Ep: 38 - Making Nonprofit Work Easier With Mollie Williams

Summary

---

Podcast Summary: The Amazing Intern - Turning Your A.I. Curiosity Into Confidence

Ep: 38 - Making Nonprofit Work Easier With Mollie Williams

Host: Cary Weston
Guest: Mollie Williams
Date: June 12, 2024

---

Overview

This episode centers on how AI, specifically ChatGPT, can be leveraged to make nonprofit work more effective, efficient, and manageable. Host Cary Weston welcomes Mollie Williams, Executive Director of The Family Van and Mobile Health Map at Harvard Medical School, to discuss her real-world experiences adopting AI in her work. The conversation covers overcoming initial skepticism, practical applications for nonprofit administrators, responsible use, and tips for maximizing AI’s value.

---

Key Discussion Points & Insights

1. Mollie Williams’ Background & Role

• [08:57] Mollie has a 25-year career in public health, focused on increasing healthcare access and equity.
• Leads two programs:
  • Family Van: Mobile health clinics reaching underserved communities in Boston.
  • Mobile Health Map: A national collaborative network for mobile clinics, focusing on evaluation and research.

2. What Is a Mobile Clinic?

• [09:49] Mollie explains:

  "It is literally a clinic on wheels. Imagine a Winnebago, outfit the inside with exam tables, maybe even separate exam rooms, a waiting space, a lab space... They typically travel to underserved neighborhoods to get health care to people who wouldn't otherwise get that care."

3. Unique Nonprofit Challenges

• [11:20] Discussion of the "nonprofit hamster wheel"—the constant cycle of raising funds to launch programs, which then require more fundraising.
• Mollie shares her drive:
  • Inspired by the HIV crisis in the 1980s and witnessing health inequities.
  • Believes "everyone should have access to the health care that they need."
  • (See also Mollie’s activism spark at [11:43].)

4. Journey Into AI: From Skepticism to Enthusiasm

• [13:59] Initial exposure to AI came from her boss and sister (who runs a small nonprofit):

  “I thought she was exaggerating... but I realized my boss was not. Not just about what ChatGPT can do, but the impact that generative AI is already having and will undoubtedly have on our future. And I was instantly hooked…” — Mollie ([13:59])

• Now uses ChatGPT "just about every day" for both professional and personal tasks.

5. Family & AI Curiosity

• Mollie’s 8-year-old son is intrigued by ChatGPT and plays with it after watching YouTube demos ([15:56]).
• Comparison to Cary’s kids, who span ages and exposure/attitudes toward AI ([15:21]).

6. Practical Nonprofit Applications: How Mollie Uses ChatGPT

A. Fundraising & Grant Writing

• [17:35] ChatGPT helps:

  • Format and adapt “formulaic” components of proposals for different funders.
  • Tailor objectives as “SMART”: Specific, Measurable, Achievable, Relevant, and Time-bound.
  • Break down requirements:

    “So when it comes to completing that grant application, I can tailor things that I already have... Give that to ChatGPT and say, make these smart.” — Mollie ([17:35])

• Grant Title Brainstorming ([17:35]):

  “One of my favorite ways to use it for a grant proposal is to help me generate a title... I'll give ChatGPT the grant and say, come up with five possible titles... typically I would have done that in-person with a big team... ChatGPT is my brainstorming partner.”

• Request for Proposals/Evaluation Plans ([20:43]):

  “I'll cut and paste just a section... and say, give me ideas for this evaluation plan, or write me an evaluation plan. I find it works better when I do it in chunks... rather than everything all at once.”

B. Document Review & Custom GPTs

• [22:08] Attempts to build a custom GPT for SMART objectives, but struggles with getting the desired output.
• Cary offers troubleshooting advice: break up instructions stepwise (Step 1, Step 2, etc.) and ask ChatGPT for help generating GPT configurations ([23:03–26:17]):

  “Do some sort of exercise in regular ChatGPT and then ask it for the configuration output. You know, what instructions would I give a custom GPT if I were to do this?” — Cary ([26:17])

C. Thought Partner for Research & Writing

• Uses ChatGPT for planning research methods; explores things like content analysis and web-scraping ([27:46], [27:53]).
• Relies on ChatGPT to format citations but is cautious—verifies each result as ChatGPT sometimes fabricates references ([31:12]):

  “I don’t ask it to give me citations. I find them myself. But it can then help me to format them.” — Mollie ([27:53])

D. Daily Admin, Emails, and Data Cleaning

• ChatGPT as a “thought partner”—drafting sensitive communications, brainstorming solutions for everyday nonprofit tasks ([16:48]).

7. ChatGPT as Writing & Interview Assistant

• Voice vs. Typing ([32:25]):

  • Mollie finds typing helps her structure thoughts; speaking feels unnatural for drafting.
  • Cary prefers speaking for “getting everything out,” highlighting different working/thinking styles.

• AI’s Effect on Critical Thinking ([34:07], [35:08]):

  • Mollie raises concern about ChatGPT “stunting” kids’ ability to think critically.

  • Cary agrees:

    “If there's one thing that I try to share... [it’s] that there are—... the value of our firm is not the deliverable... it’s being a thought-provoking, problem-solving partner.”

  • Relates to calculators and GPS:

    “Are we using this tool to fast-track the busy work so we can pay attention to the valuable, or... to get through it all together and not think about it at all?” — Cary ([38:23])

8. “Interview Me” Use Case

• [39:22] Mollie uses ChatGPT to mock-interview her in preparation for grant subcontracts:

  “I told ChatGPT, ‘I have this opportunity; I need to submit a scope of work. This is the audience. Ask me questions one at a time until you have enough information to put together a detailed and thorough scope of work.’ And it did it again and again, and almost every time it was 90% there.”

