
Creator's MBA · 2026-05-06 · 24 min
Key moments - from our scoring
Substance score
50 / 100
Five dimensions, 20 points each
Savannah Gilbo runs Notes to Novel, a signature course and membership community for fiction writers that exploded after a successful August 2025 launch brought in nearly 200 new students. Her live coaching calls ballooned from planned 60-minute sessions to 90+ minutes, and her weekly membership deliverables - masterclasses, office hours, and community calls - doubled her workload to potentially 48+ hours monthly. Rather than hiring a team manager (which she'd ruled out after corporate experience), she and host Dr. Destini Copp spent four weeks building an AI clone using Delphi AI. The breakthrough came through developing what they call a "triage framework" - training the AI to ask diagnostic questions (genre, character details, theme) before answering, mirroring Savannah's actual coaching instinct to probe before prescribing. Beta testing with 10-15 membership users revealed students asking 400+ messages per conversation, treating the AI respectfully, and crucially, arriving at office hours with higher-level questions after pre-processing with the clone. The result: reduced mental and emotional load, student confidence building, and 24/7 availability that let writers work at midnight without waiting four days for office hours. Savannah plans to roll it out to her evergreen self-study package.
Train it on your actual decision-making process by asking yourself 'How would I answer this on a live coaching call?' and documenting the triage questions you'd ask first (like genre, theme, character details in Savannah's case) before giving advice, rather than just uploading course materials.
It depends on your preferences: Savannah chose an AI clone because hiring would create management responsibilities she disliked, but hiring might suit others better; the key is recognizing that both solutions have trade-offs beyond just cost.
Savannah's beta users responded extremely positively once they understood the AI was for coaching strategy and process, not writing their books; they used it for pre-work before office hours and appreciated the 24/7 availability and judgment-free environment to ask basic questions.
The AI will likely answer questions too advanced or generic compared to your actual coaching - you'll need to teach it the diagnostic questions and context-gathering steps you normally do before diving into answers, which requires reflecting deeply on your implicit coaching process.
Yes, if built correctly with a triage framework; it handles lower-stakes questions and builds student confidence between office hours, letting you focus synchronous time on deeper strategic work while students get 24/7 supportive guidance.
Our reviewer’s read on each dimension, with quotes from the episode.
The episode offers some practical specifics about AI clone implementation (triage framework, beta testing with 10-15 users, seeing 400-message conversations) and identifies a genuine insight: students ask clarifying questions to the AI before office hours, improving office hour quality. However, much of the content is conversational setup and emotional affirmation rather than actionable insight. The core insight - training AI to ask diagnostic questions rather than give answers - is valuable but thin relative to runtime.
The AI would answer something five steps ahead of where I knew the writer needed to be... we had to figure out how to get AI to back up and ask the question
people were using it to ask questions before coming to office hours... they would go off and edit it. Then they'd come to office hours and say, 'Hey, I worked with the AI'
The specific application of the triage framework to coaching (Socratic questioning via AI) is somewhat novel within the writing coaching niche, but the broader concept of building AI clones is well-trodden territory by 2025. The insight that AI should mimic a coach's diagnostic process rather than provide immediate answers is sound but not groundbreaking - it's a straightforward application of pedagogical principles. The episode adds little contrarian or first-principles thinking.
you really helped me walk back to 'How does my brain work before the answer comes out of my mouth?' That is then what we had to train the AI to do
The foundational triaging of questions is really what makes me special as a coach, and getting that into the AI was a really big deal
Savannah is a relevant practitioner with genuine operating experience: she runs a profitable course (~200 students, recurring membership, multiple product tiers) and has hands-on experience building and deploying an AI tool. However, she is not exceptionally senior or unusual in the creator economy space, and the episode is partly a vendor testimonial (Destini helped build the clone and is promoting her services). She is credible but not a rare or elite operator.
