AI tools discussion
Hey! Curious to hear what AI tool you find most useful in your day to day work
I’m using notebooklm somewhat and I’m subscribed to gpt. However, I’m considering subscribing to some of the other tools instead/as well. I’m mostly using the models for research, summarizing large text files, and conceptually discussing my approach to an analysis - I’m of course not sharing any sensitive information with any provider. I find gpt quite useful for summarizing text and helping with excel formulas etc., but don’t find it very useful for eg. helping with analysis - I feel like it’s falling short on reasoning and sector specific knowledge, understandably.
How’s your experience with the pro-models from Perplexity, Claude, Gemini etc.? What use cases do you use the different models for, and how much benefit have you gotten from the pro vs the free models?
Based on the most helpful WSO content, AI tools like GPT-4 and others are increasingly being used for tasks like summarizing large text files, generating financial models, and even automating repetitive tasks in finance roles. GPT-4, for instance, has been highlighted for its ability to handle natural language tasks efficiently, such as interpreting comments and making changes in PowerPoint or Excel. However, as you've noted, it can fall short in areas requiring deep reasoning or sector-specific expertise.
Regarding other tools like Perplexity, Claude, and Gemini, there isn't specific feedback in the WSO threads about their pro models. However, the general sentiment is that pro versions of AI tools often offer enhanced capabilities, such as faster processing, access to more advanced features, or better integration with workflows. For example, tools like BloombergGPT are being tailored specifically for finance, offering sector-specific insights and analysis capabilities.
If you're considering subscribing to additional tools, it might be worth exploring their unique features and how they align with your specific needs. For instance: - Perplexity: Known for its ability to provide concise and accurate answers, it could be useful for quick research tasks. - Claude: Often praised for its conversational abilities, it might be better for conceptual discussions or brainstorming. - Gemini: If it offers sector-specific insights or advanced reasoning, it could complement GPT for analysis tasks.
Ultimately, the benefit of pro vs. free models depends on the complexity of your use cases and the value you place on time savings or enhanced functionality. If your work involves a lot of repetitive or data-heavy tasks, investing in a pro model might be worthwhile.
Sources: Will robots replace your consulting or financial career?, GPT-4 & Microsoft Co-Pilot, https://www.wallstreetoasis.com/forum/hedge-fund/machine-learning-taking-over-hf-research-analyst-roles-in-near-future?customgpt=1, India's Demographic Dividend | The Daily Peel | 4/20/2023, Tools of the Trade: Pro Tips
Following. Not sure why these tools aren’t being used more in this industry.
I agree, but what tools, and how?
I can see using them to assist with certain steps, fetch facts/research, etc. but nothing that is more comprehensive.
I created a chrome extension in my free time that leverage AI to analyze the articles you're reading and identify potential stock market implications.
You can google it's called Investabloom.
Tbh, it's not really difficult to use the API of gemini or gpt to create what you need. A lot of potential over there
I like using Napkin.ai here and there. It turns text into visuals pretty quickly.
The reason no one is using tools is because most people on here work on the sell-side (or big buyside) and such firms will not let us use any sort of non-firm-sanctioned AI. I mean heck my firm even blocks us from using pretty much any website where text is processed (word counters, translators etc.)
I used this app called FinPrepAI (can find it on the Apple App Store) to help me prepare for my technical interviews and it was pretty good. They had a bunch of questions and has AI providing feedback on your answers.
In PE. My issue with using AI for actual deal work is that it's still too black-box. I need to go check the work and am never certain if the answer is correct or not. End result is I still have to do the research so why bother?