• She appreciates being able to start responses on the phone and finish on her computer ([43:05–43:24]):

  “That chat was on my home computer, and I could finish it up, cut and paste it, and email it to the group that had requested it.”

---

Notable Quotes & Memorable Moments

• On starting with skepticism:

  “I was like, she... is exaggerating. Like, this is probably not as big of a deal as what she says, but I’ll have to check it out. And then... I was instantly hooked and couldn’t stop trying to figure out different ways to interact with it.” — Mollie ([13:59])

• On AI as a “team member”:

  “I don’t have the luxury [of a team brainstorm] anymore... so ChatGPT is my brainstorming partner.” — Mollie ([17:35])

• Cary’s ChatGPT philosophy:

  “If there’s one thing I try to share... [it’s] that there are—I have an agency... the value of our firm is not the deliverable... but the ability to listen, relate, problem-solve... if we’re using any tool just to produce in volume and get it done, then we’re missing the whole point of why we’re here.” — Cary ([35:08])

• Advice for Beginners:

  “You can’t break ChatGPT, so play around with it as much as you can and don’t be afraid to experiment... the more time you spend with it, the more comfortable it will get and you’ll move very quickly into really relying on it.” — Mollie ([44:27])

---

Important Timestamps

• 08:57 – Mollie’s professional background and current roles
• 09:49 – Explanation of mobile health clinics
• 11:20 – The nonprofit “hamster wheel”
• 13:59 – How Mollie discovered and adopted ChatGPT
• 15:56 – Family’s engagement with AI
• 17:35 – Grant writing, titles, and proposal adaptation using ChatGPT
• 20:43 – Breaking up input for ChatGPT and tackling requirements section by section
• 22:08 – Struggles and strategies with custom GPTs
• 27:53 – Using AI for research planning, citations, and writing assistance
• 32:25 – Thoughts on using ChatGPT by voice versus typing
• 34:07–35:08 – Will AI stunt thinking? Reflections on critical thought and tool overuse
• 39:22 – Letting ChatGPT “interview” you to build documents
• 43:05–43:24 – Workflow benefits: seamless transition between mobile and desktop

---

Takeaways for Listeners

• Start small, experiment, and don’t fear mistakes; you can’t “break” ChatGPT.
• Use ChatGPT as a “thought partner” for writing, brainstorming, and problem-solving—especially useful for overburdened nonprofit professionals.
• Break complex requests into smaller, focused chunks for better AI output.
• Custom GPTs can be powerful, but troubleshooting often requires stepwise instructions.
• While ChatGPT can be a valuable accelerant, maintaining critical thinking and personal engagement in your work remains crucial—AI is an assistant, not a replacement.
• Verify citations and critical data provided by AI; don’t let it shortcut essential intellectual work.

---

“Stay curious,” says Cary, “and try it out—even if your first result is ‘so-so,’ keep trying. The value grows the more you interact with it.”

---

Transcript

[Hiker 1]

Awkward time to ask this, but. Hey, did you download the trail map?

[Hiker 2]

Yeah, no, I don't need to.

[Hiker 1]

I don't understand. You're trusting your signal out here.

[Hiker 2]

I'm trusting T Mobile. They have the best network. And if we end up in bumtots nowhere, well, we've got T Satellite for backup. Whoa.

[Hiker 1]

I don't trust my carrier that much.

[Hiker 2]

We'll just use your phone as a flashlight.

[T-Mobile Announcer]

With America's best network and T Satellite, we're keeping you connected in places you never thought possible. And if you switch today, you get free phones for zero down and only 25 bucks a month per line for four. Find out more@t mobile.com or visit your local store.

[T-Mobile Legal/Disclaimer Voice]

Best Mobile Network Based on analysis by Ooklev Speed Test Intelligence data to age 2025 with 24 monthly bill credits and 4 eligible port ins on essentials for well qualified customers with autopay plus taxes, fees and $35 connection charge per line credits and imbalance. 2 if you pay off earlier, cancel contact US Finance Agreement example $299.99 Moto Edge 5G required T Satellite available with compatible device in most outdoor areas in the US where you can see the sky included with experience beyond or $10 a month. However, news monthly cancel anytime visit t

[Instacart Announcer]

mobile.com instacart understands that not all bananas are created equal. Some people want them green, some want them ripe, some want them ready. Right now with Instacart's Preference Picker, now available at most retailers, you can choose how you like certain items like banana ripeness, deli thickness, even avocados before your shopper even starts. So instead of crossing your fingers and hoping for the best, you get groceries picked the way you would pick them. It's a small thing, but it makes a big difference. Download the app and get bananas just how you like with Instacart.

[Molly Williams]

And so I started diving into it and one I realized my boss was not exaggerating not just about what ChatGPT can do, but the impact that generative AI is already having and will undoubtedly have on our future. And I was instantly hooked and couldn't stop trying to figure out different ways to interact with it. And now I use it just about every day.

[Kerry Weston]

Danny.

[Podcast Narrator/Host]

Hey there. Welcome to the Chat GPT Experiment, a podcast where we take all that confusing chatgpt stuff and break it down so even Grandma could understand. If you're a curious beginner, well, you're in the right place as we'll help you understand what Chat GPT is. How others are using it and practical takeaways that you can use right away. It's going to be fun, it's going to be informative and maybe surprising when you find out how helpful ChatGPT can be. Alright, let's get this thing started. Here's your host, Kerry Weston.

