To juniors,
Research associate jobs will largely disappear from this industry in 5yrs. Why? AI is quite good already and is getting better at a shocking pace.
There are AI tools I'll use today that write earnings notes, update models, take care of grunt work / heavy lifting on research. Some of these tools provide reports that are similar to that of the average analyst!
All to say, if you are at the associate level -- do not wait, try to transition to analyst ASAP. Either at your current firm or recruit externally.
I've thankfully transitioned to analyst end of last year, but I'm now looking at what gets replaced after associate -- that's going to be most analysts. Aside from superstar analysts (top 20-30% of analysts in industry today), the rest will eventually also go the way of the dodo. This is on top of shift to passive + fee pressure.
If you are a student, I genuinely mean this -- do not enter this industry (active LO equities). You almost certainly will not get a 30+yr career from here. I'm mentally planning to be in this industry for another 10+yrs (before I get kicked out), maybe an optimistic case would be 20yrs max. Don't know, but it's certainly not till retirement. After which I'm planning on doing something else, depending on how financially successful I've been.
There will always be some active LO equity seats, but with only ~40-41% of equities as active today (vs. passive) and already losing 1% per year in 20yrs that implies # of seats are cut in half. I'm not even factoring in fee pressure yet. Or AI. I would not be surprised one bit if in the US, we looked out in 2045 and only 20-25% of seats in this industry remain.
Hard for me to say I regret this industry since when I came in, everyone thought this was all very far away. And there's a few off-ramps out there. But frankly it would be incredibly foolish as a student to choose to enter this LO active equities (or brave, I don't know which). Sometimes I wish I went to med school lol
"There are AI tools I'll use today that write earnings notes, update models, take care of grunt work / heavy lifting on research. Some of these tools provide reports that are similar to that of the average analyst!"
What are these tools?
Not sure why the MS. You'd make a more convincing argument if we knew the names of these tools so prospective LO research associates, or those interested in breaking into the space, could be able to see how these exact tools will replace their jobs.
For example, Google's NotebookLM is a helpful tool for those in investment research and I see how the technology would help, say, a sub USD 250mm SM put off hiring that first research analyst.
Agreed. This is hysteria; these jobs aren’t going away.
There’s no tool that can build a better excel model than a well-trained associate, or talk to an analyst as coherently and accurately, or even write something that makes more sense.
There is not any tool, as far as I am aware, that can literally open the excel app, take a list of assumptions, and build a model.
Chat GPT and the versions being developed by data service providers
Not sure I 100% agree. The shift towards passive was largely driven by S&P outperformance in the 2010s/early 2020s bull run. Do people honestly think the s&p will continue a 10%+ annual growth in the next decade? With end of ZIRP and tariffs? P/Es at record highs? Not likely. In my view that was cause some AUM shift back to active management as it actually becomes worth money again when your alternative is a stagnating 5% a year s&p
Much of the growth this also because of like10 stocks
100% agree with your point. YTD has been a great era for active managers to really prove their worth in the industry. A lot of active managers are killing it right now relative to what the market is doing.
I mean there’s a lot of room for small cap and smid, but I can assure you no one is still winning in the large cap space. Strategies will continue to be underweight or have no exposure to NVIDIA, and it’s a tough conversation to have quarter to quarter about the underperformance or sector underweight, regardless of how you want to spin “fundamental work.”
This is spot on. At my firm we already stopped hiring associates, our PMs feel we'll rarely need to hire associates moving forward
My Director of Research thinks the same. We're first going to be cutting the associate class size moving forward to ~50% of what it is today, and he thinks longer-term it will only be a handful of associates (I work at a $20-100bl manager) will be needed after that with minimal internal talent development
VA LO?
So when the current Partners retire 10-20 years from now, there will just be no HFs because juniors were essentially removed?
Realistically the way this plays out isn't 100% capacity today to 0% tomorrow...what an inane way to look at it
It will likely go 100% go 70% to 50% to 10-20% LT (or some rough cadence)
I just got an analyst opportunity, my latest shop (which used to have 1-2 associates) is also dropping hiring of associates given AI tools are doing this work (getting data, writing basic notes, updated models)
What tools exactly?
It's amazing how many people are in denial about this, usually prospects that think AM / HF is their golden ticket or of course current jr associates. I get it, who wants to come to terms with the reality that all this sunk cost was useless? But it's wise to consider alternatives, I'm def thinking of exits
Any exits in particular?