In my last live launch, which was in August of 2025, we had our best promo and best launch ever. We welcomed so many new students. It was just under 200
I'm really regimented in my schedule. I love planning things, slotting things, recording my own podcast, speaking events
The episode provides some concrete details: ~200 students in a launch, calls extending from 60 to 90 minutes, workload doubling, 10-15 beta users, one user with 400 messages in a conversation, and membership at 4-8 hours/month of deliverables. However, it lacks hard numbers on impact (time saved, revenue impact, student retention metrics) and mostly uses anecdotes ('one person said', 'I remember at one point'). The triage framework itself is described conceptually rather than with detailed examples of actual outputs.
In my last live launch, which was in August of 2025, we had our best promo and best launch ever. We welcomed so many new students. It was just under 200
I was probably spending double that once the course kicked off... Instead of four hours a month on deliverables, it was turning into six or eight hours
Destini asks basic follow-up questions ('paint a picture', 'what did your week look like') and allows Savannah to expand, but rarely pushes back, tests claims, or drills into contradictions. The interview is warm and collaborative but lacks sharp questioning. Destini asks leading questions aligned with her consulting positioning ('do you feel like it's decreased your work?') rather than skeptical ones. There is no pushback on whether the triage framework is actually novel or whether the time savings are real.
It brings me to our next question. What kind of feedback have you gotten from your students?
Do you feel like it's decreased the amount of extra work that you have?
Computed from the transcript - who did the talking, and the words that came up most.
Send us Fan Mail Most digital product creators assume that scaling means hiring more team members. More support staff, more coaches on calls, more people to manage. My guest today took a completely different path. After Savannah Gilbo's best launch ever brought in almost 200 new students, her live coaching calls were running 30 minutes over. Her membership deliverables doubled. She was burning out fast. Instead of building a team, we built her an AI clone together. In this episode, Savannah and I walk through exactly what broke after her biggest launch, the options she considered (and why hiring more people felt worse, not better), the triage framework that made her AI clone actually coach like her instead of giving generic AI answers, and what her students are saying now that they're using it. If you've been curious about AI clones for your business but worried about whether it would actually work for your niche, sound like you, or feel right for your audience, this episode is for you.
Transcribed and scored by The B2B Podcast Index.
Intro DESTINI: Welcome back to the Creator's MBA Show. My name is Dr. Destini Copp, and I'm super excited that you are joining me here today. So here's the thing.
Most digital product entrepreneurs assume that scaling means hiring more team members, more support, more people on calls. But my guest today took a different path. After her best launch ever brought in almost 200 new students, Savannah's live coaching calls were getting out of control. Her membership deliverables were doubling, and she was burning out.
So instead of building a team, we built her an AI clone together. In this episode, we're walking through what broke first, what we tried, the triage framework that made the clone actually work like her, and what her students are saying now. Let's get right into it. Meet Savannah DESTINI: Savannah, I am super excited to be chatting with you today, and we're going to be digging into your AI clone that you and I developed together.
But before we get into all that, can you tell the audience a little bit more about you and what you do? SAVANNAH: Yes, thank you so much for having me on the show. So I help writers in a few different ways. One of them is through my podcast.
That's my free content. My signature course is called Notes to Novel, which really helps people take all their messy, shiny ideas and turn them into a first draft that reads more like a third or fourth because they've done all the hard work upfront developing their ideas. After that, I invite graduates into my membership, which is where ideally they're going through the next stages: revising, getting ready to publish, marketing, and then hopefully starting a new book right after that.
So that's the world I operate in, fiction stories all the time. That's my bread and butter. DESTINI: I'm super jealous of the work that you do. That would be my dream to live in that type of world all day long.
That is on my bucket list, to do a book like that five years down the road. Painting the Picture: The Breaking Point DESTINI: So before we get into the AI clone that we worked on, I'd love for you to paint a picture for the audience of the scale you were operating at. How many writers were you supporting? What kind of support were they getting from you?
