naomi white leader assistant podcast

Naomi is a senior executive assistant at Cohere, a provider of cutting-edge NLP models that solve all kinds of language problems; including text summarization, composition, classification, and more.

In this recording from one of our weekly Leader Assistant Zoom Chats, Naomi and I discuss language models like ChatGPT, automation and how it all can help us in our assistant roles.

Join the Leader Assistant Circle Community for FREE to watch the video of my conversation with Naomi, and get a copy of her slide presentation.


Naomi White Leader Assistant Podcast


Naomi is a senior executive assistant at Cohere, a provider of cutting-edge NLP models that solve all kinds of language problems; including text summarization, composition, classification, and more.

Naomi has spent the majority of her career in the technology sector, with a particular passion for artificial intelligence. Prior to Cohere, Naomi was at DeepMind, the London-based artificial intelligence research company owned by Alphabet.

A primary focus of Naomi’s is to learn and grow; mentoring others who share similar passions and curiosities about making work, life, and the world a better and more equitable experience for everyone.

ezCater large logo

ezCater is the nation’s most trusted provider of corporate food solutions — the best way for companies to order food for daily employee lunches, meetings, and events of any size or budget. ezCater’s simple-to-use platform provides a network of over 100,000 restaurants nationwide, business-grade reliability, food spend management tools, and 24/7 support from their highly trained customer service team.

To explore corporate food solutions or place a catering order, visit


To learn more about how you can join growth-minded Leader Assistants, check out our Leader Assistant Premium Membership for ongoing training, coaching, and community.


Download the first 3 chapters of The Leader Assistant: Four Pillars of Game-Changing Assistant for FREE here or buy it on Amazon and listen to the audiobook on Audible. Also, check out the companion study guide, The Leader Assistant Workbook, to dig deeper.


Check out our constantly updated schedule of events for admins and assistants at


Join the Leader Assistant Global Community for bonus content, job opportunities, and to network with other assistants who are committed to becoming leaders!


Subscribe to The Leader Assistant Podcast so you don’t miss new episodes!

You can find the show on Apple PodcastsSpotifyGoogle Podcasts, Pandora, and Stitcher.

Join my email list here if you want to get an email when a new episode goes live.


If you’re enjoying the podcast, please take 2 minutes to rate and review the show on Apple Podcasts here. Each review helps me stay motivated to keep the show going!


Jeremy Burrows 0:00
Hey friends, welcome to The Leader Assistant Podcast. It’s your host Jeremy Burrows and this episode is a recording from one of our weekly leader assistant zoom chats. We have these every Wednesday afternoon at 2pm Pacific Time 4pm Central Time. And in this one we had a special guest. Join us to chat all things language models, natural language processing, automation and chat GPT. So I hope you enjoy this conversation I had with Naomi white. And if you want to join us for a future leader assistant zoom chat, just join our community for free at leader And then click on weekly zoom chats on the left navigation bar, and you can join us on Wednesdays

Podcast Intro 0:59
The Leader Assistant Podcast exists to encourage and challenge assistants to become confident game changing leader assistants

Jeremy Burrows 1:13
with so much on your plate, wouldn’t it be nice if ordering food for the office were easy and reliable. My friends at easy cater are workplace catering pros helping you find food for everything from daily employee meals to staff meetings and special events. With easy caterers network of over 100,000 restaurants nationwide, you’ll have a huge variety of options near you for any group size, dietary need or budget. Your food arrives on time as ordered all supported 24/7 by easy caterers team of experts visit to find out more. All right, welcome to the leader assistant zoom chat, and today is a very special conversation we’re going to have with Naomi White. Naomi is a senior executive assistant at a company called cohere. And cohere is a provider of cutting edge natural language processing models that are solving all kinds of language problems including text summarization, composition, classification, and more. If it sounds familiar, that’s probably because it is familiar with the hot topic of chat GPT these days, but Naomi has spent the majority of her career in the technology sector with a passion for artificial intelligence. Prior to cohere, Naomi was at Deep Mind, which was a London based artificial intelligence research company, owned by alphabet slash Google, you’ve probably heard of Deep Mind if you’ve done any sort of research on AI and all that. And then the primary focus of Naomi’s is to learn and grow, mentoring others who share similar passions and curiosities about making work life in the world a better and more equitable experience for everyone. Naomi, and I connected, I believe through LinkedIn, initially, but this is our first time actually meeting each other. So I’m excited to chat with her all about Chet GPT. And she’s gonna share her screen and walk through a little bit. But welcome, Naomi, thanks for being here.

