Using AI Chatbots to Update a Resume
TL;DR - No single general-purpose AI tool did a perfect job helping me rewrite my resume. They acted as tools to unblock my creativity and help me see things I am blind to, but none of them could provide a perfect output without my involvement. That said, taking lessons from each one, I feel I can better revamp my resume without paying for a resume-specific service.
WARNING: This is crazy detailed. Use the headings to skim to get what you want/need from it.
We all hate the task of writing our resumes. We all need help, even if we are excellent at resume writing; we need an outside voice to sense-check. I have restarted my job search, but before I can get serious, I need to update my resume. After a terrible phone screen (I knew I wasn’t a fit for the role, but I took the call to jump-start my job search), I decided today would be dedicated to rewriting my resume so I could apply to roles I want. I knew rewriting would help me refine my messaging (which was awful during the phone screen) before I did anything else. In my case, I needed to bring my consumer obsession to the forefront.
What I Considered
Before deciding to do a tool comparison, I was considering (1) hiring a professional writer, (2) using an online service (think jobscan.co) or (3) using all-purpose AI tools. I decided that, since PMs need to learn how to use all these new AI tools, I should use this task to test them out. For the rewrite of my resume, I used OpenAI’s ChatGPT, Perplexity, Anthropic’s Claude, and Google’s Gemini.
Where They Excelled
I loved OpenAI’s ChatGPT for getting straight to the point without any difficult work on my part. It explained its logic concisely and then gave the practical answer, so I knew why the changes were being suggested. There were often three or four options to choose from, based on different spins I might want to consider. ChatGPT’s suggestions on what to do next led me down a different path, but they helped me feel like I had a partner. It was easy to simply follow ChatGPT’s suggestions to achieve what I wanted. I only stopped because I hit a firewall, which I later cleared. I felt a sense of accomplishment, and I didn’t have to weed through the output to piece things together. But it wasn’t pushing me where I needed to be, and I didn’t realize that until I got to Anthropic’s Claude and then eventually Google Gemini.
I found Perplexity was better suited for topics where I want help doing the work myself. Perplexity started spitting out answers without asking me to provide it with context. I started by giving it quick context, expecting a question about more context to save tokens. I was immediately worried it was wasting tokens on things I didn’t need, as it ran for 3 minutes. By contrast, ChatGPT realized it needed to ask for more information before it provided a research compendium. Since I started using the more conversational tool, I didn’t realize Perplexity required very specific first prompts to prevent analysis where it wasn’t needed. Even if I go to the default version of Perplexity, it is all about giving me “How to” instructions, but I didn’t want to do it all myself. I was looking for quick help and a nearly finished product, not instructions on how to start from scratch. By contrast, ChatGPT had told me much the same things Perplexity do, but it was in context with examples that were practical for me.
Claude is obviously meant to help writers. The output screen with a document is almost immediately placed on the right-hand side of my screen. It was the third general-purpose AI product I tried for the resume rewrite purpose, and even though it provided a format I wasn’t used to, I loved it. The output was easy to read. But I could see it was using my original work as a strong baseline and so it helped me create what Google Gemini eventually called my “master resume.” It retained the more executive-level formatting.
Google Gemini surprised me and won me over quickly. I typically encounter Gemini as the AI summaries on google.com. But I was impressed by Gemini’s ability to provide a structured approach, but much more concisely than Perplexity. It started, similar to Perplexity, with the goal of empowering me with best practices, but it kept things concise and straightforward. Then, like ChatGPT, it asked me for more information, but was extremely specific that all three things were required, which forced me to do the work I needed to get better answers: find a sample job description. If I were going for an L6 or below IC role, I might have been happy with Gemini’s final format, but it does require some fact-checking.
Biggest Fails
ChatGPT Couldn’t rewrite my entire resume and retain non-traditional elements, like “Advisory Experience” at the bottom.
Perplexity The complete output of a full resume wasn’t senior level.
Claude It was longer than it claimed. I was going to run with a modification of its output until Gemini woke me up.
Gemini Created a Google doc that wouldn’t open. I had to ask it to put the results in the chat window.
Universally Annoying
Data inputs. I wanted to have a quick setup discussion. Providing information before the models wasted tokens. But I quickly discovered that I needed to just put all my context into the first field. I probably could have written my initial request telling the chatbot that I was going to submit my data in three parts to give context, but I eventually just gave up on that. I ultimately just did a sloppy “copy + paste” and attached documents and to the initial ask.
Quality Assessment
I attempted to evaluate each product in a quick grid. This is difficult to do as it is heavily influenced my sense of asthetics, in many respects.
Grading Scale: Low 0-3. Medium 4-6. High 7-9. Clear Winner: 10
.
Public Links to Chat Experiences
Each platform will let you share your chats and/or the output with a public link. So here you go.
ChatGPT Click here Heads up, I started the convo on Vibe coding and can’t remove it, so scroll just a touch to get at what you want.
Perplexity Click here Look at ALL the sources…
Claude Click here Shows chat and resume
Google Click here I could convert some outputs to clean webpages, which was impressive.
