Google partners with Goldman Sachs in automating Investment Banking
Investment Banking has been considered an area in finance that is less vulnerable to automation, compared to others such as Trading. This no longer seems to be the case as Google is seeking to use API's to automate vast amounts of work that are traditionally done by Investment Bankers.
My thoughts: Banking is a practice that will always be built on relationships, especially at the higher level. Automation will probably never replace this even at very advanced stages. However, that isn't all that a banker does. And if the routine grunt work that analysts usually have to do can be done by AI, there will simply be less demand for Investment Bankers as a larger percentage of the overall work done by a banker is instead done by AI.
I think it will be a relatively short period of time (or at least sooner than we think) that the work of two analyst level bankers can now be done by one with the help of automation, and as they get smarter and smarter, maybe 4 or 5 to one. Seniors can then be focused on the networking, relationships, etc.
Automation won't replace Investment Bankers, it'll just make their jobs easier. You can't replace the relationship-driven part of the industry and neither can you replace the decision making abilities of bankers.
Honestly though, as an investor, would you put your capital into a company that just sent you a swarm of confusing automated valuation data or would you give it to the guys who sent you the exact same data, personally flew to your city, explained exactly why the company is a good investment tailored to YOUR needs, and then took you out for drinks/dinner afterward?
This. Might reduce headcount at analyst level at most.
There is already one analyst per deal and putting the same analyst on even more deals would put their brains way beyond capacity. I am hoping hours will reduce even if this means wages stagnate a bit.
Just to play devil's advocate: I would agree with you, but what if said "company that just sent you a swarm of confusing, automated valuation data" was much cheaper because of that automation? I think this changes things somewhat. I don't think you'll ever have confusing and unverified data ever going out to a client. This is IB afterall. But you may have a lot of the upfront work automated, thus reducing the burden on the juniors. Just my 2 cents.
Automation won't replace bankers but artificial intelligence and machine learning can. In the near future, 5-10 years, computers/AI/Tech/Automation can very easily perform many of the tasks that bankers perform. You say would you rather.... Well depending on deal size the price tag difference between your two methods could be staggering. Companies will be quick to adopt automation. What happens with an ER shop using AI/BI/machine learning produce valuation numbers on the fly from various perspectives. Large companies would ne able to freely purchase that data from multiple sources. Every client knows they are paying for the dinner and drinks in the long run.
My only point is, don't think you're irreplaceable because you're not. Yes relationships will matter but the M&A workforce could easily see 20-50% reductions in size due to technological advancements in AI and machine learning. Obviously this thread is about automation but wanted to look past automation. (Which is Stone Age compared to machine learning)
Settle down, risk
Hmmm, interesting thread to revive given what is happening with GPT being integrated into Bloomberg and Microsoft.
I remember seeing CS models from India. The outsourcing/automation have been going on for years. What people fail to realize is that banking (IB/Corp/Commercial/etc) is about relationships. This is why what makes a good analyst doesn't make a good VP and why automation alone will not be the end of banking.
Exactly. I don't think people are aware just how long the outsourced IB model has existed. There are countless firms in India who will do your modeling, write your CIMS, etc. for you at a fraction of the cost.
India - the place where everybody gets their shit from. Just make sure it isn't the Delhi Belly type. Sorry if you were eating something right now. Honestly, it's much easier to outsource if you are in a BB. If you are in an EB preftige firm or a small boutique like Qatalyst, you'll end up doing most of the work yourself.
I agree with you and Sil, but I think there will be a pretty significant reduction in the workforce even in M&A in the next 5-10 years. Relationship management will take a long time to follow but menial tasks will be quick.
http://www.businessinsider.com/goldman-sachs-wants-to-become-the-google…
The one interesting thing I am curious to see is the pace of cultural shift with increased automation. If investment banking becomes increasingly commoditized due to AI/automation, will that hurt the relationship aspect of the business? Might lead to there being less discernible difference between a historically shit bank and a historically great bank. I'm not a banker by any means but I could see this being a possibility.
Advertising used to be a relationship heavy business. It is much, much less so now that advertising placement and targeting is completely automated on the internet. Not sure if this is an analogous comparison but would be curious to hear a banker's thoughts.
Augmentation and automation are two very different things ...
Automation can do a lot and maybe it will trim the headcount a little, but I don't think it will eliminate the need for people completely. A lot of deals happen from relationships and experience so you can't automate that. You can automate excel but someone has to make the PPT just the way the MD likes it! :)
False, sir...false. I've made powerpoint macros before so yeah. Hold on while I finish building out powerpoint Siri. Her arms and legs are made of VBA
Hey, wanna join forces? I did the same thing!