[Kerry Weston]

Well, hello there and welcome to another show.

Welcome back.

If you've been here before and if you're a first time listener, welcome to the community. I'm glad you're here. The voice you heard open up. Molly Williams is our guest today. Molly is then executive director of of a nonprofit in Boston and some of the things she's going to share today is going to be relatable to many of you regardless of whether you are in a nonprofit or not. Right. I think the concept of trying to get too much done in the day and splitting between deadlines and priorities and trying to balance all of the things that are coming at us is a pretty relatable concept. Profit, nonprofit, personal, professional, doesn't matter. And I'm happy that I got to talk to Molly because there's some things that she brings out that will be very, very relatable to many of you, if not all of you listening. And I'm looking forward to sharing that. It is Wednesday. Let's see, June 12th as I'm recording this, I'm a Celtics fan. The Celtics are in the finals. It's game night tonight. Pretty excited about that and all the things that go with it. School is almost out for us. My goodness, baseball is in full swing. I just want to share with you real quick for any travel parents out there. I had four games last weekend and traveling up and down the highway and I got to do it again this weekend. It's what we sign up for, right? We only have a small window of time to be with the kids, so it's good that we get to participate in things like this. But I'm just about done with traveling. I think I want to see, I want to see the lake. The sun is out and I want to unwrap the boat sometime soon. Off to New Hampshire for Father's Day weekend. Lucky I get to stay a couple nights. I'm going to see Socks Yanks on Friday night. Kind of a cool treat there. Then we're going to head down to Manchester, New Hampshire and spend some time watching the New Hampshire Fisher Cats double a ball club there in New Hampshire of the Toronto Blue Jays and the families coming down for Father's Day thing. And then we're going to spend the rest of the weekend watching Travel Ball before we head back. And the kids, believe it or not, have school one more day on Monday. So we've got a travel back. It's about a four hour trip when we're done. Games won't get over till about 7 or 8 o' clock so you can do the math. And for a guy that likes to be in bed by 9 o', clock, you can see that I am going to be over my curfew a little bit this week. So let's get to Molly Molly's a listener of the show. She's an executive director of an organization called Mobile Health Map and the Family Van at Harvard Medical School. And here's what she shared about how she's using ChatGPT these days. She says, I'm an executive director for nonprofit and like most executive directors, I wear many different hats. Top of my day to day duties of managing people and budgets in meetings. I spend a lot of time fundraising, speaking and writing. I found ChatGPT especially useful for grant writing. It's helped me come up with some interesting project titles and write smart goals and develop scopes of work, revise content to meet character limits and reframe our work to align with the founders or funders actually interests. And one of the best uses that she has found is to ask ChatGPT to review and score grant proposals using the funders rubric and then give specific feedback on how to improve the score. So that's just one of the little things that she's found that I want you to listen for. So I think it's important to say that you've heard me say before. If you're a repeat listener. This isn't about learning software. This isn't about learning all the things that Chat GB can do. It's about trying to figure out what in your world could be replicated, made easier, scaled, made more efficient. What are some of the things that you tackle each and every day that could be put into ChatGPT to just make things a lot faster, a lot easier, Right? And a lot more efficient and perhaps even valuable? Some additional information here about Molly She's a public health professional, has devoted her career to promoting access to care and advancing health equity. She currently serves as the executive director, as I mentioned, of the Family Van and Mobile Health Map, where she draws upon her extensive experience to provide vital health care services to underserved communities. You're going to find that her experience in using ChatGPT started out in a skeptical manner, much like perhaps some of you either have been, or maybe still are. And that was why I chose the opening clip for this particular episode. Because it really does take a couple eye opening examples of how this thing works to really see that, hey, you know, there is something here for me. So if you find yourself in a place where you're still wondering how the heck to use this thing, or if you find any skepticism where you find other people using it, you're just like, I don't know how. I think Molly's got a good message and I hope the conversation give some value to you. Okay, so let's get you to our conversation. And as always, right, look at your own world, like I said, but bring your own curiosity, bring your own perspective and stay curious with this thing. Get in there, play with it, dabble, don't break it. You can't break it. That's what I meant to say, you can't. So dig in, try it out, be curious. You know, start with a limerick if you want. You don't have to go right into advanced anything.

Just have some fun.

Right? So until we talk next time. Go Celtics. Go Socks. Hopefully I make it home in time to get some sleep. And until we talk again, stay curious. Talk soon.

Hey, Molly, thanks for being with us.

[Molly Williams]

Hey, Carrie, thanks for having me.

[Kerry Weston]

So you've been listening to the show

[Molly Williams]

for a bit, maybe a couple of months now.

[Kerry Weston]

Yeah.

And you reached out and appreciate you doing that. And I always want to put the conversation in context, so would you mind sharing with folks a little bit about who you are, what you're doing, and maybe who you're doing it for these days?

[Molly Williams]

Sure. So I work in the field of public health and have done that for the last 25 years. And I'm currently at Harvard Medical School leading two different programs. One is called the Family Van and we operate a mobile clinic here in Boston. And then I also lead a network of mobile clinics called Mobile Health Map. And we're a network of mobile clinics and we help out with evaluation and research. So like an executive director of any nonprofit, I wear a lot of different hats. And that's is. That's one of the reasons I've been so drawn to ChatGPT.

[Kerry Weston]

Just for the sake of assuming that there's a question out there. What is a mobile unit? What is a mobile van? What is mobile outreach?