SAVANNAH: Okay, so this all surrounds what was also my breaking point. To paint you a picture, my Notes to Novel course is offered self-study evergreen, and it's offered live where I do live cohorts two to three times a year. We go through the course for eight weeks, and there's coaching and other elements. So that's one silo.
Then on the other side is my membership where there's a weekly deliverable from me. Week one we do a masterclass. Week two and four we do live office hours. In the middle, week three, we do a community get-to-know-you call.
In my last live launch, which was in August of 2025, we had our best promo and best launch ever. We welcomed so many new students. It was just under 200. Going into that, I was thinking, "This is great," but I was a little worried.
How am I going to support almost 200 people by myself? Luckily, it was fine. One of the things I do a good job at is staying on the calls until all the questions get answered. Most of the time, that meant my 60-minute calls would turn into 90-minute calls.
Students were really happy with that, but I was getting more and more drained. Not because I don't like interacting with people, but for course creators, sometimes it's hard to be on for longer than you planned, and then that messes with your schedule and other things. That was the event that kicked the ball and made it roll downhill. All those graduates, we had our best promo yet and the best conversion into the membership, which made those calls and deliverables balloon.
I started getting worried about how I was going to show up and serve everyone and give everyone the time and attention they need. That's what made me start looking at, what am I going to do about this? DESTINI: What did your typical week look like? How many hours do you feel like you were supporting?
It sounds like you were doing a lot. SAVANNAH: Let's go back to that time period where I was delivering it live and I had my membership. If I was meant to spend eight hours over the course of the live cohort, so eight live hours, one call a week, then I was also spending time in the community answering questions. It was, like, 24 hours a month just for this live deliverable.
But that ballooned very quickly. With more students, what I had done in the past wasn't going to cut it. I was probably spending double that once the course kicked off. Then, once the eight weeks were done, students came into my membership, and that ballooned too.
Instead of four hours a month on deliverables, it was turning into six or eight hours. I'm really regimented in my schedule. I love planning things, slotting things, recording my own podcast, speaking events, whatever. All of it's on my calendar.
When things don't go according to plan, it causes problems in more than one way. So that's what led to the explosion of "I need to do something about this." DESTINI: It was almost doubling your workload, really. SAVANNAH: Almost.
Especially when I was live. Everything else was manageable until that one cohort where we got so many new students. That's a great problem to have, right? But I wasn't ready to have a solution on the back end yet.
Why Hiring Wasn't the Answer DESTINI: What other options did you consider? You were like, "Okay, this isn't working. What else could I do?" SAVANNAH: I had seen other people run communities and run live events where they had a community manager, student ambassadors, things like that.
So I toyed with all those ideas, then I kept coming back to the realization that if I get other people involved, that's me managing people. That's still going to feel the same way to me as showing up for extra live calls or extra time. It's going to net the same result for me. So that didn't really work.
It wasn't about budget necessarily. The business could afford to hire somebody. I just didn't want to manage additional people. At the same time, I happened to partake in someone else's bootcamp where they were offering an AI clone as an upgrade, so I experienced it as a user.
I was like, "This is the coolest thing I've ever seen," because I literally felt like I was talking to that person outside of her bootcamp. DESTINI: You were very smart to realize up front that hiring other people wasn't necessarily going to decrease your workload. SAVANNAH: I can feel that. In my past life, I worked in a corporate office and I had to manage people, so I know what that's like.
That's such a big draw for why I want to run my own business. I don't have employees. I work with contractors right now. It's just not the part I enjoy, the people managing.
The Fears About Building an AI Clone DESTINI: So you experienced it through this other program. Did you have any fears about using it? What were your concerns? SAVANNAH: I might be a little bit of an outlier in the sense that tech stuff doesn't scare me.
Before I met you, Destini, I was playing with a version of my own clone, and I just couldn't get it to work. I think there might not have been a lot of great resources. I looked up all the things: How do I build this clone? What kind of things do I feed it?
