The Owner's Guide
Using AI in Your Business.
What actually works when you're running a 5–25 person company — from someone who's spent eight years building AI systems, running businesses, and learning what the consultants won't tell you.
Quick Answer
What does 'using AI in your business' actually mean for a small company?
Using AI in your business means finding one repeated process that costs you time, money, or opportunities — then building a system where AI does the work in the middle while a person reviews the output. You don't need to revolutionize your company. Start with one small win, verify it works, then build the next piece. The businesses that succeed with AI are the ones that finish small projects, not the ones that start big ones.
- →Start with one repeated process, not a company-wide transformation
- →AI prepares the work — a person reviews and approves the output
- →The sweet spot is 5–25 employees: big enough to have real processes, small enough that the owner still feels the pain
- →Real costs: $1,000 for a focused fix, $5,000 for a connected workflow
- →The #1 mistake: scoping the project too broadly
The Sweet Spot
Why 5–25 employees is where AI pays off fastest.
Below five employees, the owner can still keep the business in their head. They know which customer needs a response, which quote is outstanding, what each person is working on. When something gets missed, they step in and fix it personally.
Around five employees, that changes. The owner is no longer just doing the work — they're assigning it, checking it, answering questions, managing schedules, following up with customers, and making sure information moves between people. A task that wastes 30 minutes isn't wasting just the owner's time anymore. It's affecting several employees, delaying customers, holding up quotes, and blocking other work.
Above roughly 25 employees, many companies need more formal technology support — a dedicated person, a managed provider, or a fractional technology leader. Those professionals are essential, but their focus is often different: networks, devices, security, user accounts, infrastructure.
BotLogix focuses on business operations, not IT infrastructure. We sit with the owner and ask: where is money being made? Where is time being lost? Where are opportunities falling through the cracks? That's the 5–25 sweet spot — big enough to have real processes, small enough that the owner still carries the company personally.
The goal is not to pretend the company is larger than it is. The goal is to give a good smaller company some of the operating capabilities that were previously available only to larger businesses.
— Steffen deGraaf, BotLogix
Where to Start
The work nobody wants to do — that's where AI pays first.
Most AI content tells you to start with something exciting — a chatbot, a content engine, a predictive dashboard. That's backwards. The first place AI pays is in the boring, repeated work that eats hours every week.
When I look at a business, I start with intake and operations. How do opportunities arrive? How are leads handled? How quickly does the company respond? Where do jobs get delayed or lost? Then I look at the operational side: how is work organized, where is time being wasted, where are costs increasing?
The pattern is simple: Human input. Automated work. Human-approved output. The human provides the request — a voice note, an email, a form submission. AI processes and organizes it. A person reviews the result before it goes anywhere.
Missed call recovery
Every missed call is a potential job going to a competitor. AI can text back within 60 seconds, qualify the inquiry, and book the call.
Quote follow-up
Quotes that sit for days get lost. AI can route, remind, and track every quote through the pipeline without someone manually checking.
Email triage
Important messages buried in newsletters and junk. AI can surface high-priority items — a quote request, a complaint, a deadline.
Document processing
Contracts, invoices, intake forms. AI can extract the data, populate the fields, and flag what needs human review.
Voice note to action
Record a voice note between appointments. AI transcribes it, identifies the job, creates tasks, and puts information in the right workflow.
Client intake
New client onboarding that takes hours. AI can prepare the forms, send the checklists, and track completion automatically.
Nothing dramatic has to break for the business to lose money. The failure happens quietly — quotes fall between the cracks, follow-ups are delayed, opportunities disappear.
— Steffen deGraaf, BotLogix
What It Costs
Real numbers from an operator, not a sales deck.
A practical starting point for custom AI implementation is approximately $1,000 per consulting or development day. If the problem can be understood, solved, tested, and implemented in one day, it may be a $1,000 project. More complex workflows take longer.
But the budget shouldn't start with “how much AI can I afford?” It should start with: “What is this problem currently costing me, and what would it be worth to solve it?”
