In The Arena by TechArena - Wanjiku Kamau on Her Book: Out of the Loop, Into the Algorithm

Episode Date: January 22, 2026

Allyson Klein talks with author and Google/Intel alum Wanjiku Kamau on moving past AI skepticism, learning fast, and using new tools with intention—so readers start where they are and explore AI with hope.

Welcome to Tech Arena, featuring authentic discussions between tech's leading innovators and our host, Alison Klein. Now, let's step into the arena. Welcome in the arena. My name is Allison Klein, and today I am here with Wenghika Kamal, the author of Out of the Loop and Into the Algorithm. Wengh, you have experience working across the tech sector at Google Cloud and Intel. You're now a CEO of an independent civic initiative. And you have got a way to put a journey with AI, which is when we're here with this episode today. First of all, why don't you just go ahead and introduce yourself and then talk a little bit about why you decided to make this book. Thank you, Allison.

I'm so glad to be here and participating here in the arena. And thank you for the introduction. You couldn't have said it better, but I've got some years in tech. And most recently had a layoff from the company that I was working. at and just participated in this journey of really wrestling with my bias around AI and starting to lean into it and thought, you know, I'm writing this book and whom I really writing for and what is it about. And originally I was writing it for my parents and my parents' friends and people who are just feeling out of the loop when it comes to tech and more specifically LLMs and started out as

more of a PDF and it has evolved into a full-blown book. And the thinking there was, I think there's other people just like me who are in their careers, who are successful, who have some hesitancy in embracing AI. And I wanted to just share my own personal journey and the story that I took and wanted to be able to show that with other people. So that was the thinking around writing the book. Now this story opened in your airport with a layoff and a dog named Shant, which is an incredible narrative. She's proud with. And I think that what's interesting is just the same that you paint and what it brings up in terms for questions about what comes next. I think that's something that surprised me also is that you are deeply rooted in tech, yet you, just like many other people, found,

yourself, couple of prostrates with AI, work a little bit about them. Yeah, absolutely. So maybe give a little context as to like, why an airport scene, I lived in Sacramento for a very long time. I have my family that is still there. My brother and his wife had just had a baby. And so I'd taken some time to go spend it with them. And I try my best when I travel just domestically to always bring my dog champ. He's a nine-year-old Chihuahua mix and he's my best pal. So with all of that said, I had originally gotten the notification of my layoff some months prior. And with the layoff, you tend to get a certain amount of days to find a job internally or take a severance package. And so that airport scene was the, I shouldn't say the ending, but I had about two months into

this process of spending time with family, my newborn niece L, who's a constant theme in the book, by the way, and really just wrestling with, I had really been avoiding embracing AI. And you start looking for jobs internally and externally. And AI are in job descriptions. Prompt engineering is described. And so you start thinking, do I start my next chapter in my career by ignoring this enormous wave that everyone seems to be embracing? Or do I take a moment and go, let me figure this out for myself? And so when I thought about the book and where to start, there really was that airport scene. My dog is resting on my thigh with his head on my lap. And I'm sitting here with my phone and rather than looking at just dog memes and videos, I thought it best to go, what is AI? What are these

tools? What is this mean? And I started that journey and I opened YouTube and typed literally, what is AI? And I even added a little context of like not for work, but for me personally and how can I learn it and what does this mean? And that's really where the whole journey started, that very late flight leaving Sacramento back to New York. Now, you know, I think that one thing, that a lot of people think about is that AI is going to be dressed and a thing is the memorable. When did you first feel like you were behind? And what did that first 30 days are leaning in look like? Small steps, goals you tried, or how you kept yourself from freezing.

I'll be honest with you, Allison, I still feel behind. And I'm a user of it three, four, five days a week. But more importantly, that first 30 days, I really challenge myself, to be honest with you and said, all right, I'm going to use this almost like a daily warmup. I grew up playing soccer and you had warmups to get ready. And I knew this wasn't going to be something I was going to master. And the goal wasn't even to do that. The goal was to just understand it so I can speak about it and at least be informed of this is a tool that I want to continue to use or this is a tool that I tried and I didn't like it for the following reasons. And so my thinking is always,

I'm going to start with YouTube videos. I'm a visual learner. So when I can see other people talking about something, it helps me understand it and cement it in my mind. I started bookmarking articles, and you wouldn't believe people are, you know, your group chats with friends and former colleagues are sending you articles. So I really realized how much in the publication space there's a lot of conversations about LLMs more specifically and how they are driving reinvention or you can use it to grow your business. And so my thinking was, I'll start small with some experiments. I'll anchor on just one tool. So I chose chat GPT to just start. And I think originally, I was just using it to summarize articles and going, okay, I just tried to read through this 18-page

document plugging into chat TPT and really asking it to help me break it down and understand the human side of what I'm reading. And so what turned into that 10 minutes a day, it felt very doable. I wasn't trying to obtain mastery. I was just trying to understand my curiosity and grow it to a little bit of competence. So something interesting in the book is that you said that AI can work like a mirror for how we think.