I'd get really mixed messages from one place saying, "Just upload everything you can think of," and then other places would say, "You have to upload small amounts in a specific order." I just felt, "Okay, this is going to become a full-time job if I don't know what I'm doing." So that was one of my fears: I see the benefit. I'm not worried about the tech, I just don't know if I can spend the time I need to figure it out right now.
Then there's a whole other side of it. In the writing world, because it's such a creative space, people are very anti-AI for creative writing. There are a lot of great things writers use AI for, like helping with their book marketing or doing research. But any time you start to cross the line of "this could potentially start writing a book for someone," people will shut down.
They don't want to hear it. So I was also thinking, "What if I spend all this time, and then I'm either unable to express how it could help someone in a way that's not going to steal your creative writing power, or people just don't want it at all because it's AI related?" DESTINI: I could see that being a huge fear on your end. So you build this, and is anybody going to use it?
We're going to get into that in just a second. So you had tried it on your own. You got stuck. It wasn't giving you the results that you wanted.
Then you found me. I think I was on your email list, and I had sent out an email about helping some people work through their AI clones, and you were like, "Okay, this is for me." SAVANNAH: I'm really good at researching too. I don't let things stop me.
If I have a question, I go to Google or AI, and I'll typically find the answer I'm looking for. This is an area that stymied me. Maybe because at the time I tried, maybe a year before I met you, it was, like, early adopter time. Maybe there weren't a lot of resources out there, and I just wasn't able to find the answer or get the help that I needed.
You were coming at it like, "I've done this, and I know how to make it work." When you sent that email, there were specific pain points where I'm like, "Yeah, that's exactly what's going on with my AI clone." The Four-Week Build DESTINI: So you and I worked together. We spent about four weeks setting it up, testing it, doing the beta test with your students.
Now I would love to know, let's talk about how you launched it to your community. How did that go, and what were some of the struggles we had initially? SAVANNAH: There's something that your listeners might find interesting. I was a little worried because I'm like, "Okay, I feel like we all feel like we're snowflakes in a way.
Is this going to work for my niche?" In my niche, there are so many genres, and one of my big things I teach is the nuances per each genre, and a lot of rules that you'll hear in the writing world that differ depending on the genre you're writing. On one hand I was like, "Is the AI going to be able to capture the experience I have and the nuance that I know about coaching writers?" But on the other hand, "Is Destini going to be able to understand what I'm trying to do and what I'm trying to say?"
I was a little worried about that. But spoiler alert, it worked out totally fine. One of the things that I actually loved was that you were asking me questions from a place of "I don't know your niche, but you do." It really made me think in a different way that I wasn't able to do on my own.
I don't know if I ever told you that, Destini, but that was really helpful because you were asking me things that I'm like, "I haven't thought about it that way in 15 years." DESTINI: It's true. I really don't know your niche. I know how to teach.
I've been a teacher, a university professor for 20-plus years. But in terms of your niche and how you would coach students, that's where you have the expertise. SAVANNAH: A lot of what we went into is "How would you answer this if you were in a live call with your students?" The Triage Framework Breakthrough DESTINI: You might want to talk a little bit about that.
SAVANNAH: Yes, that was really funny in hindsight because you kept asking the same question: "How would you, human Savannah, answer this?" You really helped me walk back to "How does my brain work before the answer comes out of my mouth?" That is then what we had to train the AI to do. There was a lot of stuff that the AI got right or close to right pretty quick, just based on me feeding it my course materials, and then you also helped me write the instructions for how it was supposed to think.
But one of the things we ran into, probably the biggest issue we ran into, was that the AI would answer something five steps ahead of where I knew the writer needed to be. Every time you would say, "What would you say to the student?" my answer usually was, "I would probably ask them a few questions before I even gave them an answer." So we had to figure out how to get AI to back up and ask the question to get to a place where it could even answer the student.