Admittedly, when I first heard about automation in finance, I had a negative response but I have more a positive outlook now. Firstly, analysts (and senior bankers) will be able to focus on providing value such as by focusing on developing relationships, delivering valuable advice in grey areas, etc. instead of getting through "grunt work" which will be taken care of by software. Moreover, I think this will decrease the hours for junior bankers who spend most of their time doing grunt work. This will improve the quality of life for bankers (maybe traders too) and hopefully banks are shifting more away from "traditional" practices like demanding face time and more towards increasing efficiency, productivity, and being innovative. I think technology enhances bankers' ability to do these things. Finance is not going away anytime soon (I hope.... inb4 a communist revolution) and it's inevitable that innovation takes place - if you are afraid innovation, then you probably shouldn't be working in the field.
Analysts don't provide value - they do the bitch work alone. When an analyst moves into an AO role, that's when they start building their connections. Till then, most people on the other side just tend to look at them as the workforce. I work in the other side and have rarely sat in meetings with analysts - although it would be unwise not to connect with them if given the opportunity.
THanks for your rant, I really appreciate the bold for emphasis....now back to your textbook,
How many AI will take errr jerbz posts do we need?
Reads about IB for 2 hours on WSO
All diz shit can be done by AI, how do these finance bros not see the opportunity here!!!!!!!
Most posts here are related to Trading. Very interesting that Google had approached Goldman to target areas such as Investment Banking.
I'm genuinely curious how one could automate investment banking. Company valuation and company and industry analysis is probably the most subjective job function I can conceive of.
By removing the subjectivity element and relying on large datasets for information.
Besides, these days, investment banking often runs on similar lines to a random number generator. It's just about getting the other side to agree to your "carefully-analysed and generated" number.
I mean, maybe that works in some way in investment banking, I'll grant you that. That certainly wouldn't work in venture capital roles, however, where valuing an early stage company is so stupidly difficult that people have defaulted to convertible bonds. But even in investment banking, the "valuation" of an IPO is a balancing act between tangible numbers and salesmanship on the "roadshow" and then rewarding favored clients. Was Snapchat's IPO price, for example, purely numbers driven?
Pretty much what everyone is saying: banking is a relationship driven business that requires expertise drawn from past deal experiences; analysts provide none of the above, only grunt work.
Conclusion: analyst tasks could be largely automated, reducing the need for a 20 man analyst class to perhaps 2? Thank christ i'm gonna be in and out before this happens.
.fr ??? you read the french site? are you french ?
Um, what's the source of your news about goldman + google partnering?
/
Won't replace people, but will definitely reduce headcount in increasing numbers. Christ, is there any part of the financial industry that's growing? For a guy looking into hedge funds too, this is pretty shitty news
I think we have seen instances of this in a variety of other industries, and while jobs that exist today will be replaced to some degree through the use of automation, I also think it is safe to assume that this will create new job opportunities (i.e. more programmers with a finance focus). The other consistent result of automation has been significant innovation. IN the IB world, that may mean more flexible deal structures, easier access to capital, and bankers who are able to spend less time worrying about the model/pitch/CIM and more time focusing on winning new business.
Don't quit your day job. Automation only makes people's jobs easier. I know you're wishing to get a beach chair and a cushy monthly cash payment from the gov every month, but don't hold your breath.
I see this having a much bigger impact on Corporate driven M&A than banking IB head counts.
That picture is a joke. All of them have on collard shirts under the hoodies.
I believe those people actually work for Kensho lol. You can see some of them are wearing t shirts underneath.
take it for what it is - google being smart. if you are going to build out some next level tech (ai, machine learning, general automation), ib processes and work serve as an ideal test for fleshing out the human aspects of some software (logic, prioritizing, consistency, consuming and interpreting raw data mixed with external communication, foreseeing what needs to be done to avoid a weekend of work....etc.).
Google better build out a feature called let's "grease the wheel" because your Managing Director wants a specific EBIT/EV ratio.
So the article mentions bringing about a single node for all the data collected across the firm and I think this presents important challenges to Goldman. While Marty Chavez is on point about the risk solutions POV about banking as a business, discretion and trust are equally important aspects too.
I think that's a dimension that definitely needs to be addressed by the article and thought about when Goldman builds out this system.
Simple solution - learn to code. Computers can't program themselves, at least not yet.
For everyone who thinks that coding is hard and obscure and should be left for computer science majors- my bet is that within 15 years, programming will be a mandated part of the curriculum in most high schools and middle schools in developed countries. Programming is a tool, like math or language skills, and is becoming essential.
For what it's worth, learning advanced Excel is an OK gateway drug to introductory programming. The basic ideas aren't too different to what you already know.
