The analyst at a boutique bank spends most of a mandate moving numbers from one place to another; when software does the moving, the analyst's job becomes checking whether the numbers are right and what they mean. Data entry, spreading financials, formatting comparable tables, and logging buyer responses are the first tasks to leave the desk. What replaces them is the work the title has always promised: finding the add-back a buyer will challenge, spotting the customer concentration before diligence does, and knowing which buyer on a list is wrong for this company and why. This is a shift toward interpretation, not an end to analyst review. [1]
What an analyst's hours look like on a traditional mandate
A first-year analyst on a lower-middle-market sell-side process spends the discovery phase keying three years of financial statements into a spreadsheet, then keying the tax returns beside them, then reconciling the two. She builds the normalized EBITDA bridge by hand, formats it, and reformats it when the associate wants a different layout.
In valuation, she spreads comparables: pulling transaction data, standardizing the fields, and building the table the partner will glance at for thirty seconds before forming an opinion based on the two deals he already knew about.
In marketing, she formats the CIM, builds the charts, and checks that the figures in the document match the figures in the model after each of four drafts, which they do not, because someone changed the model on Tuesday.
In outreach, she logs every buyer response in the tracker, updates the status column, and sends the associate a count on Friday.
In offers, she builds the IOI comparison table from the letters as they arrive, retyping terms into columns.
It is possible to do all of this for a year and never once form a view about a business. Many analysts do, and many leave the industry having learned formatting rather than finance.
The tasks that leave first
Four tasks move to software before anything else, and each one is pure transcription.
Spreading financials. When an owner uploads statements to a portal and the platform normalizes them, the analyst's job becomes checking the normalization rather than performing it. She reads the add-back schedule and asks whether the owner's vehicle lease is defensible, rather than typing the vehicle lease into a cell.
Building the comparable table. When the valuation engine runs across a vertical with figures that trace to their source, the comparable set is already structured. The analyst reads it for fit: which of these companies is not really like ours, and why.
Checking document figures against the model. If the CIM generates from the same data that drives the valuation, the numbers tie by construction. The analyst's four-draft reconciliation is gone.
Logging buyer responses. When NDA, IOI, and LOI flow sits in one pipeline, the response is logged where it happens. The Friday count is a screen, not a task.
Removing these four does not remove the analyst. It removes the part of the analyst's day that any careful person with a keyboard could do, and it leaves the part that requires understanding a business.
What the analyst does instead
The returned hours go to analysis, and a managing director should be specific with a first-year analyst about what that word means in practice, because she has not done it before. Research from CFA Institute similarly describes data extraction and presentation moving to GenAI while the analyst evaluates outputs, checks validity, understands limitations, and adds insight. [1]
Analysis means verifying. The analyst is the first reader of every generated figure. She takes the normalized EBITDA and traces three add-backs to their source documents, confirms they are what the owner says they are, and flags the one that is not. She takes the comparable set and strikes the two transactions that distort the range. She takes the generated buyer list and reads the closed-deal history of the top fifty names where that history is documented, marking the ten whose last acquisition suggests they are not buying what this company sells.
Analysis means noticing. Customer concentration above a threshold, a revenue step in a single quarter that needs an explanation, a lease that expires during the likely sale window, margins that moved against the industry. These are the findings a buyer's diligence team will put in a memo, and the analyst who finds them first gives the partner time to address them.
Analysis means explaining. The analyst writes the two-paragraph note that accompanies the valuation output: here is the range, here are the three inputs that drive it, here is what would move it. The partner reads the note, not the model.
Analysis means preparing the owner's questions. When the owner asks why his business is worth a lower multiple than the one his friend got, the analyst should be the person who can answer with the specific comparable and the specific margin difference.
How to train for the new job
The old job trained itself: the analyst learned by building, and the associate checked the build. The new job needs deliberate training, because reviewing a generated output well is a skill nobody arrives with. NIST guidance likewise emphasizes validity, reliability, transparency, documented testing, and appropriate human oversight for AI systems. [2]
Pair every generated artifact with a review question for the first three months. For a valuation output: which two inputs would you challenge, and what document supports your challenge. For a buyer list: which five names should not be on it, and what in their deal history tells you so. For a readiness assessment: which finding will cost the owner the most at closing.
Have the associate grade the answers, not the formatting. An analyst who finds the wrong add-back in her first month is doing the job. An analyst who formats a perfect table and misses it is doing the old job.
Give the analyst a seat in the buyer calls early. On a traditional mandate she heard about buyers through the tracker. On a software-run mandate she has the hours to listen to the associate qualify a buyer, and that is where an analyst learns what documented acquisition history in a database actually means.
What a partner can do this week
Take the analyst's current mandate and list every task she did last week in two columns: moved a number, or formed a view about a number. Count the hours in each. Then ask her which three findings on the business a buyer would raise in diligence. If she cannot answer, the first column is why. If she can, she is ready for the second job and the firm should give it to her before the next mandate starts.
How Acquivo Pro fits
Acquivo Pro takes the transcription out of the analyst's role. Owners upload documents once to a branded portal where requests appear, the valuation engine produces a deterministic result across 21 industry verticals with every figure traced to its source, and teasers and CIMs generate from the same questionnaire and data room so document figures tie to the model. Buyer responses, NDAs, IOIs, and LOIs log in a single pipeline. The analyst's time goes to verifying the add-backs, reading the comparable set, and checking the documented acquisition history behind the buyer list, which is what the title was meant to describe. To see it on a live mandate, request a demo through the Contact page.
Key Points
- The analyst at a boutique bank spends much of a traditional mandate transcribing numbers between financial statements, models, documents, and trackers.
- The first four analyst tasks to move to software are spreading financials, building comparable tables, reconciling document figures to the model, and logging buyer responses.
- Analysis for a first-year analyst means verifying generated figures against source documents, noticing diligence findings early, and explaining what drives a valuation.
- Reviewing a generated output well is a skill that must be trained deliberately with specific review questions on every artifact.
- An analyst who can name the three findings a buyer will raise in diligence is doing analysis; one who formats a perfect table and misses them is doing data entry.
Sources and references
Source retrieval dates record research access, not article publication. Read our editorial and date policy.
Choose your next step
Acquivo is user-operated software. Owners and their professionals remain responsible for decisions, communications, and professional review.
Educational information, not individual legal, tax, or investment advice. Consult qualified professionals for your circumstances. Acquivo provides software, not representation or negotiation on your behalf.