Seeing a lot of noise around AI, tools and extensions and the power of this technology is indisputable, but I wonder if there are yet any tools applicable to the work of an analyst (thinking mainly powerpoint)?
Does anyone know any recently developed AI tools that can support finance professionals?
| +34 | ADHD ! | 19 | 13h | |
| +34 | Non-Competes Banned | 43 | 14h | |
| +30 | Being Christian in investment banking | 14 | 4d | |
| +20 | Moelis has the cutest Analysts? | 4 | 4d | |
| How do I become Sigma | 15 | 3d | ||
| +16 | Underage intern, drinking? | 7 | 4d | |
| +16 | Can someone justify why is million written mm or MM? | 15 | 3h | |
| +15 | Secretive vs Universal Prestige? | 8 | 9h | |
| +12 | What does the richest self-made woman you know do for a living? | 13 | 1h | |
| +9 | Winning - The Hindsight Insight We Actually Need | 1 | 2d |
Career Resources
Get instant access to lessons taught by experienced private equity pros and bulge bracket investment bankers including financial statement modeling, DCF, M&A, LBO, Comps and Excel Modeling.
Following
Bump
ChatGPT
I find my companies custom version of GPT-4 integrated into Teams is pretty useful
Custom models are definitely something to consider. Also, NLP tools can be used for research reports.
Excellent Use Cases:
Summarizing + “talking to” 10Ks, memos, and any long docs
Condensing outbound emails
Decent Use Cases:
Alternative search engine/research tool (often outright wrong but dependent on how niche questions are - can require investigation of sources which can negate any time saved)
Rudimentary data analysis (similar drawbacks to search but on the axiom of complexity)
Bad Use Cases:
Deck Building (no current tools have capabilities to create materials beyond organization of a storyboard - do not easily conform outputs to company/team standards and have minimum customization capabilities)
Client Level Model Generation (no tools are available to input assumptions and historical financials to create client level, reliable models - very rudimentary tools exist but reliability is unknown and would never serve an external purpose)
The deck building is getting better but it is the most difficult problem because of the widely different formats bankers use. But also likely the most painful problem.
Arkifi is working on AI tools for model generation. It should be good for building models on publicly traded companies because all of that information is publicly available and easy to get via APIs to factset/capIQ/bloomberg.
Still early!
Exciting times for sure! Will check out Arkifi.
While a deck building product for finance is hard to build, there is an opportunity for focused efforts on creating one. Right now all the tools I’ve used or previewed boast about simplicity and try to create a technology wall by making PowerPoint integration impossible or difficult. The companies are focused on a much larger market of non-finance and non-consulting users. A basic product that allows easy training on company templates to largely automate transforming storyboard/docs into visually appealing, coherent decks is possible. It has to be a combination of vision AI and gen AI though.
There are companies who claim to be able to do this, but I have not tested their capabilities as their basic products were not inspiring - bad outputs, hard to use, and only used gen AI. Would love to hear anyone’s experience if they have used a deck building product that successfully incorporated company specific deck styles.
Can you expand on how you go about the first two? (10k summaries/memos/etc + condensing emails)
For talking to documents, you simply upload them into one of the foundation models like ChatGPT and ask the document questions or for summaries. You’ll need the enterprise version to do this securely with non-public info but you can use the public version for anything else. The limitations here are if you are trying to perform calculations.
For condensing emails, the foundation models are really good at saving time and making your messages more concise and pointed. You simply input your text and ask the AI to condense or reword any long email or to summarize an attachment - can result some attachments being unnecessary all together.
Following
This whole AI craze is WAY overhyped - no one is benefitting from it right now besides the Semi's. It'll take a year or two and billions of dollars of capex later to realize such but that's the reality.
Once everyone realizes that everyone is triple ordering from NVDA, they'll come back down to reality as well.
It’s definitely hyped up but AI is being used in a real way. The question posed in the post was about IB use cases and there are not a ton of dramatic ones yet but there are plenty in other industries.
Gaming studios are using it to develop renderings and some have reduced designer headcount. Developers are using it to create code. Entire consumer outbound and inbound sales strategies are being augmented by AI. There are a flurry of drugs at the beginning stages of approval processes whose development was aided by GenAI. These are just four actual impactful use cases but there are a ton more.
working on a platform called docufi.ai -- we focus on CIMs, CIPs, Teasers, MPs, etc.
I am not in PE, but I did build an AI helper for myself. I have to read 70 conference call transcripts on earnings each quarter and provide notes to my PM’s. This helps me save a ton of time. It’s free. Google “DoTadda Knowledge”.
bump
what do you want AI to solve for your work?
I've heard about AI tools for predictive analytics in finance, which use machine learning to predict market movements, detect fraud, and optimize investment strategies. Even though I haven't tried them myself, I think the point is to spot patterns in financial data to help with decision-making. I don't think anyone should solely rely on AI tools for tasks of this type, but they can be helpful.
I read about content analysis with AI, very interesting and applicable across industries. For example, ChatGPT4 can understand natural language and extract insights from text data. Some tools even help prepare data by extracting text from web pages and formatting it into markdown, which makes it easier for AI to analyze. This markdown format is simpler and more efficient, especially when dealing with lots of data.
My general impression is - it really comes down to the user. From data input - what you gather and how you format it for AI to analyze to cross/fact-checking - how you use the results you get.
Odio sapiente est aut molestiae culpa voluptatem. Aut ut ab ab ullam quisquam totam adipisci. Eius rem dolore omnis dolorem quisquam. Repellat sed architecto eos explicabo rerum consectetur ipsa. Debitis facilis cumque eaque ut itaque rerum. Maxime est est nemo aut. Ut distinctio sed commodi tempore.
Facilis perspiciatis dignissimos ut commodi mollitia temporibus. Et consectetur similique consequatur sequi optio blanditiis fuga. Doloremque ad quidem sed et. Culpa molestiae vel praesentium suscipit facere. Nemo maiores mollitia asperiores inventore qui id autem. Consequuntur nesciunt debitis dignissimos deserunt.
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...
Ipsam est omnis repudiandae reprehenderit rerum. Enim qui tenetur et temporibus voluptatem explicabo. Aspernatur ut est omnis laboriosam veniam enim ea aut. Reiciendis dolores est sint vitae. Eum rerum recusandae velit alias quis. Voluptate cumque ut sint ab est nemo fugit.
Quis similique qui aut nihil omnis non quod. Ab distinctio non qui.
Totam neque vero voluptatum voluptatem veniam ut consequatur rerum. Qui minima sed qui qui.