$0 — Start free
Use ChatGPT, Claude, Gemini, Perplexity. Learn what each does well. Summarize documents, organize meeting notes, draft emails, analyze a repeated process. The goal isn't to learn prompts — it's to understand where AI fits your actual business.
$1,000 — One focused fix
Choose one or two specific tasks. Alert the owner to high-priority quote requests. Turn form submissions into CRM records. Convert voice notes into structured tasks. Prepare standardized first drafts. The goal is something useful in operation, not a report.
$5,000 — Connected workflow
Instead of one isolated task, connect several steps of a business process. Capture an opportunity, organize the information, create tasks, prepare a draft response, update the CRM, alert the right person. One complete workflow with a measurable beginning and end.
$50,000 — Department-level change
At this level, the project may connect multiple departments, data sources, customer touchpoints, and internal systems. Only move forward when the underlying process has enough volume or value to justify it. In the right business, this prevents or recovers a much larger amount. In the wrong one, it becomes an expensive system nobody uses.
Costs owners don't expect:maintenance (processes change, platforms change, models change), usage costs (per-request or per-user fees), testing and supervision (someone needs to verify the output), and the owner's own time (the consultant understands technology — the owner understands the business).
The size of the investment is not what determines whether it is worthwhile. The value and frequency of the problem determine what the solution is worth.
— Steffen deGraaf, BotLogix
A Real Example
Six hours to 30 minutes: what quoting automation actually looks like.
Before we built our quoting system, preparing a group-benefits quote was an extremely manual process. A person had to review the client's existing benefits contract and locate dozens of specific details. Every insurance carrier presents similar information differently — different pages, different headings, different formats.
The process included: reviewing the contract, extracting plan details, building a spreadsheet, preparing submissions for three or more insurers, sending individual emails, waiting for responses, reviewing returned quotes, rebuilding information into a comparable format, preparing a client presentation, and presenting options to the customer.
For one opportunity, the combined work could take more than six hours. We reduced that to approximately 30 minutes of active human work — a reduction of roughly 5.5 hours per quote, or more than 90 percent of the active processing time.
The system also creates a better customer experience. Information is captured more consistently. Submissions are prepared sooner. We can approach insurers more quickly — which matters because being first to register an opportunity can affect whether we obtain a quote at all.
The Numbers
6+ hrs
Manual process per quote
30 min
Active human work after automation
90%+
Reduction in processing time
The advice to another owner: examine both volume and value. A six-hour automation may not be worth building for a task that happens once a year. It can be extremely valuable when the task happens every week, affects revenue, delays customers, or requires expensive staff time.
Do not automate something merely because it can be automated. Automate it because the repeated cost of leaving it manual is greater than the cost of building and maintaining the solution.
What Goes Wrong
The failure modes consultants never mention.
AI makes it easy to begin building. It does not automatically make it easy to finish, implement, maintain, or get employees to use what was built. Here are the three failure patterns I see most often.
1. Scoped too broadly
The company tries to solve a large, complicated problem from the beginning. Automate an entire department. Replace several systems. Create the perfect end-to-end workflow before proving the basic components work.
Then one foundational part fails. The team realizes the project is more complicated than expected, progress slows, costs increase, and the tool is abandoned at 25, 50, or 75 percent completion. The better approach: build one small component that works, put it into operation, and expand from that working foundation.
2. The process wasn't understood
The documented process says one thing, employees do something different, and exceptions are handled through experience rather than written rules. The system is built in a vacuum.
It may technically perform the task it was designed for, but it doesn't reflect the real-world process people use every day. Before automating, understand the inputs, decisions, exceptions, handoffs, and desired output.
3. Tools and information are fragmented
Different people use different platforms, prompts, file locations, and methods. No central source of truth. No plan for how information moves between systems.
The result is a collection of isolated experiments rather than a functioning business system. A successful project needs shared information, clear ownership, and a structure that lets the technology become part of the company — not remain inside one employee's account.