Your first chat with AI, and it wasn't smooth, and I think that's the point. Can you share an early exchange that went sideways and what it taught you about asking three of smaller questions? Absolutely. I remember typing

very long, I can get long-winded sometimes, but I remember typing a very long, vague question, and it was a reminder of like, you ask something vague, you're going to get a vague response back. And after doing that a few times, I started to realize AI is mirroring my thinking. It's like, you know, if you put in good insights and information, chances are you going to get a good response back.

And so rather than writing these really long prompts, I was like, I'm going to just do these in really small pieces, right? I'm going to break it down and start the way I would communicate with a person. You know, you wouldn't sit here and ask a question that's three pages long. You'd ask a sentence and then a sentence and then a sentence. And so that was my evolution pretty quickly, I think, in the first month or so. And I realized there was a shift that started to happen. I mean, I talk about this a little bit in the book, but I think this goes back a little bit to the pandemic days and working remotely, not having people next to you.

And chat was our communication with colleagues. And so it was very natural for me to open chat GPT starting my day and throw in a prompt. And I realized very quickly, I started saying good morning to it. And I, you know, I almost stopped myself going, why am I saying good morning to software? But it just felt natural at some point. And so I eventually nicknamed my chat GPT Reese. And the thinking was there was I wanted just a one syllable name, something gender neutral, a name that wasn't in my phone.

I didn't know anyone personally named Reese. And it changed my thinking of how am I going to start. treating it like a colleague. And if I'm going to do that, I'm going to ask questions in a way that, you know, I'm hopefully getting much better responses back and not being vague, giving context, giving the why, right? What am I expected outcome should be? What are my goals? And that evolution really just changed my results and really kickstarted a lot of the work that I started doing with the political action committee I've been working for. Now, something that really resonating with me is that you reframe the state's experiments. Can you give us an example of an AI

misplighter and walk through how you diagnosed it, how you change your inputs, examples, or constraints, and the improved result? Absolutely. So early on, I think maybe six weeks into my AI journey, I built a QR code contact system with AI. And the thinking was, I had this idea about launching a podcast. And I was meeting people who wanted to be guests on the show. And you entered into your phone, their contact info, or they give you a business card. And I'd like to consider myself an organized person, but I was losing contacts. And so I finally approached Chat GPT and was like, I need to build a system to get organized. And I didn't know what I was needing. And one of those results was a QR code contact system. And so I was like, okay, great, let's get it going. And, man,

I failed. I don't even call it a failure. It's like experiments. If anything, my desk here almost feels like a science room where I'm with a beaker and things. And I really started just going, okay, here is my expected outcome. I want it to be smooth. I want the system to be sophisticated. And we kept trying. And, you know, I tried with a QR code. Actually, we tried a different system first. Then a QR code was the one we landed on. And then how do I make it simple? How do I make sure the inputs are working? How do I make sure it's landing in the spreadsheet? How do I make sure the person is getting an email back? How do I make this less mechanical and more human? And I continued to just adjust the prompts. I had to keep testing for tone. I continued to just iterate. And when it finally worked, I realized the win wasn't the automation itself, truly my confidence, that I could diagnose something, I could tweak it. I improved something entirely on my own. And that for me was a huge win because it wasn't doing it for anyone else. It was just doing it for myself, but it was a beautiful result. Now, I think one of the things that you'd partying upon is, like, rest in the data that you're getting.

And you kind of telling us how you iterated through practice. But why did you find out that some data in the eye is not trustworthy? And how did we get a sense for what you could trust and what you could ignore? Can you give an example of that? Yeah. So I think the official term is hallucinations. So I had a potential client that I was building out a proposal for and threw a bunch of items into, I'll just say one of the LLMs. I don't want to call one out because it wasn't the tool, you know, but I was trying to put together prospectus.

And what it gave me back was beautiful, a great title, subsections, you name it, until I got to the market research data. And I had given it some metrics on a proposed five-year forecast. And the outcome was, I can't lie to you, Allison, it was gorgeous. I was like, this is a really good document. I love the way it shape. And then I started to read the numbers and then test the numbers. And I'm like, math is not mathing here. It doesn't work.

And so part of it was the market research it did. It went out and researched competitors for this particular client. And they weren't even in the same industry, the same country. It sounded really good, but it was like comparing erasers to soft drinks. It just was not doing what it should have been doing. And then when I went through the numbers as well, I gave just really simple numbers to see what the growth could look like. It was negative. It wasn't margin neutral.

