The best AI project I ever did was free. I walked into a room, gave a presentation about AI, and walked out with three paying clients.
The worst AI project I ever did was technically impressive. Cold emails with 10 data sources, AI-generated icebreakers, automated follow-up sequences. Beautiful system. Terrible results. Because I built it without an ICP, without a clear product, and without thinking about whether anyone actually wanted what I was sending.
The difference between these two projects has nothing to do with technology. It has everything to do with implementation strategy.
After working with businesses ranging from solo freelancers to 50-person companies, I've refined an implementation approach that consistently delivers results within 90 days. Not because it's magic, but because it forces you to do the boring stuff first.
The goal here isn't transformation. It's proof. One concrete result that proves AI works for your specific business.
Step 1: Process Audit
Before you touch any AI tool, you need to know where your time actually goes.
Have everyone on your team keep a time diary for one week. Not estimates. Actual tracking. Every task, every meeting, every "quick" email check that turned into 45 minutes.
What you'll find will surprise you. Most teams spend 40-60% of their time on tasks that are repetitive, structured, and measurable. In other words: perfect for automation.
When I scan a business for the first time, I always find the same categories of waste:
Step 2: Pick ONE Thing
Not five things. Not a comprehensive strategy. One thing.
The criteria:
That last point matters more than you think. Tasks people hate are tasks people rush through, make mistakes on, and procrastinate about. Automating them doesn't just save time, it improves quality and morale.
Step 3: Implement and Measure
Use the simplest tool that works. Don't build a custom AI agent when a Zapier workflow does the job. Don't write code when a no-code tool exists.
Measure three things:
After two weeks, you should have a clear, quantified win. "We saved 6 hours per week on invoice processing" or "Customer response time dropped from 4 hours to 45 minutes."
This win is your internal case study. It's what gets the rest of the team (and leadership, if applicable) on board.
Now you've proven the concept. Time to expand.
Step 4: Map Your Core Workflows
Take your top 5 business processes and map them end-to-end. Not just the task level. The full workflow, from trigger to completion.
Example for a service business:
1. Lead comes in (website form, referral, cold outreach)
2. Lead is qualified (manually? AI-assisted?)
3. Proposal is created (from scratch? template?)
4. Meeting is scheduled (email back-and-forth? calendar tool?)
5. Service is delivered (what tools? what steps?)
6. Invoice is sent (manual? automated?)
7. Follow-up happens (when? how?)
For each step, ask: could AI make this faster, cheaper, or better?
Step 5: Connect Your Tools
Most businesses use 5-15 different software tools. They're usually disconnected islands. Data lives in one place, actions happen in another, and humans are the glue holding everything together.
AI implementation isn't about new tools. It's about connecting existing ones.
The technology stack I recommend:
Connect them. When a lead comes in, the automation scores it, notifies the right person, creates a task, and drafts a follow-up. When a project is completed, the automation generates an invoice, sends a satisfaction survey, and schedules a check-in.
Step 6: Train Your Team
This is where most implementations fail. Not because the technology doesn't work, but because nobody told the team how to use it.
Training isn't a PowerPoint presentation. It's sitting down with people, showing them the workflow, letting them try it, and answering their questions.
The biggest fear employees have about AI isn't that it will take their job. It's that they'll look stupid trying to use it. Address that fear directly: "This is new for everyone. There are no dumb questions. We're all learning."
Schedule hands-on training sessions. Not one session. Multiple. People forget. They develop workarounds. They need reinforcement.
Step 7: Measure Everything
By week 7, you should have multiple automations running. Now measure the aggregate impact:
Build a simple dashboard. Review it monthly. Share results with the team.
Step 8: Expand Intelligently
Look at what's working and ask: "Where else can we apply this pattern?"
If email automation saved 6 hours per week, can you apply the same approach to other communication channels? If AI-assisted proposal writing cut creation time in half, can you apply similar AI assistance to report writing, documentation, or client communication?
Step 9: Establish a Rhythm
AI implementation isn't a project with an end date. It's a new way of working.
Establish a monthly rhythm:
This keeps you continuously improving without the overwhelm of trying to change everything at once.
I'll be honest because I've made these mistakes myself.
Mistake 1: Going too technical too fast. Consultants (myself included) want to build impressive systems. But the client doesn't need impressive. They need effective. Start with the simplest solution that works.
