01 · Start here
Start with the operating question.
Most businesses do not need “more AI” as a starting point. They need a clearer answer to one operating constraint. Is it capability? Can the right prospect discover and understand the offer? Does every worthwhile enquiry receive useful follow-up? Does a qualified conversation move into dependable delivery? Does one owner know what needs a decision this week? Are the documents and specialist-referral boundaries strong enough to support the next move?
02 · The model
The six-pillar self-check.
Mark each statement as visible now, partly visible, or not yet visible. This is a reflection tool; it is not a maturity certification, ranking, or business-performance prediction.
1. AI capability
| Check | Visible now | Partly visible | Not yet visible |
|---|---|---|---|
| We can name the real task before choosing an AI tool. | ☐ | ☐ | ☐ |
| The people using AI understand what needs human review. | ☐ | ☐ | ☐ |
| Useful prompts, examples, or instructions are documented and reusable. | ☐ | ☐ | ☐ |
First practical move: Choose one recurring task. Define the input, desired output, reviewer, and stop condition before asking any tool to help.
2. Visibility and leads
| Check | Visible now | Partly visible | Not yet visible |
|---|---|---|---|
| A prospective client can quickly understand who we help and what problem we solve. | ☐ | ☐ | ☐ |
| Enquiries have a clear source, context, and next action. | ☐ | ☐ | ☐ |
| We know which messages or offers are intended to start a useful conversation. | ☐ | ☐ | ☐ |
First practical move: Pick one primary enquiry path and write the exact context that should accompany every new lead.
3. Follow-up systems
| Check | Visible now | Partly visible | Not yet visible |
|---|---|---|---|
| A new enquiry receives a useful response without relying on one person’s memory. | ☐ | ☐ | ☐ |
| Follow-up reflects the prospect’s stated need rather than a generic broadcast. | ☐ | ☐ | ☐ |
| We can see which conversations need a decision, next step, or respectful close. | ☐ | ☐ | ☐ |
First practical move: Define one rhythm for a qualified enquiry: acknowledgement, context request, useful next step, and owner review.
4. Client delivery
| Check | Visible now | Partly visible | Not yet visible |
|---|---|---|---|
| New clients receive a clear welcome, scope, and next-step sequence. | ☐ | ☐ | ☐ |
| Key client context is available to people delivering the work. | ☐ | ☐ | ☐ |
| We can identify recurring delivery questions, hand-offs, and exceptions. | ☐ | ☐ | ☐ |
First practical move: Map one client hand-off from “yes” to “first useful outcome,” including who owns each step.
5. Owner accountability
| Check | Visible now | Partly visible | Not yet visible |
|---|---|---|---|
| One person owns the next operational decision for each important workflow. | ☐ | ☐ | ☐ |
| We review what is moving, blocked, or uncertain on a repeatable rhythm. | ☐ | ☐ | ☐ |
| AI-generated work is checked by a person who understands the consequence. | ☐ | ☐ | ☐ |
First practical move: Create a 20-minute weekly operating review: What moved? What is blocked? What decision is required?
6. Business foundations
| Check | Visible now | Partly visible | Not yet visible |
|---|---|---|---|
| Core business records and documents are organised enough for the next decision. | ☐ | ☐ | ☐ |
| The owner understands the financial-literacy questions that affect operating choices. | ☐ | ☐ | ☐ |
| We know when a formal tax, legal, finance, or registration specialist is required. | ☐ | ☐ | ☐ |
First practical move: Create a document-readiness list and name the trusted specialist or referral route for regulated matters.
03 · Workflow visibility
Your lead-to-client operating map.
Do not automate a step until you know what good information should move through it.
| Operating moment | The question | Useful artifact | Human check |
|---|---|---|---|
| Visibility | Can the right person recognise the problem we solve? | Clear offer page, post, referral message, or service explanation | Is the claim specific and supportable? |
| Enquiry | What do we need to know before responding? | Source, need, urgency, and contact preference | Is this a real fit or a polite diversion? |
| Context | What information belongs with this conversation? | Notes, previous messages, desired outcome, relevant links | Is client information appropriate to use here? |
| Follow-up | What is the next useful step? | Helpful response, question, resource, or booking option | Does the message serve the recipient? |
| Appointment | What needs to be decided? | Simple agenda and desired decision | Does the right person own the decision? |
| Client delivery | What must happen after yes? | Scope, welcome, hand-off, delivery checklist | Are ownership and exception handling clear? |
04 · Tools with boundaries
Tool context and human guardrails.
| Tool context | Useful question | Guardrail |
|---|---|---|
| ChatGPT, Claude, or Gemini | Does this help research, draft, structure, review, or prepare? | Verify facts, protect sensitive client context, and retain accountable human review. |
| Make, n8n, or Zapier | Which repeatable hand-off benefits from a structured trigger and next action? | Document failures, exceptions, ownership, and a safe fallback. |
| SalesRobot or CRM workflows | What context should guide lead follow-up? | Avoid generic or excessive messages; review permissions and contact expectations. |
| Documents and templates | What should remain reusable instead of living in memory? | Keep version ownership, access boundaries, and source checking clear. |
The five-question guardrail: What work is this improving? What context can it safely use? What requires human checking? What happens when it fails? Who owns it after setup?
05 · Practical reset
A seven-day operating reset.
This is not a promise of transformation. It is a small way to replace scattered activity with observable learning.
06 · Honest evidence
What to measure—and not promise.
Good operating measures begin as questions. Can we see how long it takes to give a useful first response? Which qualified conversations have a defined next step? Which hand-offs create recurring delay? Is a documented AI method actually used and reviewed? Which decision remains blocked because nobody owns it?
07 · Next best conversation
Turn one page into a working operating move.
If you would like to work through one real workflow, bring the people involved, the current tool context, and the decision you want to make next. The right starting point may be practical AI training, workflow design, follow-up systems, owner review rhythm, document readiness, or a qualified specialist referral.
Book an AI System Audit ↗This guide is educational material for internal reflection and planning. It is not tax, legal, investment, accounting, registration, or filing advice. Reuse only with attribution to Gurvinder Singh and Business Automation OS™.