[Molly Williams]

Good question. So it is literally a clinic on wheels. So take a imagine a Winnebago and outfit the inside with exam table, maybe even separate exam rooms, a waiting space, a lab space you can fit a lot into. A vehicle. And they do come in all shapes and sizes. There are about 3,000 of them that we know of around the country, and they typically travel to underserved neighborhoods to get health care to people who wouldn't otherwise get that care.

[Kerry Weston]

Okay, and you mentioned this is a program of Harvard Medical, and so is this basically a Boston area service.

[Molly Williams]

The family van is mobile Health Map is a national effort, so we work with mobile clinics all across the country.

[Kerry Weston]

Okay. All right. So you're juggling a few different things at any given time. Yeah. And the executive director role that you play, from a nonprofit point of view and just from an organizational point of view, I've known and worked with a number of them, and they always seem to be trying to do more with less, be it time, money, or people. And it doesn't seem to get better every year. It seems to be a new challenge. Right. And just the time that you adjust to the new situation at the time when the new crisis comes. And is this hitting home at all? Does this feel similar to where you're at?

[Molly Williams]

Yes, exactly. I mean, I. Some people call it the nonprofit hamster wheel because you're constantly needing to keep raising money to do new programs and create new programs to raise more money. And it. It is really hard to get off that hamster wheel.

[Kerry Weston]

And what do you like about what you're doing? What drives you?

[Molly Williams]

I got into public health primarily because I grew up during the 80s, and so I was heavily influenced by the HIV epidemic and saw the way that that impacted different groups disproportionately and how much stigma there was and silence there was around the topic. And it motivated me. It lit the spark in me, sort of an activism spark. And so I saw public health as a way to address that and then have gone in lots of different directions, not just working on hiv, but on. You look at any health issue, and there are inequalities, and I believe that everyone should have access to the health care that they need. And that's what's driven my. My career choices.

[Kerry Weston]

I'm a fan of a podcast called the Rewatchables, which is they'll go back and take a look at movies that you could pick up at any point through the years and kind of break it down and have a fun conversation around it. And one of the most recent episodes, I share this on purpose because what you said just triggered it. Philadelphia was one of their movies that had come back, and I known of, you know, obviously Denzel and Tom Hanks to the largest start I've Never seen it. And so I went back and watched it a couple weeks ago from the point of view of, you know, I remember what it was like. I remember that whole episode and that time and the lack of information and the misinformation and all of the things that went with it. And the movie brought you right back to it. And it's amazing. I know things have changed in that regard. I know some things still remain the same in other regards. But it was interesting to see how large of an impact that one particular topic had on public health concerns. So I can see how you'd be motivated by that. Good for you. Can we just start off with a softball question? Then we'll get out of the, we'll get out of the heavy public health side and get to the topic of the show, which I'm sure has many benefits to you if you're juggling and efficiency sake and a problem solving partner. But just what got you started in using ChatGPT as a whole?

[Molly Williams]

You know, I heard about it from both my boss and my sister. So my boss is an associate dean of community education at Harvard Medical School. And she was going on and on about artificial intelligence and its impact on the world and how it was going to change everything. And I was like, she, she, I think she's exaggerating. Like this is probably not as big of a deal as what she says, but I'll have to check it out. And then my sister, she also runs a nonprofit, an animal rescue organization in Mexico and she's a one person shop. And so she had found ChatGPT as a way to help her juggle all these different roles. And so I started diving into it. And one, I realized my boss was not exaggerating not just about what ChatGPT can do, but the impact that generative AI is already having and will undoubtedly have on our future. I was instantly hooked and couldn't stop trying to figure out different ways to interact with it. And now I use it just about every day. And I have an 8 year old son and he is also fascinated with it.

[Kerry Weston]

Okay, now that's, that's interesting because I have, my kids are 20, 15 and 13 and the 20 year old is a junior in college and she's been a guest on the show and the feedback from her professors have been thou shall not talk, use or discuss ChatGPT in the classroom. And 15 and 13 don't know what ChatGPT is yet. So we have conversations about being curious. But it's interesting. You said your son's eight How. How is he using it now?

[Molly Williams]

He mainly likes to play with it because he'll. He sees videos on YouTube of people playing with it and tries it or make it do weird things and then he wants to repeat that.

[Kerry Weston]

Yep, the mimicking. So TikTok and YouTube is. Is constantly. It makes complete sense to me. I don't know how many trends we have in the house that started as a TikTok con conversation.

[Molly Williams]

Yeah, yeah, we don't have TikTok yet in this house. But that day, let's see if you come.

[Kerry Weston]

I am in the. I am in the limiting TikTok time spent on the phone phrase at the moment. So I know you do, Molly. You do a lot of fundraising. You're an avid speaker and I'm assuming you're writing as well in public publishing as well. So are you using it for all three of those ventures at the moment?

[Molly Williams]

All three of those. And then also just for day to day tasks like writing emails or cleaning up data, thinking of creative ways to tackle a problem, creative ways to communicate something or to communicate difficult information. I use it as a thought partner in those instances.

[Kerry Weston]

Yeah, I love that. So my three favorite words in working with ChatGPT is tell me more. And so I'm going to do the same too. So there's some nonprofit folks, there's some executive directors and some managers of organizations listening because that's the role of this particular episode. How do you use it? Tell me more. How are you using it for fundraising? Can you give me a for instance?