For example, if a student's asking the AI, "How do I show my character's wound on the page?" The AI could, in theory, give you a very generic answer like, "Oh, it needs to show up in the actions they take." That's a very general thing you could just get on Google. It's not a special answer, and it's not untrue either.
But if I were the coach, I would say, "Okay, let's think about this. I want to know what is your genre, what is your theme, because that plays into my answer to you, and also, what is your character's specific wound?" If you're writing romance and the character's wound is around trust, or you're writing horror and your character's wound is around trust, that's going to make a difference in how I answer you. So we had to teach the AI how to triage questions, and then it can't move forward until it either gets the answer from the student or gets an "I don't know" answer.
If it got an "I don't know," there's a whole other pathway it takes to help them develop those foundational things first. DESTINI: I just want to let everybody listening know, Savannah did that hard work. You went back and did all of those triage instructions for the AI. It is not anything I could have done because it was really your brain thinking through, "If I was on a live coaching call with the student, how would I handle it?"
That's what made the biggest difference for the AI. Do you agree? SAVANNAH: Totally agree. There were little things we caught here and there, like maybe it would get a term slightly incorrect to how I would use it, and those were really easy to pick up and fix.
But the foundational triaging of questions is really what makes me special as a coach, and getting that into the AI was a really big deal. Now, I used to worry, "Am I doing a disservice to my writers because they're getting an AI output that, yeah, it's based on my course material, but in theory they could take all this and put it in AI and get a similar answer?" Now that we've built in the triage system, I'm like, "No, they are getting the benefit of having me also as their coach," because it's the same thing I would ask them.
I love that. I wouldn't have known how to think about it that way if you weren't constantly asking, "How would you do it? Here, I think we need to teach it the process of how you would do it." I probably would have just thrown more information at my AI.
The Beta Test and Student Feedback DESTINI: It brings me to our next question. What kind of feedback have you gotten from your students, and how exactly are they using it? What I love about the tool we use here, which is Delphi AI for anybody who's listening, is you can actually see the conversation, so you know if the AI is answering the right way. SAVANNAH: We had a beta group from my membership because I knew those were the most AI-friendly.
They would also be willing to do a beta and know it wasn't going to be perfect. We had about 10 to 15 users as we went out with this. We tried to set up some rules, and this is just for anyone thinking about this. We set up some guidelines, like it's not going to be able to answer questions about the writing of your specific scene, and people would still ask it questions like that.
I still haven't figured out the best way to handle that upfront. Anyway, the question was: what kinds of things were people asking it? We could see on the back side. As part of the training, we had to zoom out and say, "Okay, what are people actually going to ask, and is it picking up the right things, or is it not picking up the right things?"
That was how we found the need for the triage system we built into it. Then there were other people. I remember at one point we said, "Wow, one user had 400 messages in a conversation with it," and it was so fascinating to read through. I don't know if this is true of my niche or all niches, but my students were so nice to the AI, saying please and thank you.
It was just really interesting to see, "Okay, what are they asking? What are they getting back?" That helped us fix it or make it better. The best feedback I ever got was just one person said, "I really felt like I was talking to you."
That's the goal. The other really surprising thing that happened: people were using it to ask questions before coming to office hours. Let's say I'm a student and I upload part of my outline and I say, "Can you give me feedback on these three scenes in act one?" The AI now, because we've trained it to triage things, says, "Okay, let's back up.
What's your genre? What's your theme? Tell me a little bit more about your story so I can give you more specific, actionable feedback." The student would get the feedback on their three scenes of the outline, and they would go off and edit it.
Then they'd come to office hours and say, "Hey, I worked with the AI. I got it to this point. Do you think I'm good or is there anything else you might see?" There might be a little thing where I was like, "Okay, just because I know typical challenges that thriller writers fall into, just keep an eye out for this, but overall you're good."
So it was actually helping them bring higher-level questions to our office hours. In my world, I'm like, "That's what I want." This is not meant to be like, "Ask me the basic questions." You can, of course, but I would rather help you with the deeper stuff so that you can actually get your goals accomplished.