"Learning to code" in school will bring your skill set to the level of usefulness of the English skills of Chinese students--you could possibly get by in a conversation, but your skills would be insufficient to deploy professionally.
You don't have to be a specialist in it. You just have to be able to use it as a tool.
There's a big difference between optimizing algorithms, knowing about trees, for loops and if statements, and having no idea of how code is structured. Many jobs outside of computer science currently require little above the latter, but basic knowledge and awareness will be essential in a few years.
One of my mentors was Blankfein's assistant for two years after doing his IBD stint. In talking to him, he also made the point that especially at Goldman watch how the power structure plays out in periods of transition because it signals where revenue expectations are soon trending. Marty Chavez becoming CFO is the next evolution within the firm since the J Aron contingent ascended in the mid-90s. Don't think for a second they don't expect to update the nature of the client interaction for the 21st century.
Many of you are thinking about this the wrong way.
I'm old enough to have worked for bankers who were around when computers / spreadsheets became prevalent. The view then in the early 1980s was that spreadsheets which at the time were a quantum leap in innovation and productivity (far more so than any automation or machine learning because they cut at the very heart of the job) would replace analysts and associates and eventually obviate the need for bankers as companies could do their own analysis themselves. Guess what - since then, there are probably 20x the number of front office bankers in NY and the pace of deal making has increased accoridngly. Rather than reduce the need for bankers it allowed bankers to become more productive and do more deals. I don't doubt that technology will be transformative in the future, but expect it to increase the pace of activity.
Believe me, if we had machine learning that made my analysts more productive , I wouldn't reduce the analysts. I would be scaling my business and doing more transactions. And the corporate executives and sponsors I work with would be doing the same.
I don't think that's quite the case. In fact quite the opposite should be expected to happen. I had a lengthy conversation with him over drinks about this issue and where he thought finance in general was heading long-term and I think it worth sharing with whomever finds it interesting.
The process of deal flow occurs independently of the process to create any attendant, tangible product. Bankers did more deals because, simply, there was a tsunami of capital from the baby-boom generation. Consumption slimmed and net assets soared right around the time of Windows 97.
Taken to its conclusion, with this generation retiring and Europe in a demographic free fall (avg birthrate far below replacement), from back of the envelope math it follows this deal flow "pie" will only get smaller - a an order of magnitude smaller - in the next 30+ years. So, if you can't grow your revenues any longer in existing business lines (unless at the margin), you point your ammunition at a different target on the income statement.
All I'm saying is regardless of what your view is on these things, internally a client service model driven by technology is what GS believes is the future value proposition of the firm: that you aren't getting just the expertise and experience of each vertical, but the collective intelligence of the entire company. Ultimately, at the end of the day, you'll see that a lot of the value-add in financial services really wasn't value-add at all...
Anyway, I think this is a good parallel - excerpt from NYT on HBS and disruption: "Talking about disruption at HBS and the education field in general, Mr. Ortalo-Magné spins out the possibilities of disruption even further. “How many calculus professors do we need in the world?” he asked. “Maybe it’s nine. My colleague says it’s four. One to teach in English, one in French, one in Chinese, and one in the farm system in case one dies.”
Veniam omnis adipisci et sapiente atque iusto laboriosam. Fugit nemo vitae aut quibusdam. Id quibusdam tenetur quidem. Nesciunt et repudiandae est perspiciatis et. Molestias doloremque debitis reprehenderit sunt consequatur dolores.
Deleniti quia dolorem voluptas reprehenderit quasi accusamus rem. Quod sit ut in quos assumenda eos. Aut inventore rerum sapiente inventore quae voluptatem.
Laboriosam consectetur assumenda atque facere eaque est voluptatibus. Libero sequi delectus rerum aspernatur quidem. Necessitatibus molestias ut velit adipisci qui atque.
Non fugit natus officia ducimus enim. Et cumque sunt vitae. Inventore maiores aut quam. Aut amet inventore quas qui. Aliquam sit nisi ea autem dolores excepturi ut.
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...
Ex dolorem explicabo magni aliquid excepturi rerum voluptates ut. Voluptatem ut in non voluptas. Et ut aut ducimus aut consequuntur. Quae voluptas molestiae quae sed et iste ut temporibus. Architecto ipsa sunt libero hic.
Et accusantium nam repudiandae accusamus saepe adipisci possimus. Libero quibusdam blanditiis ea odit quis.
Asperiores voluptatibus provident ratione distinctio amet accusamus omnis. Animi accusamus et ut ea alias adipisci ab. Maiores sed dolores incidunt incidunt quae ea dolorem. Omnis odio sit est pariatur qui rerum eos error. Laborum ut quia quam molestias corrupti libero.