NEOEDGE AIThe AI questions are stacking up. Let’s clear the pile.
A key process is slowing down your business. An AI pilot isn’t delivering. A vendor wants a signature. I test the options on your real work, build what earns its place, and train your team to run it.
Who you’ll work with
- WealthsimpleLed the AI strategy, from 3 to 100+ models
- GovernanceSet up AI governance with Legal and Security
- NEOEDGE AIHelping Canadian firms rethink work with AI
Click any item to see how I’d approach it ↓
Tap any item to see how I’d approach it
Illustrative examples, not a real client. Research figures are from 2026 reports. Each first read is a starting point. The real call comes from testing on your own work.
Not every idea became a tool. That’s the point.
I worked with a Canadian mortgage lender as its AI consultant, across nine teams, and I still advise them. Every idea was tried on real work first. Here’s where five of them landed.
Read the case study ↗︎“Working with NEOEDGE AI didn’t just give us a single-use AI tool; it changed the way we think about our workflows and how we can utilize AI at our core.”
Senior executive, Canadian mortgage lender
- Sales & investor relationsAI chatbot for fund questionsAnswers staff’s fund questions in Teams. Passed 30 tests before launch without a made-up answer.Live
- DevelopmentChoosing an AI testing toolMore than 15 tools compared. A small trial instead of a contract.Trial first
- AccountingCard statements, codedTyping became reviewing. An accountant checks the suggested codes.Live
- ExecutiveThe CEO’s weekly briefBlockers, momentum and decisions in one brief, each point linked to its source.Prototype
- Capital marketsFinancial lender reports, automatedAutomation of reports built from raw data, and highlighting outliers for human action.Live
AI has to earn its place in your work.
- On your work, not a demo. Your files, your exceptions, your people.
- With the checking counted. Time saved means little if someone re-checks everything.
- With someone who owns it. A named person who answers for it when it gets one wrong.
Start small. Stop anytime.
Each step is scoped and priced on its own. Not sure which fits? That’s what the first call is for.
Decision Sprint
I test one AI option on your real cases, then give you a written call on what to do.
≈ 2 weeks · fixed feeA decision record you can act on.
The call, what we tested, what to do next, and who owns it. Your team, your risk function and your board can all read it.
Clear Call Guarantee. A clear written call, or I keep working at no cost.
I run two Sprints at a time
See an example record
- Call
- Fix before rollout. Routine files worked. Two gaps need fixing first.
- Tested
- 40 real files: 31 handled correctly, 6 needed a reviewer’s correction, 3 missing pages went unflagged.
- Next move
- Add a missing-page check, send non-standard income to a reviewer, then re-test on 20 new files.
- Owner
- Underwriting operations lead, with a named reviewer for exceptions.
Controlled Workflow Pilot
When the evidence supports it, I build and test it with your team, then hand it over for your people to run.
6–8 weeksOSFI E-23 evidence work
For federally regulated institutions with a specific AI or model evidence gap. The guideline takes effect May 1, 2027.
Fixed-fee first stepWorks inside your rules
Approved tools, your data rules, risk and IT from day one
- Your approved tools. I work in what IT has already approved.
- Your data rules. Client data is used only where approved. Until then, we use realistic samples.
- Risk and IT from day one. Your technology, risk and compliance owners are part of the work, not a sign-off at the end.
- A record you can show. Every test and call is written down for your board, auditors and examiners.
I plan it. I build it. Your team runs it.
I’m Sam Talasila. I led LLMs and data products at Wealthsimple, where the firm grew from 3 to 100+ models, and data-science teams at Shopify. You work with me directly, from the plan to the handoff.
Tools and guides.
What’s in your AI in-tray?
Need help sorting out AI at your firm? In 30 minutes, tell me what’s in your in-tray. I’ll tell you where I’d start, even if the answer is to leave it alone.