Accenture Interview Guide (2026)

1Snapshot, read this first#

The one thing to understand about Accenture: the online assessment is the filter, not the interview.

• It hires freshers by the tens of thousands in a single batch, so automation does the screening.
• A human recruiter barely looks at you until you have cleared the automated gates.
• Clear the test cleanly and the interviews turn very passable, because by then they are only checking that you can think, communicate, and learn on the job.
• Nobody hands you a hard graph problem or asks you to design Twitter.
• The money question behind every round is "can we bill this person to a client in six to eight months." Almost everything maps back to that one sentence.

2Getting the call (the part most guides skip)#

Accenture recruits freshers two ways, and the path decides almost everything about how hard it is to get in front of the test.

Campus placements. The easiest door, because the eligibility bar is handled for you.

• If your college is on Accenture's hiring list, the drive comes to you.
• Keep your record clean: they often set a cutoff near 60% across 10th, 12th, and degree, or no active backlogs, though it varies by batch.
• Register the moment the TPO opens the form; the test link is scheduled on a known date.
• You compete mostly against your own classmates, not the entire country.
• Sit in the pre-placement talk and write down the exact role names and comp bands named there. That is the language you echo back in the HR round.

Off-campus / drives. Mass fresher hiring, usually themed around a graduating batch (for example a "2024 batch" or "2025 batch" drive).

• Watch the careers portal and official social channels for the announcement.
• Registration asks for degree, year of passing, percentage, and a resume.
• Register in the first hour it opens. Slots and test links are time-boxed and often fill within a day or two.
• Late registrations get dropped with no appeal and no reopened window.
• Set a phone alarm for the minute the form goes live, have your resume PDF and marksheet numbers ready to paste, and do not wait until the evening.

Referrals exist but matter less here than at product companies.

• The assessment gates everyone equally, no matter who referred them.
• A referral can get your resume looked at faster for experienced roles.
• As a fresher, your test score is the only thing that moves you to the next stage.
• Do not burn a week chasing a referral. That same week on timed quant and pseudo-code sets lifts your pass odds far more.

Resume rules. One page, keyword-mirrored, outcome-first.

• Keep it to one page. Recruiters and the ATS skim for keywords in seconds.
• Mirror the exact language in the job description, word for word.
• Lead every project line with a result and a number. A line with no outcome reads like a tutorial you copied.

ATS keywords to include naturally: Java or Python, SQL, OOP, Data Structures, Git, REST API, HTML/CSS/JavaScript, and the stack of any real project.

• List your strongest two projects with a one-line "what it does + what tech + what impact" each.
• Add internships, hackathons, and certifications (cloud basics, SQL, any Accenture/Coursera-style completion). They read as proof you can pick up new things fast, which is exactly what a services firm buys in a fresher it plans to retrain.
• One honest line about a project you can defend beats five buzzwords you cannot explain. The technical interviewer will point at the shiniest word on your resume and ask you to walk through it.

3The loop at a glance#

Treat these as typical, not fixed. Formats shift by batch.

• Some drives fold the two coding problems straight into the technical assessment window; others run coding as a separate timed slot on the same day.
• The order of the two interviews can flip, and sometimes technical and HR merge into one call.
• Read the drive email line by line. It tells you exactly which format you are getting, and showing up expecting the wrong structure is a self-inflicted wound.

Here is how the funnel narrows, roughly:

4The online assessment, sections, not surprises#

This is where most candidates are filtered, so train for it like a competitive exam, not like coding practice.

• The test reuses the same section patterns every batch.
• Master the patterns and the exact question types and nothing on the day surprises you.
• That is the whole game, because surprise is what burns your clock.

Cognitive Assessment

• English ability: reading comprehension, error spotting, sentence completion, synonyms/antonyms, para-jumbles.
• Critical reasoning & problem solving: seating arrangements, blood relations, syllogisms, coding-decoding, puzzles, data interpretation.
• Numerical / quant: percentages, ratios, profit and loss, time-speed-distance, time and work, simple/compound interest, averages, probability, permutations.
• Abstract reasoning: pattern and figure series, odd-one-out.