Naomi White 3:31
Awesome. Thanks so much, Jerry.

Jeremy Burrows 3:33
The task, what time what time of day, is it? Because you’re like on the other side of the world. And

Naomi White 3:39
technically, I feel like I’m on Eastern time. Most of the time, my company is a Canadian company. But I’m based in the UK and it’s 10pm here. So it’s bedtime for me after this.

Jeremy Burrows 3:51
Fair enough. Thanks for Thanks for doing this. At your bedtime.

Naomi White 3:56
Yeah. So before I dive into the tech itself, I wanted to give you a bit of history, which to me really adds perspective to the world of AI and might alleviate any anxiety you have around it. So if we just get back in time, the invention of writing and the printing press led to a flowering of science, literature and arts, which are all amazing things that we truly appreciate now, however, people are oftentimes skeptical of such new inventions. So let me pose the question to you might machine learning machine understanding of language be in the same realm? So there’s a really good quote from a book called the writing of the gods, the race decode the Rosetta Stone? And if you look it up later, it’s by Edward donec. And it tells the story of the invention of writing. The first symbols that humans invented to represent sounds is a super profound development. And to roughly quote from the book, the first writing must have come as a surprise and even a shock. People have to stop and wonder what this new thing could be, how it worked and what it was good for. For. So yeah, this is kind of seen as the birth of the modern world and what we rarely realize it what it is what kind of made possible for us. So yeah, we can pop into the next slide. And let’s get back to more recent times. So we’re in this new era, every 10 to 15 years, humans interaction with computers is fundamentally and irreversibly transformed. So we started off with kind of this big, bulky Mac computer back in the day, eventually progressed to kind of the first web browser mosaic. And then eventually on to the iPhone, and the snazzy new things called apps, which were like mind blowing. But what’s the next big thing? And what I think and a lot of people kind of in this field thing is that NLP natural language processing is the next big thing, Sash is the big thing. And that kind of includes chat GPT. So if we go on to the next slide, there’s just a little article there as well about kind of chat govt and how it’s kind of reshaping how we interact with computers. So if you pop into the next slide, you’re sitting here thinking, great, thanks for the history lesson. That’s all very well and good. But how can you leverage this kind of tech as an EA. So the biggest model out there right now is chat GPT, which you may or may not have heard of, there’s a lot of hype around it. It was released at the end of last year by open AI. And it’s what is known as a conversational chat model or a chatbot. The models generate conversational text that is relevant to hearing and knowledgeable given a prompt, the prompt being whatever you typed into the Chatbot. It’s the same type of tech that’s used in voice assistants. So think about your Google Home, your Alexa or Siri on your phone. Conversational AI works mainly in two parts. So machine learning and NLP. And machine learning means that the technology learns and improves, the more it’s used, it collects information from its own interactions, and then it uses that information to improve itself as time goes by. The result is a system that works better a few months after you kicked it off. And even better than that a year down the line. The second is called natural language processing, or NLP for short. You might sometimes see that acronym for Neuro Linguistic Programming, but we’re talking about a totally different thing here. So this is the process through which artificial intelligence understands language. Once it learns to recognize words and phrases, it can move on to the natural language generation. So creating language, and this is how it talks conversationally. So for example, if we took a chatbot on a shopping site, a customer might message asking you for information on the delivery of an odor. And then the conversational chatbot will know how to respond. And it will do this based on prior experience answering similar questions because it understands which phrases tend to work best in response to these types of questions. This isn’t a crazy new piece of tech, you’ll have been able to find it on websites, online stores, social media channels