If anyone has a way around this issue I would love to chat.
pm'd
I'm working on something, but I still have not figured out if to lean on public or private markets. I would appreciate some feedback if you've got time (aurelioterminal Dot com)
This is my concern about AI. How do you know if the information is correct? If you have to verify everything, it might not be worth it. I do think it is better suited to people who are more knowledgeable because he or she should be able to tell if the information is wrong and therefore the efficiency would be a positive. I also think that it is not so good for people who are less knowledgeable because these people might assume that everything is correct when it is not
Yeah agree. Smart people who have common sense will probably spot major errors/inconsistencies, and know for which tasks they should (not) use AI. I'm more concerned about people who generally blindly believe what they are being told.
Does anyone have any tools they use for proofreading and catching mistakes?
I am curious too.
If you’re interested in this area, I’m building an Enterprise AI for IB/PE that automates workflows and streamlines data silos across the deal lifecycle. We’re essentially building an intermediate layer on top of existing infras (e.g., databases, CRMs, data platforms), enabling deal teams to perform mundane tasks in seconds instead of days.
We’re also rolling out an individual plan to explore a B2C model for non-sensitive but time-consuming tasks. I'd love to hear what you are struggling with that you might be willing to pay for the solutions (and roast me with our demo).
also building in this space. let's chat
AI is great and super powerful for coding work that analysts need to do either in excel or other applications. I would be cautious using it for analysis of text and summaries given the frequency of halucinations. Given that most bankers don't have expertise deep enough in most of the areas they cover to identify all of the potential issues. AI has power, just be careful with how you utilize it.
I use this and it's quite good! https://www.ninetyfive.gg/
Hey everyone, here are my thoughts on some of the AI tools.
Claude & OpenAI
I find that Claude and OpenAI excel at research-based activities. They're my go-to tools when I need to find and track companies, process documents, or write various content snippets. However, for data-heavy tasks, I don't really think those are the best tools.
Gemini & TinyDash
I actually use Gemini and TinyDash (which I'm beta-testing from this startup) for more calculations, modeling, and data-intensive work. They both are really great in terms of creating outputs - for example, I've made countless data dashboards of company financials with them.
Happy to share if you'd like access!
I’ve tried a mix of tools for research and content workflows, and honestly, it depends on what I’m trying to get done. Claude is solid for longer context summarizing, especially when I’m dealing with articles or documents that are kind of dense. Gemini is okay but still feels like it’s catching up in terms of reasoning. I agree that GPT is great for initial drafts, catching structure problems, Excel-type stuff—but lacks nuance when the topic gets too niche or technical.
When I’m working on music-related content, especially background scores or quick loops for video projects, I’ve been using this platform that lets me auto-generate custom tracks by mood and style. It’s saved me a bunch of time hunting down royalty-free audio that actually fits.
Been playing around with Decisional (app DOT decisional DOT com) for last couple of weeks - it's pretty slick in generating a pretty good memo. REALLY good at helping verify what it quotes - gives a bounding box as a citation exactly in the document, for anything it quotes which makes it super easy for me to verify information. Can also understand spreadsheets and excel pretty well - you don't need to screenshot a spreadsheet like you have to in chatgpt. You can download the memo as a powerpoint or create financial models but I haven't tried those features yet! Def worth a try if you were unhappy with notebooklm or chatgpt and perplexity like me.
Sounds promising, I'll give it a try
Aut dolores commodi in nam iste. Deleniti rerum ex rerum fugiat voluptates pariatur. Perferendis sunt quos veritatis quos omnis est dolores enim. Consequuntur odit eum laborum ratione cum. Aut voluptatem aliquam quaerat suscipit ullam.
Ut ea et expedita rem. Repudiandae animi quia animi illo in impedit voluptate sunt. Facere animi accusantium totam aut nihil quis vel. Minus voluptatem cupiditate esse ipsam doloribus.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...
Et ut dolorem saepe ut dicta sint qui. Atque et rerum similique fugiat minus voluptas. Quas soluta accusamus eos est. Voluptas magnam culpa aliquid adipisci velit unde. Ut eos eos quos velit blanditiis adipisci.
Saepe pariatur et omnis non exercitationem. Amet quidem natus inventore voluptate voluptatum aperiam sunt.