Technical Assessment

• Pseudo-code: short snippets where you predict the output or pick the correct logic. Practise reading loops, arrays, recursion, and operator precedence carefully. This is the section freshers fumble most, almost always because they skim the snippet instead of tracing it value by value.
• Common applications & MS Office: basic spreadsheet/word tool awareness, for example what an absolute cell reference like $A$1 does or what a VLOOKUP-style lookup returns.
• Fundamentals of networking, security, cloud: OSI layers at a high level, IP address basics, what a firewall does, what the cloud is, and the one-sentence difference between IaaS and SaaS (IaaS rents you the raw machines, SaaS gives you the finished app).

How to win the assessment:

1. Time per question is tight. Many sections lock you into one question at a time with no going back, so there is no "I'll return to it later." Decide fast: if a question costs more than 60 to 90 seconds, make your best choice and move on. Banking ten easy marks beats sinking three minutes into one hard one and rushing the easy five that follow.
2. Accuracy over attempts. There is usually no heavy negative marking, but guessing blindly still wastes the time you need elsewhere, and a wrong answer banked is not a mark gained. Reason your way to a best guess, then commit.
3. Pseudo-code is pure pattern recognition. Trace variable values line by line on your rough sheet, writing each variable's value after every single loop pass. Watch for off-by-one errors, pre-increment versus post-increment (++i updates before use, i++ after), and array index bounds. Nearly every wrong answer here comes from one mis-traced loop, so slow down and track numbers on paper, not in your head.
4. Build a quant toolkit. Memorise the standard formulas so on the day you are plugging numbers in, not deriving under pressure. Know the percentage-to-fraction conversions cold (12.5% = 1/8, 16.67% = 1/6, 37.5% = 3/8). They turn a 90-second problem into a 20-second one, and that saved minute is where you clear the hard questions.

How to practise: daily timed sets, format-exact.

• Do 20 quant, 20 reasoning, 15 verbal, and 10 pseudo-code snippets a day.
• Simulate the one-question-no-back format with a strict timer, and refuse to revisit any question once you pass it, so the real test feels familiar instead of brutal.
• Two to three weeks of this usually takes you from "ran out of time with five questions unseen" to "finished with a two-minute buffer."

5The coding round & CS fundamentals#

For a services recruiter like Accenture, "system design" is replaced by a coding round plus core CS fundamentals in the interview.

• Difficulty is easy to lower-medium.
• Aim for clean working code over clever tricks.
• Nobody grades you on the optimal time complexity of a DP table here, and reaching for one is a way to run out of time on a problem you could have solved plainly.

Coding round. Two problems, your choice of language (C, C++, Java, or Python). Expect:

• String handling: reverse, palindrome check, count vowels, anagram check.
• Arrays: largest/smallest, second largest, sort, frequency count, remove duplicates.
• Numbers: prime check, factorial, Fibonacci, Armstrong, GCD/LCM, digit sum.
• Patterns: pyramid and number triangles printed with nested loops.
• Basic recursion and a simple matrix traversal.

Get something green on the board before you optimize.

• Write code that compiles and passes the sample cases first, then worry about edge cases.
• A partial solution that runs and passes three of five test cases beats a half-written clever approach that does not compile.
• The judge scores on passing cases, not on intent or elegance.

Here is the kind of problem that shows up, done cleanly in Python so the interviewer can follow it:

def second_largest(nums):
    first = second = float('-inf')
    for n in nums:
        if n > first:
            second = first
            first = n
        elif first > n > second:
            second = n
    return second if second != float('-inf') else None

# handles duplicates and an all-equal list without a second sort
print(second_largest([10, 5, 10, 8, 3]))   # 8
print(second_largest([7, 7, 7]))           # None

Two things that win points:

• One pass over the list, so O(n) time, no second sort.
• It handles the empty-second case instead of crashing.
• You do not need that polish to pass the automated judge, but saying the edge case out loud ("if every element is equal I return None here") is exactly the signal the interviewer is listening for.