It wasn't good. And so I think that was one of the first times where I felt a little bit under pressure. I had a bit of a time constraint. And I'm so glad I spent the time to go through the data and realize that it can give you something very beautiful, but very wrong at the same time. And so my outcome from that was like, I need to proof test this. I need to read it through. I need to make sure it's accurate before I share or send it to anyone. Now, you think about your AI tools as collaborators,

and, you know, I think that there's a way that you talk to your collaborators to drive to the right and output. And you give an example of a prompt that you would set the right tone, the context, defines the job to be done, that you get optimal results. Yeah. I'm always telling whichever tools. I'll pick on chat GPT or maybe I'll pick, you know, and I will tell it what I know.

need it to do. And so with the book, I thought it would be good to start a newsletter and really just for my friends and family to be part of the journey. And so I send out a newsletter every week. And I remember typing into chat GPT. You're my creative co-pilot. And I want to draft a newsletter. I wanted to sound warm, curious, written in a conversational tone for readers that are new to AI because a majority of the people on that newsletter are not active practicers in that. And I wanted it to be my creative co-pilot. And so I was asking it to, you know, suggest headline options. And let's make one a bit more relational and a little and more technical. And so those are, to me, was like a really good example of I wasn't thinking about it as just software anymore. I was thinking about it of,

if I had the means to hire someone to sit next to me and work on this newsletter together, what would I hope they're bringing to the table? And then what am I, giving to it back and how do we do this together, iteratively and collaboratively to get a result that really is me, sounds like me, but has a little bit of shape and structure that, you know, AI is able to provide. I love that. Now, in the book, we talk about how you're big first one with launching the website in 48 hours, which was incredible. What that experience teaching about the controls between human creativity and AI, support. Yeah. So let me start off by saying I had every intention of hiring a person to build

a website. And I bring that up, Alison, there are concerns about AI replicating roles that people are in them today. So for further context, I had just joined the Political Action Committee. We were in a position where funds were very low, if not, almost non-existent. And we needed a website to demonstrate that we were here and to get donations from donors. And so if people can't search you, they can't give money to your organization. And so time was ticking. And I had sent out for quotes. I was getting, you know, nine days, 10 days, 12 days, 14 days. And I'm like, I'm only in this job less than two months. Like, we don't have two more weeks to go and wait for someone to develop a website. And so this is one of those high pressure cooker moments where I truly were typing in a chat GPT. I'm C.O.O. of an

independent expenditure in New York City. I need to develop a website. I know a little bit about web design, but not a lot. Help me understand the whole ecosystem of step one to an actual website with the, you know, you can push a button that says donate now and walk me through each step, right? And this is why I always try to say, give it to me bite size, tell me how many steps there are and only give me the first two. If you give me 24 steps at once, I get overwhelmed. But if you tell me there's 24 steps and you give me the first two, amazing. And so Reese, aka, chat GPT guided me through those steps. And I got to make every decision, wrote every single word, I tested the flow myself, and really tried to figure out that balance of how do I use AI speed,

but my intuition. And the magic happened. And, you know, this was a Wednesday afternoon. And Friday morning was our board meeting. And I came in to the board and said, I know we were trying to hire someone for the website. It was going to take too long. I did it myself. Take a look. And they were very impressed. said when can we go live with this? And on the call, I released it and made it public. And so to me, that was a huge win, especially coming off of being laid off and just trying to find your footing again. And then I'm using this tool that I'm still getting comfortable with. And I'm like, we did it. Reese and I did it. And so, yeah, that was definitely your first big win.

That's amazing. As somebody who's company builds websites. I can tell you that is something that is amazing that you look at least for 48 hours, especially with the the limited experiences that you've want in web design and puts like just such a testament to the power of the tools too. I also think that there's a fitball. So one of them that you hit on is one of my topics, which is over-automating things like writing.

I feel like every day I'm reading LinkedIn posts. I'm thinking that people have gone wild and writing chat TV posts that sound very robotic and maybe not authentic and getting away from their own human voice. How do you put guardrails around that? And can you think about examples about what you have seen or experienced that are the right ways to approach this topic?

Yeah. I learned very quickly, especially actually really writing the book of, I always reread anything I write or AI produces out loud, not just reading it on the screen, but generally saying it out loud. My poor dog probably cannot stand my voice at this point. I would never write a prompt into an LLM, get an output, copy, paste, hit send. I just wouldn't do that because it's not me. You don't need me if that's the case.