A YouTube video may bring you 50 or 60 percent of the way there. The remaining 40 or 50 percent is where the real system gets built. That is also where many projects are abandoned.
— Steffen deGraaf, BotLogix
The Decision
Build, buy, or do it yourself — the owner's framework.
Start with the problem, not the preferred solution. Once the process is understood, consider the three options in this order.
1. Does a suitable off-the-shelf tool exist?
An existing tool is often the fastest option — but only when it solves the actual problem without creating unnecessary risk or complexity. Look at the monthly and long-term cost, how much control the business has, whether it integrates with current systems, who owns the data, whether the data can be exported, and what happens if the company stops using the platform.
A relatively inexpensive tool can become costly if employees need to change their entire process, data becomes trapped, or the company has to purchase several additional systems to make it useful. I reject a tool when it gives the vendor too much control over a process the business cannot afford to lose.
2. Is the process important or unique enough to justify a custom build?
A custom system makes sense when the process creates a competitive advantage, affects a large amount of revenue, or cannot be handled properly by an existing tool. The owner focuses on the outcome — explaining the business process, reviewing decisions, testing the result — without personally solving every technical roadblock.
A custom build should provide greater control over the workflow, the data, the user experience, future changes, connections to other systems, and the long-term cost structure.
3. Can the owner realistically build it themselves?
Building with AI assistance can be a great option for an owner who enjoys technology and has enough time to learn. AI makes beginning very easy. The real question is: what is the relief valve when they hit a roadblock?
Who helps when the integration fails, the data is structured incorrectly, usage limits are reached, or the tool works in testing but not in the real business? Consider how much the owner's time is worth, whether this is the best use of that time, and whether someone can help when the project becomes more complicated.
The best solution is not automatically the most advanced one. It is the solution the business can afford, control, maintain, and actually use.
— Steffen deGraaf, BotLogix
The Hard No List
What you should never fully automate.
The dividing line is whether the task follows a repeatable process or requires a meaningful opinion. When the process is predictable, automation can work extremely well. When the decision depends on context, personal judgment, client history, ethics, risk, or the way a message may be interpreted — a human should remain involved.
Final professional recommendations
AI can prepare the analysis. You make the call.
Sensitive customer complaints
The customer expects a person to understand the situation.
Major financial decisions
Errors here have serious consequences.
Relationship-based communications
Your voice, your relationships, your reputation.
Public statements affecting reputation
One wrong message distributed to hundreds of people.
Messages involving conflict or emotion
Context and empathy matter more than speed.
Final approval of mass communications
Automation can distribute the wrong answer faster than a human ever could.
Decisions with legal or personal consequences
The risk is not worth the efficiency gain.
A computer should handle the work customers expect a computer to handle. A person should remain involved where customers expect a person to be involved. Automatically sending a routine, verified invoice is appropriate. Automatically sending a sensitive email written entirely by AI is not.
AI can often bring the work 90 percent of the way there. The final 10 percent should remain human.That last 10 percent protects the company's reputation, brand, accuracy, judgment, and voice.
The First 90 Days
What a real implementation looks like — week by week.
Most successful implementations do not begin by changing everyone's job. They begin with a small adjustment. The first week is about sharpening the system, not deploying perfection.
Week 1 — Limited real-world use
Put the system into limited operation. Watch how employees interact with it. Identify unexpected inputs and exceptions. Correct confusing steps. Adjust notifications, routing, or permissions. Confirm the output is accurate. Make sure employees aren't bypassing the system. Document what changed. Real employees use systems differently than developers expect — these corrections are normal, not failures.
Weeks 2–4 — Expand and stabilize
Once the first piece works in the real environment, expand its use. Connect it to the next step in the process. Address the exceptions discovered during week one. Begin measuring the result: time saved, errors prevented, opportunities recovered.
Month 2 — Build the next piece
Take what you learned from the first project and apply it to the next opportunity. You may be able to reuse 75 percent of the technology from the first solution for another part of the business. This is how the company develops a practical, connected technology foundation.
