<\/p>\n
Kaleigh Moore is a 12-year freelance writer and editor. She\u2019s contributed to Shopify, Forbes, Vogue, Adweek, and various B2B providers, among others. Like all of us, she\u2019s now grappling with the promise and limits of AI.<\/p>\n
She calls AI a \u201cfire hose of information\u201d that greatly increases efficiency. She also cautions on what it cannot do, such as interview humans or learn from experience.<\/p>\n
She shared those views and more in our recent conversation, including her preferred AI platform, use cases for entrepreneurs, and getting started with AI-driven composition.<\/p>\n
Our entire audio is embedded below. The transcript is edited for clarity and length.<\/p>\n
Eric Bandholz:<\/strong> Give us a rundown of what you do.<\/p>\n Kaleigh Moore:<\/strong> I\u2019m a freelance writer and editor. I\u2019ve worked with all kinds of B2B and SaaS companies within the ecommerce ecosystem over the past 12 years.<\/p>\n One of the first case studies I wrote for Shopify featured Beardbrand. That was over a decade ago.<\/p>\n Bandholz: <\/strong> Does AI commoditize writing?<\/p>\n Moore: <\/strong>The emerging AI composition tools are incredible; they greatly increase efficiency, especially for the tedious parts of writing, such as ecommerce product descriptions<\/a>.<\/p>\n But remember that AI tools operate on existing information. They are not creating something new. Human composition provides original perspectives \u2014 experiences, thoughts, and feelings.<\/p>\n Do we care about those perspectives? Some people care a lot. I\u2019m a journalist first and foremost. I want to do my own homework, fact-check, and make sure I\u2019m putting out the best of whatever my name is on.<\/p>\n I worry about young people and how they\u2019ll use these tools. I\u2019m 37 years old. I grew up in a largely pre-internet, pre-social media time. I hope we always have human experience and interaction \u2014 talk to each other, go to coffee. To me, AI is a nice supplement, but it doesn\u2019t replace person-to-person interaction.<\/p>\n It\u2019s been interesting on the hiring side of things. I occasionally look at full-time in-house writing roles. Over the last 18 months, many of those roles have shifted to require AI operational skills, to hop into an AI tool and craft something. If you are not hands-on with these tools, you won\u2019t even get an interview. So the ability to learn the tools and be curious about them is an important skill now.<\/p>\n Some people say AI is just a bubble, but I don\u2019t think so. It\u2019s too powerful.<\/p>\n Bandholz:<\/strong> How does a writer or entrepreneur learn and apply AI?<\/p>\n Moore: <\/strong>It\u2019s a fire hose of information every day. I approach it as a journalist. A key skill is developing very strong prompts. The more advanced we are at prompting, the better the output.<\/p>\n Beyond that, accept a willingness to learn the new functionalities. It can be intimidating, what with all the new tools.<\/p>\n Anthropic\u2019s Claude is my go-to platform. Claude\u2019s outputs are very good. Anthropic\u2019s entire stance is open-source and transparent. The company prioritizes ethical concerns and data privacy. For me as a writer, Claude is the best. It\u2019s also a great place to start.<\/p>\n Generative AI platforms such as Claude can help entrepreneurs and marketers with promotional emails, social media posts, and LinkedIn articles, among other applications. The platforms will remember a voice and style from existing content.<\/p>\n Bandholz:<\/strong> How do you train AI in that way?<\/p>\n Moore:<\/strong> I\u2019ve been doing it for my own work. I feed Claude good, strong examples of my published articles, case studies, and guides to provide points of reference. It\u2019s akin to informing a new hire, say a junior writer or copywriter.<\/p>\n The aim supplying a lot of very specific guidelines. Lots of dos and don\u2019ts. Use this word; don\u2019t use this one. Input it once, and the AI never forgets, unlike humans. And I can update it over time, which is essential. Although more examples are not always better. AI can get confused by too much information.<\/p>\n A user\u2019s top 10 tweets would be a good, limited data set to start with, plus general instructions on likes and dislikes. Teach AI in the same way you would a human.<\/p>\n Say a merchant wanted to publish a blog post. I would enter a full brief, such as the targeted keyword<\/a>, the audience, brand names to avoid, and data sources to cite. There\u2019s quite a bit of heavy lifting involved just getting that prompt ready.<\/p>\n Certainly, the merchant could give it a paragraph and request a 500-word blog post on X for this audience. She would get a pretty good output.<\/p>\n But give it the full brief<\/a>, and she will likely receive a much better result, requiring little editing or tweaking. It\u2019s often a choice of spending time editing the output versus preparing the brief.<\/p>\n For me, editing AI text usually comes down to my writing preferences, though much of it is fact-checking. AI hallucinates; it makes stuff up. It will cite data that doesn\u2019t exist.<\/p>\n Moreover, AI cannot interview or speak with an expert. We have to integrate those afterward.<\/p>\n Bandholz:<\/strong> Where can people follow you and reach out?<\/p>\n Moore:<\/strong> My site is KaleighMoore.com<\/a>. I\u2019m on X<\/a>\u00a0and LinkedIn<\/a>.<\/p>\n<\/p><\/div>\n