[Molly Williams]

Yeah, for fundraising, a lot of speaking specifically about grant writing, it is. It can be pretty formulaic. There are certain pieces that you need in every grant proposal, like background information on who your organization is and who you serve, what the needs are in the community, what your objectives are. So that information is the same in most grant proposals, but every funder wants it presented in a slightly different way. So they may have character limits or they may require most require that the objectives be smart, specific, measurable, and so on. Some of them want detailed work plans quarter by quarter or month by month. And so when it comes to completing that grant application, I can tailor things that I already have, like ideas about what I want our objectives to be. Give that to ChatGPT and say, make these smart. And I can do that throughout the whole proposal. One of my favorite ways to use it for a grant proposal is to help me generate a title for my grant application. You always have to give it a title. And it can't just be mobile Clinic does good things you want it to be something catchy or memorable so that when people are reviewing it, and often after a grant is reviewed, there's a discussion about it. And I've been on that side of things. So when you're discussing it at the table, you want to be able to say, well, let's talk about that mobile clinic grant. Or even if you have a better name for it, say Healing on Wheels grant. You can imagine them using that name. So I'll give ChatGPT the grant and say, come up with five possible titles. I'll give it feedback and say, I like titles one and three, but make them shorter. Give me five more options and I'll just do that until I get an option that I, I like or I'll combine a couple of the options. But typically I would have done that if I had a big team in person. We'd sit around with a whiteboard and do that kind of brainstorming. And I don't have the luxury of that anymore. I work remotely most days. My team is working on lots of other things. So ChatGPT is my brainstorming partner.

[Kerry Weston]

That's fascinating. And do you have, because I'm assuming grant writing is a fairly regular occurrence for the organization. Do you have nuggets of reusable text that you have saved? You bring in and just kind of say, and so are you feeding chatgpt the actual document when you have a grant? Sometimes the requirements of the grant, sometimes

[Molly Williams]

I am, and other times I'll cut and paste just a section of the, say, request for proposals. For example, it might ask me to write an evaluation plan like how am I going to evaluate the impact of this grant? And that the evaluation plan must include X, Y and Z. So I'll cut and paste that into ChatGPT and I've already given it the proposal and say, give me ideas for this, this evaluation plan, or write me an evaluation plan. And I find it works better when I do it in chunks. Take, say a section at a time, then give it, you know, everything all at once and give it all of its instructions all at the same time. I like to break it up.

[Kerry Weston]

Gets tired, I call it. It just seems to just water down. Yeah, I just did this today, actually going through work from a client to review and instead of doing the whole document, you know, it's probably a 3000 word document. Give it sections and sections that it can hyperfocus and really kind of hone in on that. Yeah, I agree with you there. And are you using any custom GPTs to do this now. Have you gotten to a repetitive task where it made sense to kind of fire up a custom GPT?

[Molly Williams]

I've experimented with the custom GPTs, but I haven't really figured them out yet. There's one in particular I've been working on for writing smart objectives, because I thought that might be helpful for other people as well. And I don't know what I'm doing wrong, but it can't seem to figure out that the whole, that it should be one or two sentences and that whole, those whole sentences, that entire sort of short paragraph should be smart. Instead, it's giving me a specific part, a measurable part, and I haven't been able to get it to do what I want it to, but I haven't given up. I'll go back at some point and try another approach, but it's something I want to get good at.

[Kerry Weston]

So just before we jump on the call here was working with a client on this exact same issue. So if permission to ask you a few questions, just kind of go down this trail because I want to see if this resonates or has value.

[T-Mobile Legal/Disclaimer Voice]

Yeah.

[Kerry Weston]

So do you have, outside of custom GPT, when you're working with ChatGPT, do you have that same issue where you are sharing with it paragraph or section by section and you want to be smart within that paragraph? Have you gone down that journey with ChatGPT to say, make this paragraph smart in and of itself?

[Molly Williams]

I don't think so.

[Kerry Weston]

Okay. So I just did this exercise on the fly with them and I'm just going to share it with you because it's top of mind and see if it has any value. Later on we were having an issue where something wasn't happening and so we just kept revising and revising and giving it different prompts. And we got to the point where it actually did, you know, exactly as we wanted, finally. And then the phrase, typically the phrase would be, I like where we're at. I like the output, whatever it might be. I like the style, the context, the language, whatever it might be. Can you please give me the prompt that I could get here again? Right. But we decided to take it one level deeper and we said, okay, I like the style, like the format, I like the way in which we ended here. I would like, and this is going to be a repeatable thing, I would like to create a custom GPT in chatgpt. Can you give me detailed step by step directions that I could feed to the custom GPT that would allow me to get to the Same place while you are interviewing me to get the information needed to do it right. And it said, yeah, I could do that. And then it spit out a bunch of code and we paste, just on the fly, we pasted that into a custom GPT to see what would happen. And it, it did a little tiredness at the beginning, but the, it was pretty good. And we went back into the custom GPT configurations and I noticed that it was a paragraph. You know, you will do this and then you will do this and then you will do that, that kind of thing. So what really made it great is it gave us, let's call that 90%. We got 90% of the way there just by doing the exercise in ChatGPT and then saying, I really like where we're at. Give me the configuration language for custom GPT. And then we took that paragraph and we each things with step one, do this, step two. When I did the step one, step two, step three, when I broke it down to that same paragraph, same language, actually, same words, I just prefaced each sentence with step number. Bam. It, it did it. And there's something about telling it to pause, do this. And it knows that Step 1 is not going to be done until Step 2 is not going to be done until step 1 is over. That language helps it. So the combination that I would give you to try, just for kicks and let me know how it goes, is do some sort of exercise in regular ChatGPT and then ask it for the configuration output. You know, what instructions would I give a custom GPT if I were to do this? And then see if you can configure the steps and see if you get closer to where you want to be.