9 months later and the tools are still pretty sub par and don’t stack up against a good well trained associate. They’re great for some things and make some workflows and data wrangling easier but a full on replacement for 100% of associates? Not going to happen. And this is in the context of this stuff being new and shiny and management teams wanting to actively test and use the tools to create efficiencies. When that effect wears off there will be further acknowledgment of the actual capabilities. I’d say 50% getting replaced is extreme and we’ll probably end up with 15-30% fewer seats. Maybe you see a larger reduction but that likely means there are other contributing factors that are not AI tool related.
Sell side ER associate here.
The advent of AI is a fairly big reason for stopping my job search into the normal path SS ER -> LO path (given limited interest in doing L/S). There are other reasons but this is a big one. Why would I invest time into this structurally declining industry with big technology replacement risk?
It just feels like, as an ER analyst, that's an incredibly ironic choice to make.
I don't think we're quite to the point of replacement but I agree with OP (was very surprised to see who it was who was posting this, Sequoia is usually pretty balanced I think?), the speed at which firms are adopting AI (which btw is still neutered compared to non-regulatory-hindered versions) and what this stuff can already do has taken me by surprise. I thought we were like a couple years out from where we are rn (in terms of getting these capabilities at a bank). I thought regulation would really slow things down, and it has a bit but less than I thought. This stuff, in it's current neutered form, is still giving me significant time savings on certain (albeit fairly small) tasks. And it's only maybe 25% as good as base level chat GPT. Like come on.
The job is about to get harder (because the only thing associates are good for is going to be the very difficult stuff) and at the same time, more competitive. And you can bet damn sure the firms aren't passing any cost savings to your bonuses.
Happy to be given another viewpoint.
When I ask for more details on the specifics of these tools, or examples of specific tasks you've automated due to AI, I'm not asking because I'm in denial but instead to flesh out a deeper conversation around this.
My anecdotes from people at LOs is that the AI hype is overblown and a lot of asset managers are just using it as the "excuse" for downsizing they're actually doing due to shift to shift to passive and relative underperformance post-2008.
Said anecdotes also suggest that AI is immensely helpful for helping an individual research analyst think more deeply about a particular subject of research, but not for outright replacing them.
Agreed. Easy to use the excuse of 'AI dominance' when the real reason for the lack of hiring is the root nature of being a LO.
Few things
Thoughtful stuff. To be clear by title, when I say largely disappear I'm not saying entirely -- my best guess is 20-33% of the associate seats remain in 5yrs. I realize this is an expontential decay type bet, but it feels increasingly likely by the day.
The reason I'm offering a conviction viewpoint (vs. a more balanced one) in this case is the most recent data points I've come across. Latest ChatGPT model is incredible for instance and I'm using it for a lot of day to to day stuff. I'm hearing PMs increasingly say (in past 1-2mo) the quality of the research is on par with the typical analyst at the firm. I'm seeing more tools in the marketplace that build / update models and help efficiently find data (vs. having an associate run out and do it).
The pace of technological gain + adoption of AI is much faster than I'd have also expected. I think with twin headwinds of underperformance + shift to passive (and fee pressures tied to this), firms are looking to enhance productivity / cut costs however possible. All to say, do these associate jobs largely disappear in 5yrs or maybe 7yrs? I don't know the exact time / quantum, but I'm seeing more and more data points of firms wanting to move in this direction
67-80% of associate seats disappearing by 2030 is quite the conviction take.
Is there any evidence that points to your ~75% of associates not having a seat in the near future? That sort of assumption has to assume an absolute major shake up in the industry, not just because of A.I but something that re-defines the L/O industry as a whole, (unless your assumption is solely based on AI, which in that case I worry for your well-being, my friend).
Your last paragraph explains why the development of Excel and the internet didn’t stop 100-hour IB work weeks, and only led wages to decline when adjusted for inflation. Technology allows for less hiring and it allows for teams to “do more with less.”
This
In ER, this is a good piece on what has happened. There used to be more analysts, they used to earn more, and they used to cover fewer stocks.
The attachment isnt loading.
If AI can write a note equivalent to your note, you are probably a pretty shitty associate and analyst.....that said probably 60% or 70% of analysts write pretty shitty generic reports.
So I can totally see how AI could eliminate shitty and mediocre analysts putting out basic bullet points. I don't see how it can do a good job of eliminating thoughtful analysts and detailed modeling.
I completely agree. The truth is probably 70% of analysts are subpar or meh. The remaining 30% tend to be exceptional and will be even more critical
Another thought here is I wonder if research departments are thinking about it the wrong way.
Instead of cutting your associate, keep the associate, leverage the AI, and go from covering 15 to 20 names to cover 30 to 40 with with a one analyst and one associate team. If you cut out an analyst, you'll actually save 3x of cutting out the associate. You know which 10 or 15 names really need your focus and you can churn out generic coverage for the rest. Everybody already does this to some extent even pre-AI.