Mistake 2: Low hanging fruit obsession. Everyone talks about "quick wins" and "low hanging fruit." Those are great, but they can distract from the structural changes that create lasting value.
Mistake 3: No aftercare. The biggest failure in AI consulting. You implement, you leave, you send an invoice. Six months later, nobody's using the tools. The retainer model exists because AI needs ongoing attention. New tools launch. Workflows need adjustment. Teams need refresher training.
Three questions to unstick any AI implementation:
1. What's the most repetitive task in your business? Automate that first.
2. What's the most time-consuming task? AI-assist that second.
3. What would you do with 10 extra hours per week? That's your motivation.
If you truly can't figure out where to start, talk to someone who's done it. Not a salesperson. Someone who's actually implemented AI in businesses like yours.
AI implementation isn't about technology. It's about discipline.
The discipline to audit your processes honestly. The discipline to start with one thing instead of everything. The discipline to measure results instead of assuming them. The discipline to keep going after the initial excitement fades.
Every AI project at every business I've worked with followed the same pattern: excitement, implementation, frustration (something broke), adjustment, results. The businesses that succeed are the ones that push through the frustration phase.
90 days. One focused phase at a time. Real results.
How long before I see ROI from AI implementation?
Most businesses see measurable time savings within 2 weeks of starting. Full ROI, where the investment in tools and time pays for itself, typically happens within 30-60 days.
What if my team resists AI implementation?
Start with the person most open to it. Let them become your internal champion. Success is contagious. When one team member saves 5 hours per week, others want in.
Do I need to hire a developer for AI implementation?
For basic automation and AI tools, no. Platforms like Make, N8N, and Zapier are designed for non-developers. For custom integrations or complex workflows, some technical help may be needed.
What's the minimum budget for a meaningful AI implementation?
Zero. You can start with free tiers of Claude, N8N, and your existing tools. A meaningful implementation is possible without spending anything on new software.
Mitchell van Rijkom has guided dozens of businesses through AI implementation. His approach: start boring, measure everything, and never stop improving. He's the founder of AI Survivors.
The best AI project I ever did was free. I walked into a room, gave a presentation about AI, and walked out with three paying clients.
The worst AI project I ever did was technically impressive. Cold emails with 10 data sources, AI-generated icebreakers, automated follow-up sequences. Beautiful system. Terrible results. Because I built it without an ICP, without a clear product, and without thinking about whether anyone actually wanted what I was sending.
The difference between these two projects has nothing to do with technology. It has everything to do with implementation strategy.
After working with businesses ranging from solo freelancers to 50-person companies, I've refined an implementation approach that consistently delivers results within 90 days. Not because it's magic, but because it forces you to do the boring stuff first.
The goal here isn't transformation. It's proof. One concrete result that proves AI works for your specific business.
Step 1: Process Audit
Before you touch any AI tool, you need to know where your time actually goes.
Have everyone on your team keep a time diary for one week. Not estimates. Actual tracking. Every task, every meeting, every "quick" email check that turned into 45 minutes.
What you'll find will surprise you. Most teams spend 40-60% of their time on tasks that are repetitive, structured, and measurable. In other words: perfect for automation.
When I scan a business for the first time, I always find the same categories of waste:
Step 2: Pick ONE Thing
Not five things. Not a comprehensive strategy. One thing.
The criteria:
That last point matters more than you think. Tasks people hate are tasks people rush through, make mistakes on, and procrastinate about. Automating them doesn't just save time, it improves quality and morale.
Step 3: Implement and Measure
Use the simplest tool that works. Don't build a custom AI agent when a Zapier workflow does the job. Don't write code when a no-code tool exists.
Measure three things:
After two weeks, you should have a clear, quantified win. "We saved 6 hours per week on invoice processing" or "Customer response time dropped from 4 hours to 45 minutes."
This win is your internal case study. It's what gets the rest of the team (and leadership, if applicable) on board.
Now you've proven the concept. Time to expand.
Step 4: Map Your Core Workflows
Take your top 5 business processes and map them end-to-end. Not just the task level. The full workflow, from trigger to completion.