[Molly Williams]

I love that suggestion. I will try it. I will try it today. I can't wait to get that try.

[Kerry Weston]

Yeah. And, and it's, it's so simple that sometimes we forget. I just did a presentation to college faculty earlier this week and I can't imagine being a college faculty member at the moment on the beginning of this artificial intelligence. I had English professors in the room too. I'm like, ah, so talk about a challenge, right? You've got to figure out how to incorporate it, how to judge it, how to grade it, how to have people understand that this is not a do it for me but a do it with me tool. And we were talking about if you don't know, just ask it, right? That's one of the most powerful things you can do with ChatGPT is just simply, hey, this is the situation. How can you help? You know, how can you help me? Or this is my role. How can you help me? And so we just applied that to the same thing. I can't quite create the custom GPT I want. How do I do it? Can you help? And it turned around and said, sure,

[Molly Williams]

I'll definitely give that a try. That's fascinating.

[Kerry Weston]

Yeah. And it was really fast. And it's got that 4.0 that we talked about earlier, the updated 4.0 that just came out a couple days ago, in fact. In fact, as I was making that presentation to them, I said, this thing is changing so fast that something I say now won't be true tomorrow. And that was literally true because they were doing that presentation as I was doing my presentation. And so some of the stuff I was sharing was no longer relevant. It had only been an hour and 15 minutes.

[Podcast Narrator/Host]

Mm.

[Kerry Weston]

What about writing, Molly? How are you using it for writing? I know you do a lot of publication work.

[Molly Williams]

Yeah, I use it in lots of different ways. So lately I've been thinking about a research project I'd like to start, but I'm not really sure the best methods to use in its. I want to. I want to approach something. I want to try something that's called content analysis. So. And you see papers like this published from time to time, where they look at a whole bunch of documents and they essentially analyze what's in those documents. And I may ultimately use ChatGPT for. For some of this analysis, but I'm not there yet. I'm just trying to figure out which documents to analyze and how to get the data out of those documents. And I've been going back and forth this morning, actually with ChatGPT about the best way to do that. So I told it what my goal was, what my research question was, and why this question was important to me, and then asked for ideas about ways to do content analysis. To answer this question, they gave me like five or six different options. And again, I said I like one in three. And one of those was about web scraping, and I've never scraped webs. And I played around a little bit with it before, and so it gave me specific, try this software package, take these steps. And so I did that, and I realized that method was not going to work because of the nature of the documents I was scraping. So I went back to ChatGPT and said, that's not going to work. What else could I try? And it gave me other options. Now, sometimes the options it gives me are just not realistic, and I'll Go back and tell it that I would love it if it could give me the right option the first time. But it doesn't always have that information. Like for example, one of the software packages it recommended might, could have done what I wanted, but it required a $250 subscription and there was no free option or like week long trial. So it was like, nope, that's not going to work. So anyway, one of the ways I'm using it again is as a thought partner in planning this research project. Another way that I've used it for writing is with citations. So I will sometimes have a whole bunch of citations that I've used and they're all in different formats because I cut and paste them from different documents and I'll just say update this list in AMA or APA format for whatever journal it is, whatever they require. And it does it really quickly. I don't trust it, I should say, too many times it's given me fake citations. So I don't ask it to give me citations. I find them myself. But it can then help me to format them.

[Kerry Weston]

Do how do you know they're fake? Initially you verify each one?

[Molly Williams]

I, I start at the top verifying them. And in almost every instance the first one was fake.

[Podcast Narrator/Host]

Yeah.

[Molly Williams]

And then I check the second one. Yeah, they're all fake.

[Kerry Weston]

It's been programmed not to tell you. It doesn't know it's good. It's been, I believe it's been programmed never to give you. I don't know answer or give you no answer. And so I think because of that, that's my reading. I have never seen this. It's just my assumption. It tries to fill it in to give you something.

[Molly Williams]

I've asked it before, are these real? And it said yes. And then I argued with it. I said no. I checked this link you gave me and it showed me this and it said, oh, you're right, I'm sorry it's not a real link. But yeah, it doesn't want to displease you.

[Kerry Weston]

Right, right. And it's great at apologies, by the way. It's awesome. Yeah, I'm sorry about that. You are? I think it was. I think you shared with me as you are interacting with it, you are starting to use voice and talking to it. That been. Has that been helpful? I see you nodding, so has that been helpful?

[Molly Williams]

Yeah, I, I've tried it out and I'm, I'm on the fence about voice and I think that's, it's probably a holdover from my experience Writing in that I'm so used to typing. I equate typing with thinking. And when I turn on voice, I get a little bit of stage fright. Like, what am I supposed to say now? Normally I think typing slows me down and so I have more time to come up with my thoughts. And then I get nervous when I feel like I need to actually speak them at some, you know, normal pace. It's kind of silly. I think if I practiced it more and played around with it more, I might like it. But for now, I think I. I'm still mostly typing.

[Kerry Weston]

Yeah. And I think that's a personality trait. So for you, what you just shared is the exact opposite of me. Just for relatable, like, I think faster than I can type and I make too many mistakes when I try to keep up. And I also find my fingers getting in the way of me being thorough. So for me, talking is how I get everything out of my head and I don't order anything when I just get it out, get it out, get it out. Because I can just. There's always circus music going on up here. So I just get it out. But I can completely understand that, how the process of thinking and composing and the pace of typing go hand in hand for you. It's just something I can't relate to because I'm exactly the opposite way. Like, I've got to get my fingers out of the way because I'm already onto the next stop before I finish the first one.

