Summary
The ATS is the software that reads your resume before any banker does, and at a large bank it rejects more than 90% of applicants. It scores how closely your language overlaps with the job description, so you clear it by mirroring the posting's terms and making every bullet specific enough to match.
Here is something most students don't find out until it has already cost them: at a large investment bank, the first thing standing between you and an interview isn't a banker. It's a piece of software. And it throws out more than 90% of the resumes it reads before a human ever opens the file.
You can have the right school, a real internship, and a genuinely strong story, and still never get read, because you never cleared the machine. So before we talk about writing a single bullet point, you need to understand the gatekeeper. Once you know how it thinks, getting past it stops being a mystery and becomes a checklist.
The 90% Gate: The First Reader of Your Resume Is a Machine
The gatekeeper has a name: the Applicant Tracking System, or ATS. In plain terms, it's AI that "grades" your resume. Its usage is more prevalent at larger firms, the bulge brackets especially, and once you see the volume those firms are buried under, it's obvious why.
A bulge bracket can receive tens of thousands of resumes for a handful of seats. Reviewing tens of thousands of resumes manually is a poor use of firm resources, and it produces inconsistent, less reliable results. Think about what actually happens when a person tries to do it: the banker reading the 4,000th resume late on a Friday is not the same reader who opened the first one fresh on Monday morning. Standards drift. Attention fades. Two equally strong candidates get different verdicts purely because of when they happened to land in the stack.
So the firm hands the first cull to software. The machine reads every resume the same way, applies one standard to all of them, and does it in seconds. That consistency is the whole point. It is also why you cannot charm it, network around it, or hope it catches you on a good day. You clear it on its terms or you don't clear it at all.
How an ATS Actually Scores You
The mechanic is simpler than the mystique around it suggests. The system looks for certain keywords in your resume, usually quite similar to the position's job description, and rejects any resume that does not meet a pre-determined relevance threshold. That threshold will likely reject 90%+ of resumes. The remaining candidates' resumes are then evaluated by humans in creating the first-round interview shortlist.
Read that again, because the strategy falls out of it directly. The job description is the answer key. The machine is measuring how closely your resume's language overlaps with the language of the posting, scoring that overlap, and cutting everyone who lands below the bar. You do not need a flawless, word-for-word match. You need to clear the bar. Every relevant term you add nudges your score upward, and past a certain point you're over the line and into the pile a human actually reads.
Notice what the machine does not do. It doesn't hire anyone. It doesn't weigh your potential or read your story with any sympathy. It builds a shortlist of survivors and hands them to people. Clearing it is necessary, not sufficient. But you cannot win a race you were never entered in.
The Core Move: Mirror the Job Description
Read the posting closely, the way you'd read a model answer before an exam. It will tell you the group, the products, the software, the skills, and the deal types the firm cares about. Those nouns and verbs are precisely the terms the machine is scanning for. Your job is to make sure your resume speaks that same vocabulary everywhere it honestly describes what you did.
The vocabulary comes in two flavors, and you want both. The first is buzzwords, the role-specific terms that signal you know the work. This is what makes buzzwords an essential component of any student's resume rather than a nice-to-have. The second is action verbs. Every bullet point should begin with one, for two reasons at once: it reads as active and finished ("Built a model," not "A model was built"), and the verb itself is a high-value keyword sitting in the highest-value position on the line, the very first word the machine reads.
For the actual vocabulary to pull from, our Action Verbs & Buzzwords List has the full set of action verbs and buzzwords to draw on. Keep it open beside the posting as you write.
Specificity Is Just Keyword Density
So how do you actually get relevant terms onto the page in volume? You make your bullets specific. This is the part students underrate the most, because they think of specificity as a credibility move. It is that. But for the machine, specificity is something more mechanical: it is keyword density.
A vague bullet is nearly empty to an ATS. There is almost nothing in it to match. A specific bullet names industries, geographies, software, deal types, sources, and methods, and every one of those is a potential keyword hit against the posting. Watch what happens when you take a generic line a layer deeper:
Before: Collaborated with team to implement growth initiatives
After: Collaborated with 3-person team to identify global telecom industry incumbents' unique growth initiatives; liaised with strategic finance team to refine ideas for implementation
The "before" gives the machine almost nothing: "team," "growth initiatives," and that's the well run dry. The "after" hands it an industry (telecom), a team size, a named function (strategic finance), and a concrete sequence of work. Same underlying experience, several times the matchable surface area.
It works even on a bullet that already looks decent:
Before: Analyzed competitive landscape using sell-side research and public filings
After: Spearheaded benchmarking analysis of global CDMO players to draft 5-page investment committee memo by consolidating consultant call transcripts, identifying secular trends, and heat-mapping global activity
The "after" names a specific industry (CDMO), a specific deliverable (an investment committee memo), and specific methods (benchmarking, heat-mapping). Those are the exact kinds of terms a banking posting uses, so the line now reads as both more credible to a person and denser to the machine.
When a bullet feels thin and you don't know what to add, mine it with these buckets:
- Mention software used (Bloomberg, Capital IQ, etc.)
- List receiving parties (to whom you submitted/presented your work) and duration of project
- List industry, geography, business model and investment mandate
- Mention number of participants or team size
- List information sources (e.g., sell-side research, public filings, etc.)
Each bucket is a place to dig out a concrete, matchable term. And look closely at what those categories are: software, industry, geography, methods, sources. They are the same categories a job description lists. Reading keywords off a posting and mining your own experience for specifics are the same activity run from two directions, which is exactly why specificity is the engine that drives this whole strategy.
Your Skills Line Is Prime Keyword Real Estate
There is one spot on your resume built for nothing but keywords, and most students waste it. The Skills line, down in your Additional Information section, is the most concentrated place to name the software a posting asks for. Use it for software only, and primarily financial software:
Non-Financial: Microsoft Office Suite, Python, PowerBI, SQL
Financial: Bloomberg, Capital IQ, Refinitiv Eikon, Preqin, Pitchbook, FactSet
If a posting names Capital IQ or FactSet and you've used it, putting it here is a clean, unambiguous match the machine cannot miss. And the system does not care whether a term lives in a bullet point or on the Skills line; it only cares that the term is present somewhere. So if you've already woven these platforms naturally through your experience bullets, a standalone Skills line becomes optional. If you haven't, keep it, because the keyword has to appear somewhere.
Past the Machine, Into the Room
Clearing the ATS will not get you the job. It gets you a reader. The machine builds the shortlist; people decide who gets the first-round interview from there. But hold onto the number: more than 90% of applicants never make it through this gate at a large firm, so simply clearing it already puts you ahead of most of the field.
And here is the reward for doing it right. The specificity that beats the machine is the very same specificity that earns a human's trust. Nothing you did to pass the filter has to be undone for the person who reads you next. Once a banker is actually reading, the work shifts to quantifying the impact of your bullets, writing them tight, and being able to speak to every line convincingly when you're sitting across the desk. Those are their own disciplines, and each is worth mastering. But none of them get the chance to matter until the software has let you through.
So get the keywords right first. Mirror the posting, make every bullet specific, name your tools, and clear the gate. Then go win the room.
Enjoyed this article?
Click on a star to rate it.
Free guide
The Behavioral Interview Handbook.
In-depth frameworks, templates, and real answers for every behavioral question you’ll get asked in interviews — and how to apply them to your own story to stand out and get the offer.
We will never spam or sell your info. Ever.