As an LO associate who has used ChatGPT to write research notes and has shifted most of their work to thematic and deep research rather than taking notes and doing model updates, I...:
Also NoEquityResearch I agree most analysts write really sucky reports 60-70% of the time. We also just went through a RIF and lost one analyst and two associates on my team. So I think it's true that not all analysts' jobs are safe while assocites are not. I think it's more about what can be moved to "passive coverage" as we are calling it. In other words, what can ChatGPT cover and what can't it.
Yeah agree on analysts that suck being in trouble. If you are an analyst that stays at the surface level, leans on sell-side and is lazy, adios. Deep and differentiated research is going to matter more than ever. That analyst probably wouldn't hack it anyway but it will be more obvious with AI
Problem is that 'really good research' is rare. At my prior much larger shop, 20-25% of the analysts were exceptional and produced work that the best AI just can't come close to (with actionable and high hit rate recommendations)...this will remain differentiated
The problem is that AI is pretty good at replacing the 'fat middle' of mediocre research. Fact is that this is the bulk of most research. Much of this is generated by associates (the table stakes stuff) and easily outsourced to AI
My current firm has held off on hiring an additional analyst since with AI there's no rush to do so. We may in a few years if AUM grows more but for now none of us feel overly stretched. Agree that many firms will move slower on this, but that just means it's a persistent headwind in the mid-long term even if it has low impact in the next 1-2yrs
I'm joining one of the larger LOs that gets mentioned on this site, and also have a CS background, so this is something that I've been thinking a lot about. I would be curious to hear what others have to say about the following thoughts that I've heard from others or reflected about myself:
Would appreciate people's thoughts :)
Isn't this short-sighted because how will you build a pipeline of future PM's etc. I guess you can just hire directly at the analyst level.
Also there is a lot of "art" in fundamental investing (assessing management, understanding qualitative factors) that still needs to be parsed through. The liberation day sell-off was a great example, everything got tagged but there was a lot of opportunities in the carnage of companies better suited for this environment.
Not quite. Why? Because only 20% of associates ever get the promo to analyst and beyond these days (though ability wise, I think 1/3 are probably good enough). I'm not making the case that 100% of associate jobs go away, but rather 70-80%. Moreover, you're forgetting the pipeline coming from other areas of finance where people start as analysts (i.e. GS IB kid goes to Citadel after 2yrs and spends 2yrs there). So there are other avenues to tap as well
There is an art. Absolutely, which is why the profession will be around. The best PMs and analysts who are skilled at the art of investing will be given tremendously greater leverage via these tools. The rest are gradually culled
Had some more thoughts on another problem of cutting out the associate with AI.
Let's say for the sake of argument that your analyst spends 100% of his time talking with buy side clients, management, and traveling. Awesome, that's basically what you want from a revenue perspective.
So, you eliminate his associate with AI. Is he still spending 100% of his time on client interaction. I would guess not. Sure, maybe AI can eliminate the associate but now your revenue generator is spending 80% of his time talking with clients and 20% of his time dicking around with Chat GPT. Yes, the productivity increase is huge for one person but it's still extra work on the analyst.
As I noted above, I think a better use of AI would be to leverage Chat GPT with an associate to increase the coverage universe and maximize the analysts time talking with clients etc.
I don't know a single analyst spending 100% of his time doing that. I don't know a single analyst that spends even 30+% of his time doing those things in aggregate. That job may exist, but it's not what overwhelming majority of analysts are doing. The premise itself is wrong
That's why I said "for the sake of argument"
Let's say that an analyst spends 50% of their time talking with clients etc. Do you want that time to increase or decrease?
If you've gotten rid of the associate and replaced him with AI, now the analyst is spending 30% or 40% of his time talking to clients as he is spending more time now dicking around with AI.
FYI. I know that most analysts don't spend that much time talking to clients but the few rockstar revenue generators that I've known have been a combination of a strong associate cranking out reports and an analyst just talking to people non-stop.
An interesting POV from Ken Griffin as a counterpoint to the replacement of Analyst+ roles:
~18 minutes in.
I work as an analyst at one of the mega firms. The trajectory of AI tools is just one more reason to echo the advice I give to anyone that reaches out which is to start your career doing IB/PE and then move to LO afterwards if you still want to. The grind is worth it. The general stress of this job is a lot lower knowing I have a wide network of people from IB/PE that now span across banking, corporate, investing and PE backed strategic roles that think highly of me and know I have hard skills to work for them and/or their portcos / companies in their network.