[Molly Williams]

One thing I've been wondering, thinking about that typing and thinking is I heard someone talking about generative AI and they were worried about students about young people not learning to think in the way that we did because writing is thinking or you think something about the writing process stimulates thinking. That wouldn't happen if you just let Chat GPT even write the first draft for you. You didn't have to do that thinking work to put it on the page. So I'm. I'm curious to know what you think of that especially. We both have kids, and once yours start using Chat GPT, will it stunt their ability to think?

[Kerry Weston]

So let's just, let's not even filter this for kids. I have seen. I'm a big advocate of telling our team here to experiment. If we're going to be conversant, if we're going to be value added, if we're going to guide and have any strategy and be a valuable partner, we've got to make the mistakes ourselves. We've got to figure things out ourselves. And ChatGPT is no different. However, I have witnessed an overdependence from adults on using the tool. So I don't even. I think your concern is absolutely valid but I don't think you need to be a kid in order for this to be a concern. I see that if you don't harness it a little bit and interact while you see it happening or at least manage it, there is a tendency to just have it be efficient and nothing else. Have it be a fast way to an end and then move on. And if there's one thing that I try to share over and over again is there are, I have an agency, I have a, I have a marketing PR agency and the value of our firm is not the deliverable, it's not the stuff. The production shop is not the valuable side. That's the outcome, like that's the end product. But the real value is the ability to listen, to, relate to problem solve and to be valuable on a, on a meaningful level, whatever that means in the situation, the client in the business to be a thought provoking problem solving partner that can help navigate and solve problems and get to an end result. Yes, we might produce something, but that's not the value. We can, there's producers, there's people that can do things all over the place. So I share that if, if we are just using chat or somewhere, doesn't matter, it just happens to be chat GPT for this instance. But if we're using any tool just to produce in, in volume and get it done, then we're missing the whole point of why we're here. And so yeah, I, I can see there being a concern. It's the same conversation we had with the calculator, it's the same conversation we had with a computer.

[Hiker 1]

Right.

[Kerry Weston]

It's just another generation of a tool and showing your work, right. And thinking about your work is going to be more important moving forward. So I think it's a valid concern, an absolute valid concern. And look at where you are, you're, you're in public health. The person that is taking care of a human being better know what they're doing. If the tool doesn't work, like if the software doesn't work, if the instrument doesn't work, there should be some sort of base knowledge and understanding as to what we're doing and why. Right? If we're just relying on the computer to do it, then we've got an issue. Right? So I'm sure you face that too.

[Molly Williams]

Yeah, exactly. It reminds me, I have a friend who was on the other side of Boston, and she was trying to get home recently, and her GPS stopped working all of a sudden. And she called me and she said, what do I do? How do I get home? Do I go to a gas station and get a map or ask someone? It had been so long since she had had to navigate without a gps, even just within the city of Boston, that she didn't know what to do. I was like, yeah, we are so reliant on these machines and you don't realize it until they're gone.

[Kerry Weston]

How long have you been in Boston?

[Molly Williams]

6 and a half years.

[Kerry Weston]

Okay, so this may be. This may be dating that. But I remember when you could drive through Boston two times in a month and it'd be a different route because there's so much construction, you didn't know what the roads were anyway. Right. So that's funny to hear. I don't know how to get through Boston. But no, we are so dependent on things now, and my kids will. They don't understand what a paper map is. Right. I remember taking trips with a paper map and all. And so, yeah, there's absolute concerns as to, are we using this tool to fast track the busy work so we can pay attention to the valuable, or are we just using this tool to get through it all together and not think about it at all? And that's a. Yeah, that's a real concern. Moving forward, which is the last topic I wanted to share or ask you to share is you're also using chat. Speaking of figuring out if you know what you're talking about, you're using chat to interview you and ask you questions too. Yeah. And the work that you're doing, can you share a little of what you've been doing there?

[Molly Williams]

Sure. And it's something that I learned in your podcast I'd never heard about, that I could ask ChatGPT to interview me. And so a few, maybe about a month ago, I had, in the span of a week, five different proposals do where we were not the prime proposer. We were going to be subcontractors. And it was a totally different kind of proposal than we'd normally be a part of. It's. There's this new big government agency called ARPA H which is funding rapid development of health technology solutions. So people had asked us to work with them on their proposals, and they were asking us for things we had never produced before. I've written a scope of work, but not for a government grant. So I told ChatGPT I have this Opportunity. I need to submit a scope of work. This is the audience. Ask me questions one at a time until you have enough information to put together a detailed and thorough scope of work. And it did it again and again, and almost every time it was 90% there

[Kerry Weston]

just by gathering answers from you. And then how does it know it has enough? I'm interested in that comment. You said until you have enough information, how does it know that it's qualified enough?

[Molly Williams]

That's a good question. I think in my mind when I looked at it, I could say that's enough because each section, there were sections of it that I didn't give it the sections, but there was like a background deliverables, a. I can't remember the other sections, but they were all filled out with a paragraph or a number of bullet points. So the basics were there.

[Kerry Weston]

And this was for a grant or is this for another document?

[Molly Williams]

This was for. This was going to be an attachment for someone else's grant. We were going to be a subcontractor.

[Kerry Weston]

Yep. And so the sections were predefined by that document.

