How We Use AI and How We Don't
AI Transparency Policy
Human judgment and machine speed with nothing hidden.
Every company Kelly O'Neil has founded is values-based. Integrity and transparency sit right alongside exceptional client service and excellence in everything we deliver. Trust with our market, our clients, and our team is of paramount importance to us, so we're putting our use of AI right out in the open.
Yes, we use AI. We use a lot of it, on purpose, with a lot of rules.
If you've ever wondered whether the expert you're paying is a chatbot with a nice headshot, this page is for you. Some people think AI is the devil. Others have handed their whole business to a chatbot and gone to Bali. We sit in neither camp. The FUSION™ methodology is built on integration and that includes this one: human judgment, intuition, and wisdom, paired with machine speed, each doing the part it's best at.
We will never be sneaky about anything, including AI. It's not our style. You deserve to know where the line is, so here it is.
1. Where Our Thinking Comes From
Every concept, strategy, and system that comes out of Phoenix Growth Partners starts in a human brain. Specifically, Kelly O'Neil's and the team of human rock stars who support her.
Kelly has spent more than 32 years in rooms most consultants only read about: boardrooms, war rooms, media rooms, analyst rooms, and investor rooms. She has helped the companies she advised generate more than $750 million, and we stopped counting in 2024. She has made tens of millions of dollars in her own companies and has run her own business for more than 25 years.
She has worked with thousands of clients across a long list of industries: technology, hospitality, professional services, medical devices, web and digital agencies, home builders, interior designers, lawyers, accountants, coaches, consultants, and many, many, more. That range matters. A pattern that looks new in one industry has usually already played out in another and Kelly has seen most of them up close.
Category of One positioning and business architecture for visionaries come from being on the ground inside organizations, often working with the same clients for years. Marketing to the affluent comes from Kelly's own lived experience. Everything we teach or advise on comes from both.
Kelly's strategic ability is part gift and part training. She was mentored by millionaires and billionaires in Silicon Valley's tech world and trained by some of the sharpest minds in its brand and PR agencies, analyst firms, and technology firms. She practiced what she learned, succeeded with it, and then brought it to the entrepreneurial market.
Kelly values continued learning. She has invested roughly half a million dollars in her own education and is still investing. She is also staunchly against lifting something from a seminar or a book and regurgitating it to clients. Nothing gets taught here until Kelly has perfected it herself and can translate it across every client and every industry we serve. Quite frankly, if we're being honest, she's most likely to go in the opposite direction of whatever she's been taught. She's a bit of a contrarian like that. The status quo has never been her thing and she'll likely vastly improve on whatever she's learned.
Our methods are our own. We built them because the business mentorship industry has a lot of room for improvement. There were several important holes in the market that entrepreneurs needed filled and nobody was filling them. Visionaries and innovators need those gaps closed to be epically successful. That's what we built Phoenix Growth Partners to do.
AI didn't invent any of that because it wasn't in the room.
2. The 25-50-25 Rule
“Give me six hours to chop down a tree, and I'll spend the first four sharpening the axe.”
That line is often credited to Abraham Lincoln, though he probably never said it. (See? We fact-check even the inspirational quotes.) The earliest known version came from an anonymous woodsman, and he had a point. We've spent years sharpening our axe, and every AI-assisted piece of work at Phoenix Growth Partners follows the same formula.
25%: Our original thinking: This part has nothing to do with AI. It's strategic and critical thinking, pattern recognition, trained expertise, life experience, and hard-won wisdom. Kelly is a futurist. This is her working in her gifts and her genius, backed by training and decades of applying it across many companies and every area of a business, mostly with great success, plus a few train wrecks she learned from the very hard and expensive way. It's seeing problems most advisors miss, and solving them before other people know they exist.
Kelly captures that thinking by recording it out loud, into Otter or straight into Claude, the same way she dictated every point in this policy. That original thinking becomes the context the AI works from.
50%: The sharpened axe: We took decades of original thinking, proprietary methods, and skills and baked them into our Company AI Brain and all of our LLMs (large language models), like Claude. It holds:
our detailed business knowledge
projects built with specific context to achieve specific outcomes, based on detailed instructions
proprietary skills encoded from the methods and processes we've used with clients for decades
25 years of real deliverables that are still relevant
deliberate and custom prompting
Then we programmed the AI to perform to our standards, with guardrails and fact-checking built directly into our AI solutions:
It must cite its sources, including when it goes out to the internet to research at our request.
If it can't produce a citation, it has to flag it.