I generally agree that RA jobs will likely decline over time. The bull case is RA being used as a lower cost option to cover the long tail of small caps that don’t have direct analyst coverage at most funds. Small cap is a tough business with lower caps to fund sizes and inability to economically cover all stocks in the index with high priced analysts. Whether RA roles are reduced is likely a function of how much value a given firms puts on this value prop.
I’d add that I don’t believe analyst jobs are going anywhere. If your firm is reducing its analyst count with AI, your firm is on a path towards being obsolete. If you think AI is anywhere close to being able to replicate the work you do as an analyst, you aren’t built for this business
Agree on small cap staying power.
US small and international developed will both be the better areas going forward
EM theoretically is attractive but it's messy & frustrating to cover / you have the China underwieght or overwieght problem / clients don't want to crappy returns of the asset class
Associate jobs will still exist but in 5yrs they will be half or less of what they are today. Which is fine because even today only 20% of associates ever convert into analysts anyway so you'll still have a pipeline with only ~40% of the associate seats in 5yrs that you have today
What should a student be looking at for jobs in this industry?
Given firms will shrink headcounts, but in no way is the industry 'going away'.
When computers came, it upheaved the world of Audit & Tax in Big Accounting. But it didn't get rid of their jobs - it just let the companies work employees harder. Rather than spending 2 weeks on 1 client, you could spend 1 week on 2. Same thing is going to happen with AI.
NotebookLM makes work faster - so what? That just means an analyst can do more work than before. Yes, a fund that caps their investments at maybe 20 at a time might reduce or cap headcount since they have no use for more people. But that just means the companies considered 'too small' can get more coverage from Equity Researchers who have more bandwidth and small headcount funds can cover twice as many stocks, making their funds competitive with big firms.
To add to this 'AI is taking over' mantra, I have a similar human-first view of markets. Let's say market makers and algo traders create a perfect equilibrium price, where no short-term trader can make real money. That just means humans are the only alpha generators we have - a program can scrape the web for news and analyze 10-Ks once they come out, but there's no way they can project and perceive the same human emotions and intonations when interviewing company leadership or tell the mediocre consumer sentiment for a company when they're at the grocery store.
If machines eliminate any arbitrage in the markets, it's even more important for firms to raise analysts into leading associates so they can capitalize on the human element.
lol only if you knew how many markets professionals are being laid off right now. with pay ever decreasing. AI is going to kill off so many roles. Algo’s are absolutely CRUSHING market-makers in the last 2-3yrs alone. many of the best in the business are losing their jobs - research is going to be hit way harder
Hard but true: juniors should start planning their next move early because AI and passive flows are rapidly changing the industry.
if you are smart. you would avoid research like the plague (equity or credit). pay will crater and seats are going to continue to vanish
Research positions are rapidly being replaced by AI; advance quickly or stay out of this dwindling field.
Excellent point: juniors should prepare to leave or advance quickly because AI and passive flows are rapidly shrinking this industry.
Brutally honest: AI and passive flows are killing active equities, and associates need to move up or out fast.
The jig is going to be up soon for newcoming folks. The AI tools are already superior to most research associates
Very wise observation: AI is changing research roles more quickly than most people realise, and your counsel to juniors to change course early is excellent.
Would appreciate someone to humour me with my argument. How could this just now become a thing? Haven’t systematic trading strategies already proven their place? and yeah I get it, we’re talking about different things, GenAI is not the same as systematic, but surely if systematic strategies aren’t even 100% model based, accounting for black-swan events and require human oversight? (I hear Renaissance Technologies is the only systematic fund that’s 100% model based).Also GenAI is as good as the person using it. Yeah I might not know all the ins and outs of what the job entails (ie just learned from the thread that you guys compare earning calls reports discussions) and i see how GenAI helps w that but surely that should free more time to generate more alpha, something beyond ensuring a 5% return levered at 3x.
This is true. My current firm isn't firing any existing associates due to AI, but as they cycle out most of the analysts (incl myself) will not be backfilling them. AI can do their work
if you don’t mind. can you be way more specific on what an analyst vs associate vs an RA is? like years of experience and level of responsibility? without that context i have no idea what you’re talking about. sell side is intern -> analyst -> associate -> vp etc
RA is research associate
I’m at one of the large T1 LOs and my shop has actively added RAs.. sounds like the smaller boutiques are the ones making the cut/deciding to not hire more RAs which makes sense when you have fee compression headwinds over the LT and inability to compete with the large shops
what is an RA (years of experience/responsibility). this is an insanely vague response. and what product? equities? credit?