But I always feel like authenticity, it should be a filter, right? We shouldn't think about it after the fact. And one trick that I do, not even a trick, just something I do, if you can't tell, I like to talk. And I've always found it a little bit easier to communicate verbally than to write things down. And so I use voice notes constantly. I've got to walk my dog. It's New York City. You're always outside. And so I can talk into my phone and say, I'm thinking about writing a book or I'm thinking about, you know, starting a podcast or whatever it is. And I voice note. And sometimes the voice notes are 90 seconds. Sometimes they're 20, 25 minutes. It just depends. And then I take that. I can put that into chat, GPT, and give it context. I was thinking about this on my walk today and here's an idea. And so the starting point always starts with me. It's not just the prompt question. It really. really starts with me. And so I feel that's for me has been a good way to extend my voice into the LLMs in a way that they're not trying to imitate me and I'm not trying to imitate it.

Does that make sense? It makes perfect sense. And I think it's lovely that you've been up with that. I think if people want to lean into one thing, this would be one thing that really set them free. Something else that you talk about is sustainable AI habits. Can you tell me what you mean by that and can we give some recommendations of habits? Absolutely. So when I first started off, I think I got excited. Well, actually, at first I was very nervous. I'm still nervous.

Let me be clear. But I started small and I was like, I'll use chat GPT. And then I had used Gemini a little bit when I was at Google and then tested co-pilot, then perplexity and clawed. And you can get overwhelmed very quickly with the amount of tools there are. I'm a strong proponent of use free until you get to a point that you're coming up against a wall repeatedly that you feel like you want to pay a little bit more for a tool. But I started to figure out there are certain tools that work for certain reasons.

If I'm just needing to brainstorm or sparring, I use certain tools. If I'm using it to help me write something, I'll voice note first and then use different tools. And so it took me a moment to find my footing. And then I'm a lover of a spreadsheet. So I started to document them in a spreadsheet and kind of did a quick mental audit once a week of like, wait, I've gone in two weeks now and I didn't use this one tool. And so cross it off my list and more importantly, deleted the app for my phone so I didn't go to try use it again. But my thinking was like, I'm going to just use AI very simply, you know, coming back to what we're saying about our voice,

I didn't want it to be an outsource for my thinking. That is never the goal. But I wanted to be able to use it to think about things faster, get information. There's so much to be learned in industries, and I wanted to use it as a way to close some gaps when I'm trying to engage with a new client or in the community activism space of, okay, I know a little bit about what's going on in this organization. Let me do some research. And so this was a chance for me to try different tools in different mediums, figure out which ones works, go back to my spreadsheet, cross off the ones that weren't really working and keeping that circle fairly small, because you can get really overwhelmed very quickly. And there's always a new update. There's

There's always a new tool. There's a few teachers that I follow on YouTube that I like because they go through all of them. And I like their teaching style. So they may talk about the new latest GPT or they'll talk about a completely new tool coming out of somewhere else. And so I try to spend my 10 to 12 minutes in the morning just looking at what's new and seeing could it be a replacement for something or is it just more hype and I just ignore it. But that's kind of how I've chosen to manage that. You know, I had a lot of feelings reading your books from tip to tail, and I'm sure other readers will as well. What do you want readers after reading your book? Hmm. Hopeful and that we are capable. I would hope a reader would come away saying, I have a little bit more hope,

and I am capable because whatever you're needing is already inside of you, the whole point of the journey of out of the loop into the algorithm, I wanted to create a subtitle that is how I finally made friends with AI, right? We don't need to be besties, but I really came to terms with how to make it work for me. And so I would want readers to see themselves the way I now do, not behind, but just they're at the beginning. And more specifically, you know, AI isn't about replacing who we are. I fundamentally don't believe that.

I know I'm not using it for that and I don't want it to replace me. But I will say on this journey for myself of rediscovery, it has shown me I am limitless in ways I never thought possible. And for that, I am super grateful. And so I hope people get that they can do whatever they want to. And for me, AI was that tool. It may not be for them, but they get a little bit of hope and capability coming out of reading the book. I love what you've done here. And I want to wait to see what we're first.

You're a published shopper. Congratulations. Where can people find your book and where can they connect with you if they want to talk to you? please connect with me on LinkedIn, drop me in notes. I try to stay on top of my LinkedIn every couple of days at most. And then I do have a little copy of the book, so please check out the book. It's going to be available everywhere. Paperback, hardcover, audio book, and ebook. Pre-orders are available on Amazon and Barnes & Noble,

but anywhere that you get your books online. And if you're local in New York, you've got a couple of bookstores here that plan to carry the book as well. So yeah, please find me on LinkedIn and would love to connect. Well, congratulations. Thank you. Thank you. Thanks for joining Tech Arena. Subscribe and engage at our website, Techorina.a. All content is copyright by Tech Arena.