CS fundamentals (the technical interview core). Be ready to speak clearly on:

• OOP: the four pillars (encapsulation, abstraction, inheritance, polymorphism) with a real one-line example each; overloading versus overriding (compile-time, same class, different parameters versus run-time, subclass redefines the parent method); interface versus abstract class and when you reach for each.
• DBMS / SQL: normalization basics up to 3NF; primary versus foreign key; the common joins (inner, left, right); the difference between DELETE / TRUNCATE / DROP (DELETE is row-level and rollbackable, TRUNCATE resets the table fast and is not easily rolled back, DROP removes the table structure entirely); and writing a simple SELECT with WHERE, GROUP BY, and a join on the spot.
• DSA basics: array versus linked list (contiguous memory with random access versus pointer-chained with cheap insertion); stack versus queue (LIFO versus FIFO); what a hash table does and why average lookup is O(1); and time complexity in plain Big-O terms.
• OS: process versus thread (separate memory space versus shared memory within a process); deadlock and its four conditions in one breath (mutual exclusion, hold and wait, no preemption, circular wait); and what paging is at a high level.
• Networking: TCP versus UDP (reliable and ordered with a handshake versus fast and fire-and-forget); what happens when you type a URL and press enter (DNS lookup, TCP handshake, HTTP request, server response, browser render); and HTTP versus HTTPS (HTTPS is HTTP over TLS, so the traffic is encrypted).

How to win: explain as you go.

• Say what you are doing, state the time complexity, and name one edge case out loud, for example "what if the array is empty, I return None here."
• Interviewers here grade clarity and correct basics, not elegance.
• A calm correct walk through a simple answer beats a mumbled attempt at a hard one, every single time.
• If you do not know something, say "I am not sure, but my best guess is" and reason from first principles. That reads as honest and coachable, which is the entire hiring thesis for a fresher the firm plans to train for months.

A calm correct walk through the easy answer beats a mumbled stab at the hard one, every single time.

6Behavioral & culture fit#

Accenture frames its culture around its core values: Client Value Creation, One Global Network, Respect for the Individual, Best People, Integrity, and Stewardship, all under the company promise to "let there be change."

• You do not need to recite these back word for word.
• Your answers should sound like someone client-focused, team-oriented, honest, and ready to keep learning.
• Accenture is a people-and-delivery business, so flexibility (shifts, locations, new tech you have never touched) and stability (you stay and grow, not jump in a year) matter a lot in the HR round.

Prepare four or five STAR stories (Situation, Task, Action, Result) from projects, internships, group work, or fests.

• Common prompts: a time you worked in a team, handled a conflict, met a tight deadline, learned something new fast, or failed and recovered.
• Reuse the same three or four real experiences across prompts and emphasise a different angle each time, so you are never inventing a story cold under pressure.

Worked STAR example, "Tell me about a tight deadline."

• Situation: In my final-year project, our 4-person team was building an inventory web app, and two weeks before the demo our database design broke under test data.
• Task: As the one who knew SQL best, I had to redesign the schema and get the app stable without slipping the submission date.
• Action: I normalized the tables, added proper indexes and foreign keys, rewrote the three slowest queries, and split the remaining UI work across two teammates with a shared checklist so we ran in parallel.
• Result: We shipped two days early, the app handled about 5,000 test records with page loads under one second, and the project scored the highest grade in our section.

Keep stories tight and land them on a number.

• 60 to 90 seconds each, always ending on a number or a clear outcome.
• Never badmouth a teammate.
• If a prompt is about failure, own the mistake plainly and finish on what you changed afterward. The result they want to hear is that you learned and adjusted, not that you were secretly blameless.

7Your prep plan#

4-week plan

• Week 1: Lock quant formulas and reasoning types; one timed section daily.
• Week 2: Verbal + pseudo-code daily; start revising OOP and DBMS.
• Week 3: Full timed mock assessments (cognitive + technical) three times; one coding problem daily.
• Week 4: Interview mode, write five STAR stories, mock technical + HR, polish spoken English, read your own resume aloud so every line is a sentence you can defend without hesitating.

2-week plan

• Days 1-4: Quant + reasoning + verbal timed sets, two hours a day.
• Days 5-7: Pseudo-code + one coding problem daily; revise OOP, DBMS, SQL.
• Days 8-10: Two full mock assessments; fix your two weakest sections.
• Days 11-14: Technical + HR mocks, five STAR stories written, communication practice out loud and recorded.