for years now. And the AI tech issues here to speed up and streamline. Streamline you so it’s here to help you. We move on to the next slide. So what can I use it for? So if we’re thinking in the context of chat, GBT, and shortly I’ll sort of show you a couple of actual examples. But it’s really, really great. Well, the over kind of overarching theme here is that you can use it to save time, which is great when you got 1000 things to do as an EA. It’s also great to create things that you’re not kind of a subject matter expert on. But I would definitely do this with caution. So one of the things that you can use it for is content creation, so drafting emails, blogs, social postings, they can really do like a whole range of different things. So for example, on email, you can give the model information such as the subject of the email, sort of a few bits of key information, and then the model can generate a response specific to the prompt. And the model can be training your past emails and responses. It will improve usability over time. So when you sign up to chat GBT, you will have your own profile and it will be kind of learning on that engaging with kind of the information that you provide it and learning on that over time. Another thing you can use it for is consolidating information. So running through transcripts, meeting notes. You can also use it for planning travel, helping to create itineraries looking at accommodation options, all of that type. thing, idea generation and research. So you’re thinking about your exact going to a brand new city, and you want to find out a bit of information about it. So they’re going to New York, I want to find out the best restaurants for fresh fish there. And you can type that in and get some really good responses. Another thing it’s really good for is data entry. And also generating reports as well. A couple of words of caution here, though, always proofread it can sometimes the model can spit out some very interesting responses. And also can spit out outdated, outdated information, and also sometimes toxic content. So toxic content might be, let’s say, it’s International Women’s Day. So maybe not sort of gender equitable response, what you can do that is flagged that to open AI, there’s like a thumbs up thumbs down. piece on there, we’re on chapter UBT. And you can just report things there. It’s a model, it’s learning all the time, your feedback is going to help it. Yeah, so we’re moving to the next slide. There are a couple of examples here that I typed in earlier, super easy to sign up to chat GPT, you kind of you do, you have to enter a bit of information on yourself, such as your phone number and your email address. But if you’re happy sharing that with open AI, you have access to all these different things. So is basically just typing in a query a prompt in here. So for me earlier, I was like, I want to write an email on a corporate tone, letting staff know that has been changed. So they’re working schedules. It’s super cool. You watch it, and it kind of types this up in like real time, obviously, someone is typing something back to you. And it creates a very comprehensive email back. Also for the idea generation on the other side. I wanted to find out about work social over 10 people in London, this is like just a snippet of the response. I’m not kidding, you went on for ages. And this was great. Because sometimes my mind is just like, I can’t think of anything right now. I’m brain dead. I just want to, I want someone else to help me out. And this is like a great, great resource for that, and can come up with a bunch of different responses. So yeah, these are just a couple of examples. I will say. So people wondering, is chat GBT going to take over the world? No, probably not. I think it’s going to be here to like, help us and work hand in hand with the assistant profession, our kind of our role is very much about knowing your lead, really about like human touch, personalized approach. And really kind of, I think, chat GPT will be like the differentiator between like a good and a great assistant. If you can use this to your advantage, then brilliant. Like, you can use this to kind of optimize save time. And it’s a really great resource. I’ve talked for a really long time, just a little bit about okay, here. We also have a platform very much like chat GPD, slightly different. But you can sign up to this for free as well. I will send a link after share with Jeremy. And again, it kind of spits out information very much in the form that chat GPT.

