Practical AI for businesses that don't have a technology team.
Most AI projects stall because nobody owns the unglamorous parts. I help small and mid-size businesses in the GTA work out what's actually worth doing, then build it so it still runs after I leave.
If any of this sounds familiar
“We've tried a few tools. Nothing has really changed how we work.”
“Everyone says we need AI. Nobody can tell us what to do first.”
“Someone built us something. Now nobody knows how it works.”
“We don't want to depend on something we can't maintain.”
Not sure where you stand? Take the 5-minute readiness check.
These are the conversations I have most often.
How it works
- 1
Conversation
30 minutes, free. What you're doing now, what isn't working, and whether there's anything worth pursuing. Within a day you get a one-page proposal: your gaps ranked and the next step I'd recommend. If there's nothing worth doing, I'll say so.
- 2
Start small, or start with the full picture.
You choose.
- 3
Build
Six to twelve weeks. Implementation with the logging, version control and documentation that keep it working. Before I leave, one person on your team owns it, and the people who use it every day have written routines.
Optional: Keep it working.
Monthly. I check each system against standards we agree on and send a one-page scorecard: what ran, what failed, what changed.
What you get from an Assessment
A written report you keep, whether or not you hire me to build anything.
- The three things worth doing first, on one page.
- How work moves through your business today, mapped.
- Where time and money leak, with hours and dollars estimated from your own numbers.
- Every opportunity ranked by value and effort.
- What not to do yet, and why.
- What has to be true first: your data, your software and ground rules for AI.
- A 90-day plan and a 12-month roadmap, with what each step takes.
- A walkthrough meeting with you and your team.
Two to three weeks, depending on the size of your business. Fixed fee, agreed before we start. More than 10 staff? Start with one department: Sales and Marketing, Operations, Finance or Customer Support.
Simplify recurring tasks
Estimate what you could free up
Enter a few operating assumptions. The hours and the dollar value of that released capacity update as you change them.
These figures use the starting assumptions. Changing them needs JavaScript. Book a 30-minute call
Count purchase orders entered from email. Waiting time does not count.
Average employee minutes on that work for one order.
This is the cut in the targeted manual effort. It is not the share of work that becomes fully automatic.
Use a blended cost for the people doing the work, including employer costs. This is not a wage and it is not a billable rate.
Released time is not cash. Include only what you can capture through less overtime, avoided hiring, lower outside processing or more completed work.
Entering orders from emailed purchase orders
About 33 hours a week back.
Close to one full-time person.
$29,167
Estimated annual value before implementation costs
Based on your assumptions. Excludes implementation, licensing, integration, maintenance, human review and other ongoing costs.
Current annual effort
3,333
hours
Hours released annually
1,667
hours
Gross annual labor capacity value
$58,333
How this figure is calculated
- Current annual effort is 3,333 hours. That is 20,000 orders at 10 minutes each.
- Hours released annually are 1,667. That is 50% of the current effort.
- Gross annual labor capacity value is $58,333. That is 1,667 hours at $35.
- Estimated annual value before implementation costs is $29,167. That is 50% of the gross value.
- Residual manual hours are 1,667.
Results are illustrative estimates based on the assumptions you enter. They are for information only. Actual outcomes depend on the workflow, how suitable the work is for automation, volume, exception rates, implementation cost, ongoing operating cost and whether released capacity can be used. Labor capacity value is not a cash saving. This calculator is not a financial guarantee or a formal business case.
Bring these assumptions to a 30-minute call.
Book a 30-minute callWhy me
Twenty years building the systems businesses run on. I still write the code.
15 years, 500 users, 10 countries
I designed and built the ERP that ran a medical and optical distributor working from a bonded warehouse in Dubai. It never needed replacing while the business grew to US$100M in revenue.
4 days to 4 hours
I cut dealer onboarding at an automotive marketing company from 4 days to 4 hours, moving dealers to a new website platform and setting up their campaigns: 400 dealers in 2 months with a team of 5.
Optical, automotive, distribution
I advised a global lens manufacturer's regional business on lab production and the optical shop channel, and built software for car dealers' online marketing.
I published STRATA Protocol, a method for governing AI-assisted development, because most of what gets built with AI quietly loses the things that make software trustworthy.
Based in and working across the Greater Toronto Area. More at florantepascual.com.
Start with a conversation.
Thirty minutes, no charge, no pitch. If I'm not the right fit, I'll tell you who is.