Here's the conversation I have with nearly every business owner who reaches out:
Them: "We need to automate X process. Can you build it?"
Me: "I don't know yet. Let me shadow your team first."
Them: "Can't you just tell me how much it'll cost?"
Me: "Not until I know what you're actually doing."
This frustrates people. They want a number. They want a timeline. They want certainty.
But here's what I've learned after building dozens of automations: the process you think you have is never the process you actually have.
The workflow chart on your wall? That's version 1.0 from three years ago. The actual process includes the workaround Sarah invented when the CRM broke last month, the manual check Mark does because "the system doesn't catch everything," and the edge case that only happens on the third Tuesday of months that end in 'r'.
If I quote you an automation based on your description of the process, I'm quoting you for something that doesn't exist. When we start building and discover the real workflow, we're already off track—budget blown, timeline extended, trust damaged.
This is why I split every automation project into two distinct phases. It's not to drag things out. It's to make sure we're building the right thing.
Phase 1: Discovery & Architecture (Week 1-2)
Phase 1 has one job: understand reality.
Shadowing
I sit with your team and watch them work. Not for five minutes. For hours.
I'm looking for three things:
1. The actual steps
Not what the manual says. Not what you think happens. What literally happens when someone processes a document on a Tuesday afternoon.
2. The decision points
Where does your team use judgment? Where do they say "it depends"? Those moments are either automation opportunities or automation landmines.
3. The edge cases
"Oh yeah, when the vendor sends a handwritten invoice, we have to..." There it is. The thing that happens 2% of the time but breaks 100% of naive automation attempts.
Shadowing reveals the real workflow—including all the workarounds, exceptions, and tribal knowledge that never made it into any documentation.
API Research
Next, I research your tools.
Your CRM, your email system, your document storage, your industry-specific software—I need to know what's possible.
Some systems have robust APIs. Some have terrible APIs. Some have no APIs at all.
This determines what we can build. If your management system doesn't expose the data we need, we're not building that integration. We're building something else.
This step saves weeks of wasted development time.
Architecture & Proposal
Finally, I design the automation and create a detailed Phase 2 proposal.
This isn't a vague estimate. It's a complete implementation plan that includes:
- Exactly what gets automated and what stays manual
- The specific tools and integrations we'll use
- Setup cost and ongoing maintenance cost
- Success criteria and metrics
- Timeline for implementation
At the end of Phase 1, you receive a comprehensive Phase 2 proposal. You know exactly what you're getting and what it costs before any development starts. You decide if the ROI makes sense. No surprises, no commitments until you see the full plan.
Phase 2: Build & Implementation (Week 3-6)
If the Phase 1 proposal makes sense, we move to Phase 2—actually building the automation.
Development (2-3 weeks)
I build the system based on the architecture from Phase 1.
You're not in the dark during this period. I send progress updates, share preview links, and flag any issues that come up.
Development timelines vary based on complexity:
- Simple workflows: 1-2 weeks
- Moderate integrations: 2-3 weeks
- Complex multi-system automations: 3-4 weeks
Testing & Validation (1-2 weeks)
The automation is built. Now we test it against reality.
I train your team, they start using it, and we watch for edge cases. The handwritten document. The vendor who formats data differently. The scenario we didn't anticipate.
We find the gaps, we fix them, and we validate that the automation actually works in production—not just in theory.
Success Verification
Finally, we check the metrics.
Did we hit the success criteria from Phase 1? Are you actually saving the time we projected? Is the automation stable?
If yes, project complete.
If no, we keep adjusting until we do.
The Timeline
From first meeting to completed automation: 6-7 weeks on average.
Week 1: Initial meeting, Phase 1 begins
Week 2: Shadowing, research, Phase 2 proposal delivered
Week 3: Proposal review and Phase 2 kickoff
Weeks 3-5: Development
Weeks 5-6: Testing and validation
Week 7: Success verification and project close
Can it go faster? Sometimes. Simple workflows can be done in 3-4 weeks total.
Can it go slower? Often. Complex integrations or scheduling constraints can push it to 8-10 weeks.
But 6-7 weeks is the realistic target for most projects.
Why This Process Works
No surprises: You know exactly what you're getting and what it costs before development starts.
No wasted effort: We're building for the real process, not the imaginary one.
No abandoned projects: The testing period ensures the automation actually works in your environment.
No ongoing dependency: I train your team and document everything. You're not stuck calling me every time something changes.
The worst outcome isn't paying too much or waiting too long. The worst outcome is spending 6 weeks building the wrong thing.
This two-phase process prevents that.