2-day crash plan

• Day 1 AM: Quant formulas + one timed quant/reasoning set. PM: 20 pseudo-code snippets + revise OOP and the four pillars.
• Day 2 AM: one full mock section + two coding problems (one string, one array). PM: SQL joins, DELETE/TRUNCATE/DROP, three STAR stories, say your intro and your "why Accenture" out loud five times until it stops sounding rehearsed and starts sounding like you.

8Why people get rejected at Accenture#

1. Ran out of time in the assessment: got stuck on hard questions instead of banking the easy marks first. The single most common fail in the whole loop, and it is pure practice, not intelligence. People who finish with a buffer are not smarter, they trained the one-question-no-back format.
2. Weak spoken English: failed or scored low on the communication assessment, which is a hard gate for client-facing roles. People underestimate this one because it is not a "technical" skill, and it quietly ends more candidacies than bad code does.
3. Blank on fundamentals: could not explain the four OOP pillars, a basic join, or process versus thread. The easiest marks in the entire loop and the most painful to lose, because losing them signals you did not revise at all.
4. No real project to discuss: only coursework, nothing they could speak about with ownership and detail. If every answer is "we did it as a group and I am not sure who wrote what," the interviewer cannot credit you with anything concrete.
5. Red flags in HR: unwilling to relocate, work in shifts, or clearly likely to leave quickly, which worries a company hiring and training at scale and planning to put you on a specific client project.
6. Pseudo-code traced wrong: misread loops and increments and lost a section they could have passed with careful line-by-line tracing on the rough sheet.

9The offer, comp & negotiation#

For freshers, the Associate Software Engineer package is typically around Rs 4.5 - 6.5 LPA.

• Usually structured as a fixed base plus a small performance/variable component, with occasional one-time joining or retention elements.
• Higher brackets (sometimes labelled Advanced ASE or specialized tracks) go above this range.
• Those are usually tied to how you scored in the assessment or to specific in-demand skills you showed.

Fresher offers are largely standardized, so expect limited room on the base number. The band is decided far more by test performance than by anything you say after the offer letter lands. Still, three things worth doing:

1. The test score is your leverage, and it is set before you ever talk comp. The way to land a higher band is to score well on the assessment and the coding/technical interview, because that score maps you to a premium track. Post-offer haggling will not move a standardized fresher band, so put the effort upstream.
2. Ask about role, project, and location, not just salary. Clarity on the tech stack you will touch and the growth path is worth more in your first job than a small base bump, because the skills you build in year one decide your market value in year three.
3. Be professional if you have a competing offer. Mention it calmly and factually, with the actual number and the company name, and never bluff, because a bluff that gets called ends the conversation on the spot. For standardized fresher bands a competing offer rarely changes the figure, but it can sometimes influence your role or location placement.

10FAQ#

Q: Is there negative marking in the Accenture assessment? A: Typically the core sections are not heavily penalized for wrong answers, but formats change by batch. Treat accuracy and time management as what matters, and do not leave easy questions unattempted, because an unattempted easy question is a guaranteed zero while a reasoned guess at least has a shot.

Q: Which language should I use in the coding round? A: The one you are fastest and most accurate in, usually Java, Python, C, or C++. Fluency beats picking a "fancier" language. Python saves you boilerplate for string and array problems, but use whatever you can write bug-free under a running timer.

Q: How hard is the coding compared to product companies? A: Much easier. It is mostly strings, arrays, numbers, and patterns, the kind of problems you would see in a first-year lab, not a Leetcode-medium. Clean, working, well-explained code is enough to pass.

Q: Do I really need good spoken English? A: Yes. The communication assessment is a genuine filter because many roles put you in front of a client or an onshore team within months. Practise speaking clearly at a steady, unhurried pace out loud before the test, and record yourself once so you hear exactly where you rush or mumble.

Q: What if I have a backlog or lower percentage? A: Eligibility varies by drive. Many set a cutoff near 60% and prefer no active backlogs, but the criteria shift between batches. Read the specific drive's rules carefully and apply wherever you are eligible, because the rule that blocks you from one drive may simply not exist in the next one.

Want this turned into a personalized, dated prep schedule that fits your exam date and your weak sections? Build your own plan and walk in ready.