[Molly Williams]

Well, not really. The subcontractors were pretty open. I mean the. Sorry, the proposers simply emailed us and said, send us a scope of work. And. And they also asked us, some of them asked us for. I think it's called a competency statement. A competency document which I look read about is. Is kind of like a CV or a resume for your organization, which as a nonprofit's not something we've ever been asked for.

[Kerry Weston]

Yeah, I just learned something right now.

[Molly Williams]

Yeah, I used the same technique and I double checked it because I asked it to do it and it did kind of look like a resume. So then I googled what's a competency statement? And it was that. So I double checked ChatGPT's work.

[Kerry Weston]

Yeah, trust but verify. That's pretty cool. And so the, the asking allows you. Now you said earlier that typing allows you to think. It allows you to slow down and pace yourself and think. So when it's asking you questions, are you typing it back or are you talking about

[Molly Williams]

in this instance? This was interesting because the week that all these were due, I was also really busy. And so I was on the bus a lot. So I was using my phone and typing in my answers. I wasn't in a place where I

[Kerry Weston]

could talk out loud. Yeah, yeah.

[Molly Williams]

But the fact that I could do it on my phone and I could see basically on the small screen what was coming out and think, okay, that's close enough. When I got home, that chat was on my home computer and I could finish it up, cut and paste it, and email it to the. To the group that had requested it.

[Kerry Weston]

That is a great point that I've never said before, but I appreciate you bringing that up. Is the cloud based system that is ChatGPT. You could start a conversation on your phone knowing that when you pull it up in your desktop later, the conversation is there because it's the same account. Yeah, that's very interesting. I'm going to bring our conversation to a close, but I've had a lot of. This has been interesting. I learned a lot and I appreciate you sharing because I know there's a lot of folks that are busy and juggling and doing important that were they are time strapped as well. And you've given them some ideas as to how they could approach theirs. You have any advice for someone that's just getting started or wondering how the heck I'm supposed to use this thing at all?

[Molly Williams]

I think I heard this on your show, but I would repeat it, which is you can't break chatgpt, so play around with it as much as you can and don't be afraid to experiment with it. The more time you spend with it, the more comfortable it will get and you'll move very quickly into really relying on it. So if it doesn't do what you want the first time or you're disappointed with it, which is what I hear from a lot of people, the first time they try it, they're like, so. So keep trying because it, it does. Does get better.

[Kerry Weston]

Yeah.

Yeah. That's awesome.

Well, good. It's been a pleasure having you and thank you for reaching out. I've enjoyed the conversation.

[Molly Williams]

Likewise. I did too. Thank you for having me.

[Kerry Weston]

And if you would, when you try out your custom GPT, just get back to me, let me know how it went. I want to share.

[Molly Williams]

I will. I'll let you know. I'm going to try it this afternoon.

[Kerry Weston]

Awesome. Very good. Well, Molly, thank you. We'll talk again soon. Okay.

[Molly Williams]

All right, take care.

[Podcast Narrator/Host]

Bye.

[Hiker 1]

Bye.

[Molly Williams]

Bye. Bye.

[Podcast Narrator/Host]

All right, folks, that's a wrap for today's episode of the ChatGPT experiment. Thanks for hanging with us and diving into the world of ChatGPT. We hope you found it helpful and fun and informative. And speaking of helpful, we'd sure appreciate you taking a moment to rate and review us on your favorite podcast platform. By the way, five star ratings are our favorite. Until next time, keep experimenting, keep learning. And remember, Chat GPT may be Smart, but it can't make a decent cup of coffee yet. See you soon.

[Hiker 1]

Awkward time to ask this, but. Hey, did you download the trail map?

[Hiker 2]

Yeah, no, I don't need to.

[Hiker 1]

I. I don't understand. You're trusting your signal out here?

[Hiker 2]

I'm trusting T Mobile. They have the best network. And if we end up in bumtots nowhere, well, we've got T Satellite for backup.

[Instacart Announcer]

Whoa.

[Hiker 1]

I don't trust my carrier that much.

[Hiker 2]

We'll just use your phone as a flashlight.

[T-Mobile Announcer]

With America's best network and T Satellite, we're keeping you connected in places you never thought possible. And if you switch today, you get free phones for zero down and only 25 bucks a month per line for four lines. Find out more@t mobile.com or visit your local store.

[T-Mobile Legal/Disclaimer Voice]

Best Mobile Network Based on analysis by Ooklab speed test intelligence data 2H 2025 with 24 monthly bill credits and 4 eligible port ins on essentials for well qualified customers with autopay plus taxes, fees and $35 connection charge per line credits and imbalance due if you pay off earlier. Cancel Contact Us Finance Agreement example $299.99 Moto Edge 5G required T Satellite available with compatible device in most outdoor areas in the US where you can see the sky. Included with experience beyond your $10 a month auto renews monthly. Cancel anytime.

[Ryan Reynolds]

Visit t mobile.com hi, Ryan Reynolds here for Mint Mobile. Are you looking for a beach read this summer? May I suggest your big wireless bill? It's got suspense, mystery, a slightly flat emotional arc, and a shocking twist where you realize you've been overpaying the entire time. Fortunately, though, Mint's story is better. Every plan $15 a month, even unlimited. That's it. Happy ending. Zero tears. Give it a [e-mail] Switch upfront payment

[Molly Williams]

of $45 for three months, $90 for six months, or $180 for a 12 month plan. Required $15 per month equivalent taxes and fees. Extra initial plan term only greater than 50gb. Me slow when network is busy.

[Instacart Announcer]

See terms.