
From pro snowboarder to AI CEO: Nicole Donnelly’s founder story
with Nicole Donnelly, AI Smart Marketing
From pro snowboarder to AI CEO: Nicole Donnelly’s founder story
Show Notes
Nicole Donnelly is the founder of AI Smart Marketing and AI Smart Insiders, an AI training and consulting firm that helps businesses at every level - from solo consultants to large enterprises and government agencies - implement AI across their operations. Before founding AI Smart Marketing, Nicole built Baby Legs, a baby legwarmer brand she grew from a personal parenting hack into a global product sold in Target, Nordstrom's, and 85 countries within three years before selling it in a seven-figure exit. She then ran a marketing agency, got into AI in 2015, and pivoted the agency to focus on AI training and consulting as it became clear the technology would reshape every business function.
Nicole is also a former competitive snowboarder who trained with the Park City Snowboard Team and competed in Big Air, halfpipe, border cross, and slope style - with at least 20 documented concussions along the way. She lives in Vancouver, BC, travels globally to deliver AI training (Germany, Switzerland, Istanbul, Lisbon, Vietnam, Thailand on the schedule), and runs a YouTube channel, AI Smart Marketing, and a weekly newsletter where AI drafts the content and Nicole edits it to add her own perspective and occasional contrarian opinions.
Why Small Business AI Adoption Is the Real Stakes
Nicole's core concern is not AI warfare or existential risk - it is the 75% of small and medium businesses that have not yet adopted AI. Roughly 25% of small businesses, per the latest surveys, have done AI training and implemented it meaningfully. The remaining 75% represent an enormous population of employers and employees who are sitting out what she describes as the most significant competitive window small businesses have ever had.
The argument is structural: Google search volume declined for the first time ever this year because AI search is eating into it. Small businesses that build AI discoverability now - optimizing for how AI search surfaces and recommends them - are establishing positions that will be much harder to claim once the window closes. The macro backdrop of instability makes the timing even more urgent: companies that spend this summer hardening their AI operations will be positioned to accelerate when conditions stabilize, while slow-moving large enterprises are still approving the initiative in committee. The small company advantage in AI is speed and agility, and it is temporary.
The House Renovation Framework for AI Strategy
Nicole's most useful framing for business AI adoption is the home renovation analogy. You might only have the budget or appetite to renovate the bathroom right now - but you need a strategy for the whole house before you start, because the bathroom floor connects to the hallway floor, which connects to the living room. Decisions made in isolation create mismatches you'll pay to fix later.
Applied to AI: start with something small and low-disruption (painting a bedroom rather than gutting the kitchen), but design the entire organization's AI roadmap first. Know where you want the technology to reach across every function - operations, sales, marketing, HR - and sequence the implementation from least to most disruptive. Nicole's team will sometimes decline an engagement or back a client up to the strategy phase if they want to buy a single tool without thinking through how it connects to the rest of the org. The whole-house view is not optional; it is what prevents wasted spend and failed rollouts.
How to Build a Weather-Based Sales Forecasting Agent in 30 Minutes
In an upcoming AI Lab session, Nicole's team is building a sales forecasting agent for a small outdoor gear company whose revenue is seasonal and weather-dependent - warmer winters mean less gear sold, El Niño years look different from La Niña years. The build: create parallel custom agents in ChatGPT and Gemini, each fed the same knowledge base of historical sales data and weather data. Include a link to the Farmer's Almanac in the agent instructions for forward-looking seasonal forecasting (Nicole recommends it over Weather Channel for full-season regional predictions, citing her snowboarding background as her source).
Running both agents in parallel serves a specific purpose: Gemini consistently outperforms on math accuracy (in a recent test, Gemini got the numbers right where GPT-4o did not), while ChatGPT often delivers stronger qualitative insights. Neither is universally better - the right approach is to test natively, pick the model that performs best for the specific output type, and then chain them via API if needed. This live parallel-testing methodology is what distinguishes Nicole's training from generic “here is what AI can do” content.
What Actually Differentiates a Real AI Consultant
Nicole is direct about the noise in the AI consulting market: many people calling themselves AI consultants lack the business acumen to think about implementation holistically. The tell is whether they understand how AI plays out across the whole organization - across ops, sales, marketing, finance, HR - not just within one function or tool. Someone with deep AI knowledge but limited business experience will solve the wrong problem or solve the right problem in a way that creates problems downstream.
Nicole's four-D consulting process starts with strategy before tools. The team will walk through the entire organizational workflow, identify where AI and automation can replace the most repetitive work, and assess how changes in one function affect others. Her estimate is that for most knowledge workers, a thoughtful AI implementation can reduce the effort of the job to 30% of its current load or less - but that kind of transformation requires mapping the whole organization, not installing a chatbot in one department and calling it done.
The Business Model: Courses, Live Labs, and On-Site Intensives
AI Smart Insiders runs three overlapping revenue streams. The entry-level offering is open enrollment courses: the flagship four-week AI Your Agency / Ops course is $997 per seat, run live via Zoom one hour per week, and participants walk away with at least one piece of their workflow automated or AI-enhanced. Weekly AI Labs (open to the public one to two times per month) are where participants bring real business problems and build solutions live - the weather forecasting agent being a current example. The self-serve recorded course at $197 exists but sells poorly; live cohorts with human interaction are what people actually buy.