Month 3 — Evaluate and decide
Look at the numbers. Has the system reduced hours? Prevented errors? Recovered opportunities? Is the owner happier? Are employees actually using it? If the answer is yes, build the next piece. If not, understand why before investing more.
Start smaller. Finish the project. Confirm that it produces the result you expected. Then connect the next small project to the first one. The goal is not to build more. The goal is to finish more.
— Steffen deGraaf, BotLogix
The Real Problem
You don't need another app. You need your existing apps to work together.
Most companies did not intentionally design their current technology environment. They accumulated it. Email, office software, spreadsheets, cloud storage, social platforms, design tools, accounting software, CRM, scheduling tools, project management, messaging apps, industry-specific software.
Each tool may have solved one problem. Over time, the company became scattered across ten or twenty different systems. Employees spend much of their day moving digitally from task to task: email to spreadsheet, spreadsheet to document, document to PDF, PDF to CRM, CRM back to email.
The tools are not necessarily the problem. The missing connection between the tools is the problem. The future is not that every business needs another application. The future is that the owner or employee can ask for an outcome and the existing systems can work together to prepare it.
Do not automate your responsibility. Automate the unnecessary work surrounding it.
— Steffen deGraaf, BotLogix
Common Questions
What owners actually ask.
How do I know if my business is ready for AI?
You're ready when you can explain the process you want to improve — the problem, the input, the repeated steps, the people involved, the decisions, the expected output, and the exceptions. If you can't explain the process, fix that first. AI cannot repair a workflow that nobody understands.
What's the one mistake to avoid?
Scoping the project too broadly. A large scope causes every predictable problem: the project takes too long, costs too much, employees struggle to understand it, the technology changes before it launches, and the team loses confidence. Think big about where the company could go. Scope small enough to get something working.
Should I just hire a tech person instead?
Sometimes. An internal hire makes sense when there's enough continuous work to justify the salary. But don't hire a full-time technology employee simply because AI feels important. Hire one when the volume and value of the work make the position economically sensible.
What about platform risk?
Build your business on assets you own. Use outside platforms to extend your reach, not to hold your entire business together. Your website, CRM, customer records, documents, workflows, and operating knowledge should not exist exclusively inside a social network, an inbox, or a single AI provider.
How do I get employees to actually use the system?
Don't begin by telling employees everything is changing. Begin by removing the parts of their work that are repetitive, frustrating, inconsistent, or easy to forget. The system should feel like help, not disruption. Preserve the process people already understand while quietly removing unnecessary steps from it.
What's the most satisfying result you've seen?
In 2018, we built an early keyword-triggered DM automation. In our experience with that campaign, it brought in approximately 500 new contacts within 24 hours. What made it satisfying wasn't the number — it was watching a complicated backend system turn into an effortless experience for the customer. The technology almost disappeared. That's still what I find most satisfying today.
Go deeper
How Much Does AI Implementation Cost?
Real numbers: $0 to $50,000 budgets.
The 7 AI Mistakes That Cost Money
What actually kills AI projects.
What to Automate First
The one-week exercise that reveals your best opportunity.
Why Your First AI Project Should Be Boring
Shiny tools become expenses. Boring systems create value.
Build, Buy, or Hire
The 7-step decision framework.
The First 90 Days
Week by week: what a real implementation looks like.
How to Hire AI Help Without Getting Burned
Red flags and the right questions to ask.
AI Strategy Day
$1,000 on-site. A real plan you own.
30-minute strategy session
For real businesses with real problems.
Not for tire kickers. Not for tech tourists looking for a demo. For business owners who already know something in their operation is broken — and want to know if AI can fix it.
This isn't about replacing people.It's about killing the bad business processes your team hates as much as you do — the manual work, the missed calls, the after-hours admin.
Strategy sessions are free. Strategy Days are $1,000, capped at 4 people per session. Larger teams or multi-location workshops by special arrangement — ask when you book.
Your session is with
Steffen deGraaf
Founder, BotLogix · Burlington, ON