Rogo AI
Wanted to hear everyone’s thoughts on Rogo - my firm do not use them but I have heard mixed views on performance of the product and use internally.
Also curious about the company itself rather than the product capabilities. I can’t wrap my head around how they have received a 2bn val and the funding they have as well as recruiting from extremely impressive backgrounds.
Model ML is more fun / flexible to use, honestly both are pretty close having tried both. In terms of val its really just following the Harvey / Legora wave, but you are seeing products gaps now vs generic Anthropic / OpenAI offering
On recruiting that's pretty much just people herding towards investors...and investors are just herding towards anything not the big labs (if they don't have allocation...)
Got you, what’s your take on where the legora/harvey/rogo etc. names are going. I just can’t seem to wrap my head around why they are growing and valued at the rate they are.
From my experience, most banks institutional banks are still using copilot as their core for AI internally and don’t necessarily see them moving towards or adding a rogo to their stack for example
Co-pilot internally has to be the dumbest thing, Like when was the last time Microsoft made a good product?
Its still a little hard to say but personally my opinion shifted quite a bit from 6 months ago. Back then it was all rave for full enterprise commitment to Anthropic or OpenAI. But now, especially with the fact that latest frontier models have safeguards / delays / usage cost skyrocketing...feels like people are pulling back a second.
Once that happens...seems like things shifted back to the apps folks (i.e. Harvey now doing post-training and will offer a software stack more like a inference + deployment in addition to consumer seats)...vertical finance guys (rogo / model ml / hebbia) will likely follow. Also obviously a bunch of noise with both China and / or US open source. So still a ton of things in flux. I just laugh at any banks trying to build their own stuff (at one of the BBs and our tech team couldnt even build a stupid buyer log internal software...they are not building good internal AI)
My firm has Copilot Premium (not sure if this varies based on what firm pays for, but models currently accessible are Opus 4.7, Opus 4.8, "Chat GPT 5.6 Deep Think," and "Chat GPT 5.5 Instant") and Rogo.
Regarding Copilot, great to use as a smarter Google, and it recently became capable of automatically referencing data / materials on our network without having to be prompted (which, it couldn't even access the drive prior to the upgrade to premium). As far as having Copilot put together a full pitch deck, credit memo, model, or other deliverable, it's a grand waste of time. It's great if you're doing research or have a question that can be answered by Google but would probably take awhile to find.
Rogo on the other hand is interesting. Right now, my firm has two tiers of Rogo, one of which they are super selective with who they provide a license to. The tier that a wider group has access to (I am in that group) is god awful. Occasionally, if pigs are flying, Hell's frozen over, and you won the lottery, it spits out something brilliant. The rest of the time, you'd be better off asking George Milton to do something for you. The "super" tier of Rogo, aka Felix by Rogo, is actually brilliant from what I've seen. Our model is super complex (probably overengineered) and not easy to work with; one day, I got a late Friday staffing and wasn't feeling it so I had one of my friends with Felix throw it in there. The thing one shotted it - minimal prompting and solid output (including sensitized cases). One thing it sucks at though is formatting, so I guess we all still have a job for at least a while.
Rogo is dogshit. Especially in excel. Decent for research but honestly think copilot has more use cases.
I never really get this model wrapper companies. At my MBB I almost exclusively use Claude / Cursor. Both are very good. I would say like a first year consultant. Recently evaluated the peak sales of a pharmaceutical drug and the model it pulled me together in like 20-30 min back and forth was similar to what we would normally have after 1-2 days manual work.
Very clean excel, detailed assumptions + linked. Very solid.
That being said, I also don‘t like how sycophantic Claude is at times ans this catches me always offguard.
More recently I tried to generate a new workout routine and it habitually changed the routine over and over by the slighest challenge. („You are right!“). That is still very much the nature of an LLM. It provides you an answer with incredible confidence, but will resort to a different answer with the same confidence a few seconds later. Often times a simple „are you sure thats feasible..“ is enough.
Rogo is so overrated. Claude continues get better and hope it just destroys these shit gpt wrappers
Rogo is great for PPT. Extremely underrated in getting a first draft of bullshit your seniors are writing out
This is what I dont get.. why cant you just build a workflow on claude code to do this? Why would anyone want to use rogo...?
I think a lot of banks don’t have Claude enterprise throughout the IB
Rogo, Harvey, Model ML all raised a ton of money and have to prove their business case. Claude and GPT will eat their lunch. Some law firms bought Harvey, just to tell their clients, hey we have the latest and greatest. But do they actually use it.
Hebbia is also so horrible it’s insane they’ve raised so much in VC funding.
It takes so long and you can’t even adjust the intelligence of the model even though it’s LLM-agnostic
Codex is the best, if you leverage it correctly you can save a lot of time
I can literally do at a minimum 60-70% of my job solo with Rogo now, without an analyst, unless i have to put together the entire CIM overnight
Is it as good as Fable? No, but if you're saying it's dogshit and not helpful then you're either a terrible prompter or don't know how to use AI and create skills / build workflows
Dude, it still makes so many errors, you literally tell it to extract specific info of a deal from a link, and it will take the wrong date, wrong EV etc.
Also, it so so fucking obvious when someone has used Rogo on their slides, the language is not up to par. Currently have a few incompetent people (surprisingly both are Associates in my team) who otherwise would've been fired, now essentially living in Rogo and barely surviving. What a curse for us and blessing for the regards.