We click the links and check every reference ourselves.
It's programmed not to make things up, and we check that it doesn't.
It's programmed to operate critically, challenge thinking, and keep improving.
This is where it does the manual heavy lifting, turning our original thinking into pretty little client-ready documents. Generic AI pulls from the whole internet. Ours pulls from us. It works because we spent years teaching it and we keep training it the way you'd train a high-performing human in your own organization.
25%: Human precision: We review, fact-check, edit, and polish so it's right before we ever turn it over. Kelly is fanatical about her voice, her words, and anything with her name on it meeting the top level of excellence. She won't even let her own team members reply to people in her name without reviewing it first, let alone an AI.
Quite frankly, our trained AI outperforms what a junior team member could produce, along with many senior team members and more than half the agencies we've ever worked with, because we trained it so thoroughly. That frees our human rock stars for the work only humans can do.
So when someone asks whether AI wrote something, the honest answer is that AI did about half the labor, using our thinking.
Every part of the process runs on our intellectual property, and the judgment at the end is ours alone.
3. Our Guardrails
We started learning AI before most people had heard of ChatGPT, and we never stopped. We take courses and trainings, and our team trains every week to keep our skills sharp for today's AI.
Beyond the checks built into every AI solution, our systems follow a few more rules:
Guardrails protect against hallucinations, the confident-sounding errors AI is known for.
Our systems won't use data or statistics that can't be documented.
Every data-based claim we publish must be citable, because so much of our work is built on research and data.
Our strategy systems give every answer a confidence score from 0 to 1, based on the strength of the evidence behind it. Below 0.8, the system has to name its weakest points, rethink, and tell us what information it's missing.
Every answer also lists its key caveats: the assumptions it made, the open questions, and anything we need to verify before we rely on it.
Our AI advisors are blocked from answering questions outside the scope of their training (as should most business coaches, but we digress), just like Kelly and our team won't answer questions outside our scope of expertise.
Several people test every AI system before it reaches you, including our Fractional Chief AI Officer.
We only load trainings that are relevant, current, and built to stand the test of time. We keep documentation of everything our AI is trained on, so when information needs updating or becomes irrelevant, we can make an immediate shift. We review our AI systems regularly and keep training them. We don't dump our entire Google Drive into AI and hope it lasts. Some people brag about dumping 17 years of their knowledge, trainings, and products into their AI, and you could probably ask it how to market a teleseminar and get a plan. Ours won't do that.
On top of all of that, human eyes are on everything. Qualified human review means Kelly or a trained member of our team reviews everything before it goes out.
Even so, AI makes mistakes, just like humans do. Think of it as a brilliant junior analyst who never sleeps and occasionally lies with total confidence. That's why it has a boss. It takes someone with real discernment to tell whether the output is accurate, and that discernment is what we sell. Our business is built on judgment, advisory skill, and wisdom. Scraping the internet and tossing the results into a post has never been our model.
Some skeptics might wonder whether a firm that builds AI solutions will push AI on its clients. We recommend AI only where it serves your business, the same way we recommend anything else. Our goal is to positively impact humanity. AI helps us serve more genius visionaries and innovators who need to get their work out into the world than we ever could without it. It also gets every person on our team back into their Genius Zone (we call it the FusionZone™, which is the Genius Zone with a big twist). It's the philosophy our company is built on.
4. What AI Does For Us
Research and intelligence: Strategic positioning takes serious market research and competitive intelligence. That work used to take weeks. Our trained systems now produce, vet, and fact-check it in minutes, and a qualified human checks it again before anyone relies on it. Clients get better strategy, faster.
Formatting and production: We've programmed AI to build client deliverables in our own formats, using our proprietary methods. It handles the assembly work a junior analyst used to do. The thinking stays with us.
Turning Kelly's brain into content: Kelly records her thinking out loud and tells AI exactly what goes in. AI turns her dictation into polished writing. That's exactly how this policy was made: Kelly dictated every point, including what we stand for, what we believe, and why we're doing this in the first place, and AI assembled it. The ideas, opinions, and jokes are hers. (For the record, Kelly is neurodivergent and was using em dashes long before they became an AI tell. Her editors have confiscated them, and she'd like them back.)
AI photos: We used to do a photo shoot every 45 to 90 days to keep fresh images coming. Now we create AI images built from real photos of Kelly and our team. We keep them as close to real life as humanly possible. AI does tend to make us look like we just left the spa after a great hair day with a stylist and a makeup artist, and we've made our peace with that.