DESTINI: I feel like you saying that made this whole project worth it. Just getting those higher-level questions in your office hours, that is so very powerful. Do you feel like it's decreased the amount of extra work that you have? Mental and Emotional Load SAVANNAH: I do, because it almost feels like it's taken some of the mental load or emotional load off my plate.
When people are trying to figure things out, they have a lot of emotions too. If they can go to a more neutral AI that's still supportive and helps them work through their thoughts, they don't have to worry about coming to office hours and being embarrassed about asking a question. I'm a very sensitive person, so I will pick up on that too, and I try to hold space for the emotion. Now, the lower-energy stuff can happen with the AI, and they like that because they're not feeling exposed or vulnerable for not knowing what they don't know.
In a way, it's rewarding for them and for me to be like, "Yeah, you did that, good job. You are ready to move on to the next phase of whatever you're doing." DESTINI: It sounds like it's giving them more confidence too. SAVANNAH: Yes.
It's fun because it almost feels like they're coming and they're proud of their work versus, "I don't feel good about this and I need you to help me get out of this rut." It's almost like they're doing that with the AI. Of course, it's fine if they come to office hours and do that with me, but I like that they can do that when they need it versus waiting four days to come to office hours. If they're working at midnight on a Saturday night, they can keep going.
Advice for Creators on the Fence DESTINI: Any last-minute tips for the audience? If they are thinking about this or maybe on the fence or trying to figure out, "Would this work for me?" what would you say to them? SAVANNAH: I would say definitely give it a shot because, A, you deserve to have that time back in your calendar.
And if you need help, like obviously people like Destini are there to help. Destini was great to work with and very easy. It made the stress of it go away. I really appreciate you for that, Destini, because it is something that can be overwhelming if you let it be.
Just having Destini there to slow you down and not have you get overwhelmed and also say, "This week we're going to do this. Next week we're going to do this. Let's worry about one chunk at a time," was really helpful. If you need help, go find someone who can help or work with Destini.
And don't let the tech scare you because even if you're not tech-friendly, A, there are people to help, and B, things are just getting smarter and easier every day. By the time you listen to this, onboarding into your AI clone is probably way easier than it was six months ago when Destini and I did it. DESTINI: Isn't it crazy how fast tech is changing? SAVANNAH: It's crazy, but it's also cool.
I'm so into all the AI developments. I would say just go for it because your time is so worth it, and you deserve to do the things you want to do. Where to Find Savannah DESTINI: Savannah, can you let folks know where they can find you? I know there are a lot of people listening that have this on their bucket list too, just like me.
Let folks know where to find you, and I believe you have a free gift for them also. SAVANNAH: Yes. If any of you are fiction curious and you want to write a novel, come find me at savannahgilbo.com.
I also have a free quiz. If you are fiction curious and you're like, "Okay, where would I even start?" or maybe you have started something and it's been a couple years since you touched it and you're stuck, my quiz will tell you and give you resources. "Okay, you are here.
Here's what you need to do to get to your next milestone," whether that's finishing a draft, starting a draft, whatever. So go take the quiz. I know Destini's going to put the link in the show notes for that. DESTINI: Yes, I will definitely have it there.
Savannah, thank you so much for joining us and sharing all of this with us. I'm so happy that this is doing so well for you. SAVANNAH: Thank you so much. Maybe we'll have to chat again because one thing I didn't say is that we are about to roll it out to our evergreen Notes to Novel package, like, "Here's the benefit of doing the self-study version."
TBD on how that goes, but I will keep you posted. DESTINI: Definitely keep us posted. Outro DESTINI: Thanks for listening all the way to the end. I hope you enjoyed this episode today.
If you love the show, I'd appreciate a review on Apple Podcasts or your favorite podcast platform. Have a great rest of your day, and bye for now.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.