, the structure, the variables and the expected output, which is what instructions for a model do.
From Decision Rules to Sorting
If the email mentions 'refund' or 'cancel', escalate to retention. If it mentions 'pricing' or 'quote', route to sales. If it mentions 'bug' or 'error', route to support. If it mentions 'invoice' or 'payment', route to billing. Everything else goes to the general inbox.
Categories: [{name: 'retention', keywords: ['refund', 'cancel'], action: 'escalate'}, {name: 'sales', keywords: ['pricing', 'quote'], action: 'route_to_sales'}, {name: 'support', keywords: ['bug', 'error'], action: 'route_to_support'}, {name: 'billing', keywords: ['invoice', 'payment'], action: 'route_to_billing'}, {name: 'general', action: 'default_inbox'}]
Your rules set the categories and the routing, which was the hard part. An AI model adds one thing the rules lack: it reads the whole message instead of matching words. “I’ve been trying to reach someone about my account for three days” never says “cancel”, and a model can still flag it as a retention risk.
From Procedure to Workflow
Your client onboarding procedure already lists every step, and today a person runs them.
New Client Signs Up
A form submission starts the workflow
Create Accounts
CRM and billing accounts set up automatically
Send Welcome Email
Personalized onboarding documents sent at once
Schedule Kickoff
Calendars checked and the call booked
Assign Account Manager
Matched on industry and team capacity
Each step is already in your procedure. The system runs them in order within seconds instead of across a day, and nobody skips step three because they were busy.
Finding Your First Candidates
You can find the first candidates yourself, in what your team has already written.
If your team has written down how the business works, the brief for automating it already exists, in your shared drive.
Which template does your team fill in most often?
Book Free Assessment