And yes, aware of workflows and projects - still complete ass. Better than an intern, but not a competent analyst.
As I said, no offense but you are likely a terrible prompter. Re: AI language, a) it has been solved, there are a million skills available that will re-calibrate the language to sound more human and b) I said i do 60-70% of my job, not 100%; it is great for giving me ideas, answering questions, verifying my work etc. if it did 100% of my job I would probably not be needed anymore? You guys can't possibly be complaining about having to refining the work you get from an AI tool a bit?
Dude, you’re literally a post-MBA associate. You are the lowest, dumbest, most retarded rung on the totem pole. You can’t tell what is right and wrong, good and bad. Listen to the 2nd year analyst once in a while. He’s forgotten more than you will ever know and it’s telling that he sees shitty work product in Rogo and yet you can’t even tell.
Rogo sucks. Claude or ChatGPT are the only viable path forward.
I continue to be astonished by the progress of OpenAI/Claude frontier models. We've stumbled upon an amazing technology.
Rogo is my dog's name. Literal dog shit
Sometimes infuriates me how lucky some people are. Founder defo had connections, insane of Lazard to sign as first client. We have used it at my bank for over a year, it was awful. Only just got better now thanks to them being able to afford better Claude models...
Fable blows Rogo out of the water.
Rogo is dogshit
Their latest agent Felix is quite good.
Prior to Felix, the product felt pretty underwhelming. Felix caught them up to the frontier.
I have access to Rogo, GPT, and Anthropic (including Fable) through work - I much prefer to use Rogo’s Felix
We have Felix - Rogo and Copilot integration. I've been liking Rogo for completing mundane, monotonous tasks. In Rogo I'm able to create a "Project," dump a ton of files in the Project tab, then as long as I prompt it with specific enough instructions it's able to reference the correct files and fill out whatever I ask it too. Of course, there will always be a few mistakes, but I've been liking it. It saves me a good amount of time.
Copilot has been pretty helpful in Excel too as long as my prompts are really specific.
The best is to contract with an AI consultancy... These platfroms are super expensive and not tailored made to your process etc. AI is the best when the context, memory is well defined.. Usually they build around Claude so whenever a new model is out it will be updated so you get the best engine
Every person I know who has both Rogo and Claude strong prefers Claude. Ask any Lazard analyst, pretty sure they are going to move off of Rogo. All the PE associates who came from banks with Rogo are absolutely blown away by Claude
What are some complex use cases you use it for?
For AI ball-knowers… can anyone make the case for Rogo being more than just a harness? Took a demo. It looked like basically a worse version of systems I’ve built myself. Claude user with basically unlimited tokens and have build extensively in openclaw. Not impressive that you can email Felix.
Most cynical view is that they only exist so banks can pretend they get more safety than using Claude/Chat enterprise. Wrong?
Amazing a VC fund gave them money to build something you can make in a couple afternoons.
Rogo started pretty great, and there's a lot to like. It can be pretty good at scraping data, research, and putting some daily things together. The more I've used Rogo though the worse it's seemed to get. I think my frustration often comes from the fact that it wildly overcomplicates tasks and asks. It will overexplain and make things super unintuitive. If I have it try a task or give a shot at an analysis, its a lot of coaching, only for it to miss a simple thing, or not present it in a way I want (despite asking explicitly). When it does get it right, it often comes with a few paragraphs and lists of nonsense. The term "concise" and "to the point" mean nothing to Rogo, neither does "make no mistakes".
Not terrible like Copilot, but increasingly underperforming.
Similar sentiment here. Tend to just have it churn data out into a simple format, force it to put checks on everything to verify it did it correctly, then I have excel pull the rest w/ fx, pq, macros, etc.
Dolore eligendi repudiandae odio cumque ipsa eius tempore at. Molestiae voluptatem voluptate magnam minus reiciendis nemo. Dolorem ut nam ducimus animi soluta. Eligendi doloremque porro ipsum id.
Aperiam dolore aliquam minus impedit aliquam delectus. Sit est placeat libero aut. Quaerat iste maxime inventore eum accusantium voluptatem. Nemo et est quis inventore quis voluptatem.
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...
Cum eveniet corporis perspiciatis occaecati voluptas quo aut. Ut laborum perspiciatis sequi. Quis animi ea aperiam aut est laboriosam. Laboriosam nam facere nesciunt aut eius rerum voluptatem.
Ut autem et dolorem nihil labore. Adipisci qui fuga nobis modi in fugit. Recusandae omnis cum aliquid alias ea corrupti adipisci.
Sed alias aut qui necessitatibus nisi vero molestias quasi. Et magnam deserunt sed. Facilis quo vel et. Ea sint et neque itaque et est.
Aperiam velit magni voluptatem voluptatem ex reiciendis ut. Vitae id ut ducimus id cumque ut omnis iste. Deleniti at eveniet maxime autem labore. Saepe sint magnam quaerat laborum quidem. Modi explicabo consequatur in enim nihil. Molestias ut hic et qui eos. Fugiat nulla aspernatur sit dolores id quam ipsam.
Veritatis accusamus molestias temporibus architecto incidunt similique. Porro fugiat nostrum delectus eum facilis suscipit odio non. Sit molestiae ea repudiandae.