Resources
What to automate first.
Pick the boring, repeated, already-ruled work. Leave the judgment with a person.
- Fixed scope
- Human approval
- Built around your tools
- Documentation you own
Quick answer
Automate the boring, repeated, already-ruled work first. Inbox triage. A follow-up draft. A CRM field after you confirm. A weekly number pull. Leave the judgment with a person. If you cannot name the reviewer, the rule, and the tool, stop.
The first workflow should be work you already do the same way twice a week. Not the once-a-quarter exception. Not the project you are still inventing. That is how a build stays small enough to review in one glance.
Decision framework
Start where the rule already exists. If two people do the same job two different ways, you do not have a workflow yet. I will write that on an Audit. I will not pretend a model will invent the SOP.
Start where a person can review in one glance. If the reviewer needs a meeting to understand the draft, the workflow is too wide. Shrink it until yes / edit / discard takes a minute. That is why I like follow-up drafts and inbox triage as first builds more than “automate operations.”
Start where the tools already exist. I build around the inbox, CRM, docs, and calendar you already pay for. A new platform is not the first move. Desktop chat can be enough for one person — see Claude Desktop vs a custom agent. A shared path needs a named reviewer and usually an Audit.
On the Audit I score each opportunity Value × Difficulty × Risk. The first build is usually high value, medium or low difficulty, and a risk I can gate with a named person. The sample Audit shows that scoring without a fake client name.
Example · EXAMPLE / COMPOSITE
A shop wants “AI for the business.” The pile is actually three jobs: inbound mail, CRM typing, and a Monday packet. The first build is not all three. It is the inbound draft reply with a named approver. CRM write is second. The packet waits until the fields are less of a rumor. That is the dealer-group composite and the professional-services composite on the homepage — different verticals, same order.
When this applies
It applies when you can point at a repeating path and a person who will review. It applies if you are choosing between a $199 Fit review, the $1,250 Audit, and a $1,495 personal assistant. The Fit review answers “is this even a first workflow.” The Audit ranks the list. Personal is when the work is one human’s files and calendar, not a team system.
When it does not
It does not apply to judgment work: pricing exceptions, first-time complaints, hiring, money movement, legal sign-off, unsupervised public posts. Those are in when not to automate and the ungated list of ten.
It does not apply if your real ask is a full ops overhaul. I do not sell that. One practical workflow at a time.
Cost and risk
Current prices: Fit review $199, Audit $1,250, Personal $1,495, Business typically $3,750–$7,500 after the Audit. Details on AI agent setup cost and Audit cost.
The risk of picking the wrong first workflow is not a sunk vanity metric. It is a reviewer who stops reviewing, or a path that needs judgment you never wrote down. I would rather tell you to leave it alone. Custom AI engagements commonly run well beyond this range. Effortless Agents is intentionally scoped around one practical workflow at a time.
FAQ
Should I start with Personal or the Audit?
One person’s files and calendar: Personal. A shared company path: Audit first.
What if I cannot pick the first workflow?
That is what the $199 Fit review and the tell-me form are for. I will give you a straight answer.
Do you pick the idea with the biggest payoff?
I score value, difficulty, and risk. I do not invent an hours-saved number to rank the list.
Next step
Write the repeating task, the software, and who reviews. Tell me your workflow, or buy the Audit if you already know you want the ranked map. Then read the matching example page before you widen scope.