Diving In A Bit More
ChatGPT
Key Observations: Mixes concrete, customized examples along with guidance on where to go next and explanations of why the suggestions are different. I probably need to adjust these just as much as the other outputs, but I know quickly why the changes were made. Even when I don’t love the wording, the explanation of the intent works as a thought partner to push me in the right direction.
Offers Help and Suggests Next Step
If you’d like, I can help you: (1) Rewrite a resume bullet to emphasize leadership or metrics (2) Benchmark your resume language against a sample PM job posting (3) Build a tailored resume for a specific company (e.g. Google, Meta, Amazon). Just drop in your current resume text or a job link.
Recognizes It Needs More
Great — please go ahead and paste the current resume entry you have for your role at Amazon Fire TV, and I’ll help you rewrite it for clarity, impact, and alignment with top product leadership roles.
Explains Why it Works or How Improved
Strong verbs (“defined,” “partnered,” “authored”) convey leadership and ownership. Quantified impact is prioritized early in bullets and tied to business outcomes. Clear separation of initiatives gives space for each to shine. Terminology like "customer journey," "ad surfaces," and "companion app" speaks Big Tech fluently.
Offers Up A Different Spin
Let me know if you'd like a slightly more technical, people-leadership-heavy, or GTM-focused version — or if you're ready for me to help polish your intro/overview paragraph next.
Shortened from ~95 words to ~65 without losing meaning. Clearer structure: what you do → where you've done it → what you’re great at → what drives you.Punchier phrases (“scaling high-impact teams,” “delivering measurable outcomes,” etc.). More senior tone — reads like a VP+ resume summary.
Perplexity
Key Observations: When AI was new, I needed and wanted proof. Perplexity feels like I am talking to the 60+ year old teacher who explains the why and then makes me read all the research myself. While ChatGPT feels like a fun and practical coach who gives me examples that get me going, the examples teach me more in 2 sentences than the summary of 22 different sources.
Links to Sources
Lists anywhere from 2 to 22 sources, regardless of the size of the tweak.
Provides Generic Instructions
Key Strategies (1) Use Strong Action Verbs: Begin bullet points with verbs like led, orchestrated, spearheaded, drove, or coordinated to signal authority and proactive leadership. (2) Specify Teams and Stakeholders: Explicitly mention the range of groups you led or collaborated with: engineering, product, design, go-to-market, marketing, finance, and more. (3) Quantify Impact: Where possible, tie your leadership to measurable business results—such as increased revenue, improved product launches, or enhanced user engagement.
Claude
Key Observations: It was designed as a writer’s assistant, which is immediately apparent.
Keeps a nice record of the final document's versions, allowing you to go back and find things you might like from earlier versions.
It explains why it did what it did. Perfect! I've strategically trimmed the resume to fit exactly 2 pages:
Key Changes Made: (1) Core Competencies: Reduced from 12 to 6 most essential keywords (2) Professional Summary: Shortened while keeping key value proposition (3) Lending Club: Reduced to 2 bullets, kept most impactful achievements (4) Bottlerocket: Moved to "Earlier Career" section with just company/title/dates (5) Advisory Section: Condensed to 2 bullets, removed research practice detail
What I Kept: All major revenue impacts and quantifiable achievements, Leadership and team management experience, Technical expertise and platform experience, Most recent and relevant role details
What I Removed: Less critical bullet points from older roles, Detailed description of 8+ year old role, Some redundant keywords and phrases
This version maintains your strongest selling points while fitting the 2-page constraint. The "Earlier Career" approach is professional and shows your full career progression without taking up valuable space. Would you like me to adjust anything else?
Key Observations: It gave me the instructions. Then it told me three things I had to give it for it to help me. The key of which was asking for a job description to customize my resume to my role. Which is truly best practice in today’s environment.
Explained the changes inline. Showing Original, Proposed and Rationale
Original: Launched Yelp's North Star metric for use by 13 teams across three product organizations, helping the CFO to unite product and sales functions
Proposed Tweak: Defined and launched Yelp's "North Star" user engagement metric, unifying 13+ product and engineering teams around a single vision for the core consumer experience.
Rationale: This reframes the achievement from a business-centric one ("unite product and sales") to a user-centric one ("unifying teams around the core consumer experience"), which is exactly what this Tinder role is about.
Nailed the missing link for me. I get caught up in what is easy to talk about, or what I talked about last time I pitchec myself. Google was able to help me see how I was a fit (or wasn’t by things it made up)
Tool or Chatbot?
I will probably try a professional tool eventually, but for now, I feel I have a strong set of recommendations to work with to improve my resume. At the very least, I hope it helps you. If I try a resume tool or service, I will share my learnings.
Note: I did have to pay to upgrade some of the tools if I didn’t want to wait hours for more credits. I have paid versions of Claude ($20 per month) and Google Gemini ($20 per month), but not yet ChatGPT or Perplexity (yet). The results from all four products were huge leaps above the last time I tried the resume re-write process, nearly a year ago.