Jeremy Burrows 13:28
Awesome. Thank you, Naomi. Yeah, it’s fun, because, you know, I work for an artificial intelligence support automation company called capacity. And we’ve been building in natural language processing bot, since you know, the end of 2016. Really, we launched in January of 2017. And so I’ve been kind of soaked in this world of natural language processing, language models, all that stuff for the past six plus years. And now everybody is all gung ho talking about Chet GPT. And like, this isn’t really a new thing. It’s just a pretty large language model, like you mentioned, that has done a really good job of being practical. But there are certain things and there’s a couple of people will open it up for questions. It’s couple of people in the chat that said, you know, how, how confidential can it be if I’m entering private information? And since it’s not connected to the internet, how do we account for the most current data and feedback? These are great questions. I think one thing that you’re seeing is, you know, we’ve even had, you know, security teams say, Hey, you’re not allowed to put client data in to chat GPT because it’s a security risk. And so there’s certain things you have to be like you mentioned to just be cautious about but I think that if you’re looking for general things like like the examples you show showed about London, you’ve got to work of it with with people in London, what are some good ideas or if you’re trying to just draft a email about x and y, and then you want to customize that and take that as a as a step one, to customize to your own private company information internally, it’s a great way to have instead of staring at a blank screen, you can stare at a draft. That’s actually pretty helpful. So anyway, what are your thoughts on those couple of comments there about confidential information? And how? How do we account for the most current data and all that?

Naomi White 15:43
Yeah. So I think first part focusing kind of the security aspect of things, obviously, input, anything you feel comfortable with, if it’s something that you don’t feel comfortable sharing, again, you’re sharing it with a company that will be using your data to train so just bear that in mind. And like Jeremy said, if you can use it for more of a general thing, or like, for example, for an email draft, you can get away with providing kind of minimal information that you can edit afterwards, if you’re having kind of this, this draft response that you can edit, you’ve got kind of the the base to kind of get you thinking that’s at least saved you some time there. So yeah, just use at your own discretion. I think in terms of you the information being up to date. Again, it’s a language model that is trained kind of on data up until about 2021. So some things will be outdated. However, you might have heard some of kind of the hype around this being chat piece of that they’ve also bought out, there’s a waiting list that you can sign up for access to it. And that the intention there is to have a conversation with the internet search. So it will be able to pull more recent information. But that will kind of maybe be not quite the same as chat GPT and not be able to have the same capabilities. So yeah, obviously bear that in mind, when you are asking for more kind of like recent information, like if you’re asking it for like, let’s just say the prime minister in the UK changed yesterday chat, GBT probably won’t be able to tell me the information. So yeah, again, use at your own discretion. Just bear that in mind. Just be really cautious around kind of, if you’re then using it kind of, I guess in like a company wide forum.

Jeremy Burrows 17:35
Yeah. And Tara has a question about a similar note about, you know, seeing chat, GBT and AI touted as a good way to kind of clean up rough notes into something more finished. What I’ve seen in that in those situations, and we do a lot of intelligent document processing, or IDP capacity. And, you know, you can take out the personally identifiable information before, you know, we’ve got your templates, and you’ve got your kind of general forms, and you can take out the PII and, and still get a lot of helpful, you know, extract data from it and help build out different models. And then I think the, the other thought is, you know, something that we’re trying to do that capacity in particular is, you know, we, we don’t send any of our customers data out to chat GPT, or open AI, or whatever. And we have our own language models internally that we’ve that we’ve built out over the last six, seven years. And some of them are based on, you know, there’s some stuff that we’ve done, and we’ve trained or with our own data, there’s stuff that we’ve used on this open source, but all of that data that gets kind of put into that model is not sent out to a third party. And so there’s, there’s certain companies that do send it out to the third party and then there’s certain tools that don’t and so you just have to be really careful and make sure that you’re doing your your homework. I did, I did appreciate your slide Naomi with the history and I love that started in 1984. That’s the year I was born. So I’m not I’m not as advanced as NLP, but some days I like to pretend it’s anybody. Oh, here we go. Dana has a question in terms of travel does it integrate with booking tools?