Example for a service business:
1. Lead comes in (website form, referral, cold outreach)
2. Lead is qualified (manually? AI-assisted?)
3. Proposal is created (from scratch? template?)
4. Meeting is scheduled (email back-and-forth? calendar tool?)
5. Service is delivered (what tools? what steps?)
6. Invoice is sent (manual? automated?)
7. Follow-up happens (when? how?)
For each step, ask: could AI make this faster, cheaper, or better?
Step 5: Connect Your Tools
Most businesses use 5-15 different software tools. They're usually disconnected islands. Data lives in one place, actions happen in another, and humans are the glue holding everything together.
AI implementation isn't about new tools. It's about connecting existing ones.
The technology stack I recommend:
Connect them. When a lead comes in, the automation scores it, notifies the right person, creates a task, and drafts a follow-up. When a project is completed, the automation generates an invoice, sends a satisfaction survey, and schedules a check-in.
Step 6: Train Your Team
This is where most implementations fail. Not because the technology doesn't work, but because nobody told the team how to use it.
Training isn't a PowerPoint presentation. It's sitting down with people, showing them the workflow, letting them try it, and answering their questions.
The biggest fear employees have about AI isn't that it will take their job. It's that they'll look stupid trying to use it. Address that fear directly: "This is new for everyone. There are no dumb questions. We're all learning."
Schedule hands-on training sessions. Not one session. Multiple. People forget. They develop workarounds. They need reinforcement.
Step 7: Measure Everything
By week 7, you should have multiple automations running. Now measure the aggregate impact:
Build a simple dashboard. Review it monthly. Share results with the team.
Step 8: Expand Intelligently
Look at what's working and ask: "Where else can we apply this pattern?"
If email automation saved 6 hours per week, can you apply the same approach to other communication channels? If AI-assisted proposal writing cut creation time in half, can you apply similar AI assistance to report writing, documentation, or client communication?
Step 9: Establish a Rhythm
AI implementation isn't a project with an end date. It's a new way of working.
Establish a monthly rhythm:
This keeps you continuously improving without the overwhelm of trying to change everything at once.
I'll be honest because I've made these mistakes myself.
Mistake 1: Going too technical too fast. Consultants (myself included) want to build impressive systems. But the client doesn't need impressive. They need effective. Start with the simplest solution that works.
Mistake 2: Low hanging fruit obsession. Everyone talks about "quick wins" and "low hanging fruit." Those are great, but they can distract from the structural changes that create lasting value.
Mistake 3: No aftercare. The biggest failure in AI consulting. You implement, you leave, you send an invoice. Six months later, nobody's using the tools. The retainer model exists because AI needs ongoing attention. New tools launch. Workflows need adjustment. Teams need refresher training.
Three questions to unstick any AI implementation:
1. What's the most repetitive task in your business? Automate that first.
2. What's the most time-consuming task? AI-assist that second.
3. What would you do with 10 extra hours per week? That's your motivation.
If you truly can't figure out where to start, talk to someone who's done it. Not a salesperson. Someone who's actually implemented AI in businesses like yours.
AI implementation isn't about technology. It's about discipline.
The discipline to audit your processes honestly. The discipline to start with one thing instead of everything. The discipline to measure results instead of assuming them. The discipline to keep going after the initial excitement fades.
Every AI project at every business I've worked with followed the same pattern: excitement, implementation, frustration (something broke), adjustment, results. The businesses that succeed are the ones that push through the frustration phase.
90 days. One focused phase at a time. Real results.
How long before I see ROI from AI implementation?
Most businesses see measurable time savings within 2 weeks of starting. Full ROI, where the investment in tools and time pays for itself, typically happens within 30-60 days.
What if my team resists AI implementation?
Start with the person most open to it. Let them become your internal champion. Success is contagious. When one team member saves 5 hours per week, others want in.
Do I need to hire a developer for AI implementation?
For basic automation and AI tools, no. Platforms like Make, N8N, and Zapier are designed for non-developers. For custom integrations or complex workflows, some technical help may be needed.
What's the minimum budget for a meaningful AI implementation?
Zero. You can start with free tiers of Claude, N8N, and your existing tools. A meaningful implementation is possible without spending anything on new software.
Mitchell van Rijkom has guided dozens of businesses through AI implementation. His approach: start boring, measure everything, and never stop improving. He's the founder of AI Survivors.
.svg){kind=icon}