AI video avatars and voices: Kelly would rather poke a pin in her eye over and over again than film 30 reels a week. So some of our reels and trainings use an AI video avatar of Kelly, and some trainings occasionally use a clone of her voice, delivering content she wrote and has taught for decades. After 25 years of teaching this material, we don't need to reinvent the wheel, and Kelly doesn't need to re-record Module 4 for the ninth time. Our leaders and brand ambassadors would rather spend that time with clients than hire a spokesperson to pretend to be us.
Kelly's long-form content is usually Kelly, live. Reels and trainings are a mix. We'll do our very best to label every AI-generated piece, and you'll usually be able to tell anyway, because it'll look like AI.
AI advisors: We're building a suite of AI advisors trained on our original works, systemized solutions, and lived experience. Every one is introduced and labeled as AI. Each has guardrails that keep it inside its training, so when you ask something outside its lane, it's built to say “that's not mine to answer.” Each advisor also keeps memory for its client, so it works the way your own AI would, remembering your context from one conversation to the next.
AI assets on our team: We're building a full team of AI assets under the guidance of our Fractional Chief AI Officer. Anything repetitive that doesn't require strategic thinking or a human touch gets delegated to them. We're in the business of solving problems. The less busywork and admin shit we do, the more time we have to leave a legacy on this planet.
Tier-one support and sales help: AI answers basic questions 24/7, even while we're asleep, and a human always sits behind it. Clients also have a direct text line and email to our support team. High-end relationships are run by humans. Convenience is handled by AI.
5. What AI Never Does For Us
It never does our thinking. Strategy, judgment, and advice come from humans.
It never sends a client deliverable unreviewed. Nothing leaves our office without qualified human review, and we often work a deliverable over many rounds until it's chef's kiss perfect.
It never handles client financial data. Financials, account details, and private records stay out of AI.
It never sees your trade secrets in the open. Many of our clients hold trade secrets and sign serious NDAs, so at their request or our suggestion, we change client names, products, and identifying details before anything goes into an AI system.
It never trains on our work or yours. Our AI tools are set so our data is not used to train models, which keeps our IP and yours from leaking into the market.
It never pretends to be human. If you're talking to AI, we tell you.
It never invents testimonials, statistics, or numbers. Every client testimonial is real and approved by the client who gave it.
6. How We Handle Your Information
We currently use Perplexity, Claude, and ChatGPT, with settings that keep our data out of model training.
Call recording. We currently use Fathom, and sometimes Zoom, for AI note-taking and transcription on sales calls and client calls. Both are visible on the call, so you'll always know they're there.
Intake forms. Your intake answers go through AI, including your birth date. That's how we build your FusionCore™ DNA analysis.
AI advisor memory. Our AI advisors remember each client's context, so they work the way your own AI would.
Team accounts. Team members are not permitted to use personal AI accounts for client work. Client work stays inside company systems.
Your client agreement spells this out, including your consent to session recording and AI processing of intake materials. You can ask us to delete your materials at any time by emailing [email protected].
7. When Something Goes Wrong
Errors happen, whether from AI or team members. Yes, even sometimes Kelly, and she'll be the first to admit it. We take every one seriously. If you think you've found an AI error from us, tell us right away at [email protected] and we'll fix it.
Responsibility for our advice, including anything AI helped produce, is governed by your client agreement, the same as any advice we give.
8. What You're Paying For
Clients pay for results, and those results come from decades of experience. AI lets us deliver more, faster, and that means you get to market faster. Ultimately, it also makes our work a bit less expensive for you, because we can deliver more of what you need to create the results you deserve.
AI is built into how we work, so we don't offer an AI-free version of our services. Ownership of deliverables follows your client agreement.
9. If You Disagree
We fully respect that. You don't have to like how we use AI. We just won't hide it. This policy is linked on all our websites and social channels, so you always know where we stand.
10. What Changes and What Doesn't
AI is moving at warp speed, and what's cutting-edge on Monday is old news by Friday. We reserve the right to update this policy as the technology changes, and we will.
This policy covers Phoenix Growth Partners. Our partners and affiliates have their own AI policies. When this policy changes, we update this page, and private clients hear from us directly.
A few things won't change, no matter what the robots learn to do next:
exceptional work
exceptional client care
fierce protection of your privacy and trade secrets, and of ours
That's how we roll.
Kelly dictated every point in this policy, AI assembled it, and it took thirteen rounds of revisions to get it just right. That's 25-50-25 in action.
Last updated October 12, 2026.