Naomi White 19:33
I am not personally aware of companies that have used chat GBT to integrate with booking tools, but I imagine that something that will come very, very soon now plenty of companies that are kind of working using kind of the chat GPT platform to be able to provide services. So I definitely kind of get keep a lookout

Jeremy Burrows 19:55
for that. Yeah, and I think that more and more, you’re seeing Seeing and you’re seeing in the news every day, you know this there’s a company called intercom we we have chat GPT integrated into our support tool and then drift came out with something this week, same deal. And then, you know, Microsoft investing in open AI and all this it’s like, all these in Salesforce just came out with something about having a, essentially the the open AI or the chat GPT, NLP, whatever built into their tool so that sales team members can have basically AI suggest ways to respond during sales calls or to respond to to emails. And then I think that Tara has a good question, what’s been your most satisfying, unexpected result with AI? Naomi, do you have any interesting use cases that you’re like, oh, man, that saved me a lot of time other than the examples you share?

Naomi White 20:55
This is a really good question. Actually, I’m trying to think, I think I’m gonna answer this question in two parts. Something that we do at co here is a lot of kind of annotation input ourselves, and the language models that we love to play around with them, no matter like who you are, if you’re a machine learning engineer, or if you’re a member of like, recruitment staff. And just generally seeing the models improve over like the space of like, a couple of weeks, we had this kind of like big challenge over the holiday period at the end of last year. Like training the model, like asking it to generate various different things. Sometimes it was about some very interesting and kind of quite funny responses. But kind of going on it, let’s say like two weeks later, and you’ve got like a much more coherent, like, for example, generating like a poem, which I think is just a really fun one, because I’m like, oh, like, you just don’t think of AI to be like that creative. And when it comes up with a really cool response, you’re like, great. So that’s kind of maybe one part of it. And something that I am also super impressed with is a tool called reclaim. And you can input things that you need to do in your calendar. And I the two leads that I work with it cohere use it, and it basically drops things into their calendar, I’m always consistent up there are certain gripes I do have with it. And I don’t think it’s ever it’s not at the level of an assistant. But sometimes where it can kind of specify on the calendar like, this is where you should be like doing exercise and your day or whatever. And it can recognize a really good timing for even though it’s not kind of given a specified slot. That can be really kind of, yeah, it kind of blows my mind.

Jeremy Burrows 22:45
Yeah, I’d say my kind of WoW, use case is with my podcasts. So there’s a company, I think it’s called, I don’t remember what’s called maybe swell AI or something, have to look it up. But basically, you can, I can upload my podcast, audio file. And it used to be where I’d have to spend about $1 $2 per minute, just to transcribe my podcast episodes, which is why I don’t have full transcripts of my of my 200 Plus episodes on the on my website, because that’s a lot of money. And now, there’s this tool, and I can upload the file, it’ll spit out a transcript, AI generated. So it’s not as clean as some of these other transcript tools where there’s an actual human going through and cleaning it up. But it’ll spit out a transcript in a matter of seconds. And then it’ll also spit out a blog post based on the based on the episode based on my podcast episode. And then it’ll also have social media posts, like LinkedIn posts based on that episode. And then it’ll have a little timestamp outline. And I mean, it’s just crazy. All the stuff that that used to take a few hours, at least, to do for each episode. I can now do several episodes in a matter of minutes. And it’s just pretty fascinating how these AI tools are. So anyway, that’s great. Thank you so much again, Naomi. I know we’re out of time. I appreciate you taking time to share your your journey with language models. And I know I was interesting and helpful. And it’s always good for us assistance to be curious about all the new latest technology as we go go about our careers and try to level up. So thank you everyone for joining hope you have a great rest of your Wednesday. Thanks again, Naomi. And hope you hope you have a good night’s sleep.

Naomi White 24:48
Thank you.

Jeremy Burrows 24:50
Thanks everyone. Have a good one.

Podcast Intro 25:01
Please review on Apple podcasts

Download FREE Chapters