At the top of the offering stack is a ten-week deep-dive program at $5,000 that covers advanced agent building, full implementation methodology, and an Applied AI Certification for LinkedIn. On-site custom training engagements - where Nicole or a trained instructor travels to the client - start at $10,000 per day and represent the highest single-transaction revenue. Nicole stacks these travel gigs by geography when she's already traveling, booking engagements in Germany, Switzerland, Istanbul, Lisbon, Vietnam, and Thailand around existing trips.
From Baby Legs to AI: The Throughline Is Emotional Connection
Nicole's entrepreneurial path is unusual enough to be instructive: professional snowboarder → baby product founder → marketing agency → AI training company. The connective tissue is not the industry but the mission. Baby Legs was born from a parenting hack (cutting off snowboard socks to protect her daughter's knees while crawling) and grew to 85 countries in three years because Nicole was purely focused on making parents' lives better - and the product delivered an obvious, specific benefit that gift-givers could understand instantly.
The marketing agency came from caring about the emotional connection between brand and customer. The AI pivot came from discovering that AI, used correctly, could deepen that emotional connection rather than flatten it. Her framing: AI done right makes humans more human. It flips the 80/20 rule - where 80% of work is the grind you tolerate and 20% is the work you love - so that 80% is the meaningful creative work and 20% is the necessary overhead. That is the actual promise of the technology, and it is the reason Nicole's mission (spreading joy to a billion people) feels coherent across a career that looks disjointed from the outside.
Tools & Resources Mentioned
- AI Smart Insiders - Nicole's training and consulting platform; aismartinsiders.com. Free AI Labs every Wednesday, open enrollment courses, resources page, weekly newsletter.
- AI Smart Marketing YouTube - Weekly how-to videos on AI tools and implementation; free resource for anyone learning to apply AI.
- GoHighLevel / OS CRM - All-in-one CRM, LMS, pipeline, and automation platform Nicole uses for course delivery and student communication. Recommendation: always access via an agency (like OS) rather than directly - same price, much better support.
- Gemini Gems - Google's custom agent builder; outperforms ChatGPT on math accuracy in Nicole's testing.
- ChatGPT Custom GPTs - Better qualitative insights in Nicole's testing; recommended for knowledge-base-heavy agent builds.
- Farmer's Almanac - Recommended data source for full-season regional weather forecasting in agent instructions; Nicole prefers it over Weather Channel for accuracy.
- Perplexity - Nicole explicitly does NOT recommend it due to hallucination and incomplete answers.
- Claude - Mentioned as a fallback when ChatGPT or Gemini underperform on a specific task.
Frameworks
The House Renovation Analogy
Before implementing any AI tool, build the strategy for the whole organization first - just as you'd plan a full home renovation before starting the bathroom. Floors connect. Decisions made in isolation create mismatches you pay to fix later. Start small and low-disruption, but design the full roadmap before touching anything.
80/20 Flip
In most knowledge work, 80% of the job is the grind you tolerate and 20% is the work you love. AI done right inverts this: the repetitive, tedious work compresses to 20% or less, and the creative, high-judgment work expands to 80%. The result is a happier, more productive workforce - not a smaller one.
Test Natively Before Chaining
Before building multi-model workflows, test the same task natively on each model. Gemini is better at math; ChatGPT often produces richer insights. The right model for each step of a workflow only becomes clear through direct comparison - not vendor documentation or benchmark rankings.
The Small Business AI Window
Small businesses have a structural speed advantage over enterprises in AI adoption. Right now, that window is open: Google search volume is declining as AI search rises, large companies are still in approval mode, and the infrastructure is cheap enough for any size business to deploy. The window is temporary - companies that move this summer will have a compounding head start by Q4.
FAQ
Who is AI Smart Insiders designed for?
Anyone from independent consultants to large enterprises and government agencies. The open enrollment courses attract mixed groups from many different companies - participants learn from each other as much as from the curriculum. Custom training is available for larger organizations that need tailored implementation across specific workflows.
How do I know which AI model to use for a task?
Nicole's answer: test natively first. Build the same agent on Gemini and ChatGPT simultaneously, run them in parallel, and compare outputs. Gemini currently outperforms on math accuracy; ChatGPT often delivers richer qualitative insights. The right choice depends on the type of output you need - and it changes as models are updated.
What is a custom GPT or Gemini Gem, and is that an agent?
Yes. A custom GPT or Gemini Gem is a form of agent: a set of instructions plus a knowledge base that constrains and directs the model's behavior toward a specific task. You don't need a $750/month agent platform to build useful agents - the native custom agent builders in ChatGPT and Gemini handle a wide range of real business use cases.
What does Nicole mean by ‘AI your organization this summer’?
The macro environment is unstable; sales cycles are slow. This is the ideal time to invest in AI infrastructure rather than waiting for conditions to stabilize. Companies that spend the next 90 days implementing AI across their operations will have capacity built and compounding when the market recovers - while slower competitors are still getting started.
How should I think about AI and workforce impact?
Nicole's framing: thoughtful AI implementation can reduce most knowledge work to 30% of its current effort. The goal is not to eliminate jobs - it is to redirect human energy from repetitive work to creative and relational work. Employees trained to use AI well are consistently happier and more productive. The risk is not adoption; it is non-adoption.
What is the ‘cucumber’ referral?
If you heard Nicole on this podcast and want to connect with her on LinkedIn, say ‘cucumber’ in your message - she'll know you came from this show.