RA = research associate if you couldn’t tell from the title of the post
Across both equities and credit
I feel like that's something in line with that others have been saying for years about the industry anyway...over time smaller shops will stop growing and reduce / not hire RAs, while bigger shops will. It just seems like AI is what's making that thesis play out...
...but is it really? I usually am skeptical when AI comes into place and suddenly is the justification for why a thesis was right. Not saying it's wrong or that there's thesis creep going on, but sometimes I wonder if this is a case where (i) something else is at play and AI is just "getting the blame" or (ii) AI (or some process efficiency) was already priced in and ChatGPT / AI chatbots just enable that efficiency.
I think it's the second case here...people didn't know about ChatGPT in 2021 and for like a year and a half after it came out there wasn't this same level of concern about workforce reduction...
...but ultimately to me it seems like the smaller players will just continue to not be able to compete and the large players will dominate more...are we entering a world of "all hail BlackRock and Fidelity"?
A close family member of mine is a partner at a large equity research firm and they’d beg to differ. You still need people skills to succeed in equity research.
You said nothing here....that is a non-sequitar to the arguement
Yes, you need people skills to do well in ER. Who said otherwise?
The important shift here is that AI helped an analyst complete the generic ~50% of basic research on a name pretty quickly (which is where the associate spends ~80% of their time). You just don't need to have many associates anymore for that grunt work. The entire point of associates was 1) outsourced grunt work and 2) talent development. If premise #1 no longer holds true, then you only need a small % of associates compared to what you had before to train into future analysts
We know that only 10-30% of associates (depending on what shop, so ~20%) ever become future analysts. Compared to say ~50% of whom are 'good enough' to become an analyst. So you really only need a class of ~20 associates today vs. ~100 before (maybe ~30 assuming that 1/3rd don't meet the bar from a quality perspective). 30 associates going forward vs 100 before is a HUGE difference
Keep in mind continued passive flows + fee presssure are also shrinking the number of seats on top of this math. So the 30 today is probably ~20 if we look out into 2035
Would this thesis hold for fixed Income (HY/LL) and public opportunistic credit roles as well? I'm thinking about getting out of private equity/credit investing.
Think you should stay
This was a great discussion, so I'm reviving it with some news: https://www.cnbc.com/2025/12/22/asset-manager-janus-henderson-gets-bought-by-trian-general-catalyst-for-7point4-billion.html
Ordinarily, this wouldn't be that interesting, as private capital has already been moving into the wealth and asset management space. But what I do find interesting, and is relevant to the discussion in this thread, is that General Catalyst (a VC firm backing a lot of AI start ups) has been hard at work buying traditional businesses.
What do you make of this buyout? What direction do you think Janus takes?
AI replaces average analysts faster than it replaces curiosity, domain intuition, and accountability. The bar is rising but it’s not disappearing.
It's not at all. But it'll eat up most associates first (as 90% of associates are below of the bar of an avg analyst), and then the bottom 50% of analysts (over time, could take 10yrs). If you are in the top 50% of analysts, you will most likely be fine in terms of AI (assuming you are at a shop that can at least grow AUM enough to offset downward fee pressures)
This analysis is simplistic, particularly in the following respects:
Crazy coming back to this post almost a year later. Feels like AI has only gotten 10x as good since.
Labore id totam voluptatem ut magni aut magnam enim. Reprehenderit sint id eaque consectetur blanditiis optio. Non nobis aut velit culpa magnam quis.
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...
Molestias aut aperiam aut ut quam ea et. Non corporis ut tenetur libero. Dolor quis earum eligendi enim consequuntur.
Et doloremque et sit provident. Nulla et error beatae et explicabo. Magni dolores sed dolore impedit nihil laboriosam dolor. Consequatur nulla atque ut laudantium qui magnam quia. Laborum veritatis natus et.
Odio eum ad eum possimus. Ad libero quas ea natus alias et maxime. Eos perferendis sed dolores odit suscipit dicta recusandae. Voluptatem iure neque doloremque molestiae.
Voluptas omnis tempore veniam. Minus voluptatem et delectus eum cum. Neque earum expedita quae esse suscipit officia. Placeat quam et aut magni et officia. Nemo dolore sit et.
Error quia odit qui non unde eaque ut. Inventore quia expedita cumque nesciunt soluta. Doloribus quisquam hic laudantium quam assumenda ratione. Et dolorem a ea. Nobis explicabo ducimus eius pariatur. Impedit voluptatem mollitia laborum modi saepe magnam consequuntur qui.
Dolor sapiente eaque et suscipit iste nesciunt. Et aliquid molestias repellat odit eius ratione rem fugiat. Consequatur maxime labore velit quia facilis quis. Ratione exercitationem non aliquam aut incidunt id enim.