Candidate Experience Survey: templates + AI-driven improvements

Angle + intro: what candidate experience surveys do and when to send them

Candidate experience surveys are standardized questionnaires sent post-apply, post‑interview, and post‑offer/rejection to quantify how candidates perceived your process and where it broke. According to the Talent Board’s 2024 CandE research, North America’s overall candidate experience NPS (the “CandE Score”) sat at 17 in 2024, with a 14% “resentment rate” among candidates reporting very poor experiences; top withdrawal reason: “the recruiting process took too long.” ¹

Candidate experience directly affects referrals and brand. The CandE program reports that “employers depend on referrals and on average 20%–40% of their hires can come from referrals,” and that roughly 90% of surveyed candidates are rejected—so the experience you deliver to non‑hires still drives word‑of‑mouth and future pipelines. ²

“Candidate-experience auditing helps employers design the optimal candidate journey…and identify and correct the problem areas,” notes SHRM—underscoring why surveys should be embedded as a permanent feedback loop, not a one‑off project. ³

When to send:

• Post‑application (within 24–48 hours) to spot drop‑offs caused by job description clarity, application length, or technical friction.
• Post‑interview (same day or next morning) to catch fairness and preparedness issues while memories are fresh.
• Post‑offer or post‑rejection (within 72 hours) to learn what moved acceptance/decline and what felt fair, even for those you didn’t hire. CandE winners pair “consistent and timely communications” with timely feedback requests across stages. ²

Copy‑paste templates by funnel stage

A candidate experience survey is most useful when short (≤10 questions), mostly 5‑point Likert items, plus an open text and a candidate Net Promoter Score (0–10 likelihood to recommend your hiring process). Greenhouse’s built‑in candidate survey uses nine default questions and caps at 10, which is a good target for completion. ⁴

Scale guidance and targeting

• Use a 5‑point Likert (Strongly disagree → Strongly agree) for speed, clarity, and fairness. Talent Board maps 4‑ and 5‑point ratings to NPS consistently, supporting trend tracking. ²
• Send stage‑specific surveys automatically (e.g., “Applied” or “Interview Complete” triggers) via your ATS; ATS and ATS both support timing/recipient controls. ⁴
• Keep one optional open‑text field per block; AI can auto‑tag themes (next section).

Application experience (10–12 items; 5‑point Likert + cNPS)

• The job description accurately reflected the role’s day‑to‑day work.
• The required qualifications were clear and specific.
• The salary or pay information (if provided) met my expectations for this role/market.
• The application took an appropriate amount of time to complete.
• The application form worked well on my device (no technical issues).
• Communications after I applied were timely and set expectations for next steps.
• The career site made it easy to find the right role.
• The application asked only for necessary information (no duplicate data).
• I felt respected throughout this part of the process.
• Optional: Please share one thing we could improve at the application step. [Open text]
• Candidate NPS: How likely are you to recommend this company’s hiring process to a friend or colleague? (0–10)

Interview experience (12–15 items; include accessibility/DEI)

• My interview(s) were scheduled promptly after initial outreach.
• Interviewers were prepared and familiar with my background.
• The questions were relevant to the job and fairly assessed my skills.
• I received a clear outline of the interview process and timeline.
• I had equal opportunity to demonstrate my ability to do the job.
• The interviewers treated me respectfully.
• I received timely updates between interview rounds.
• The assessment(s)/take‑home work were reasonable in scope for the role.
• The interview format (virtual/in‑person) worked well from a technical standpoint.
• Accessibility: I was offered appropriate accommodations (e.g., captioning, extra time).
• Accessibility: The physical/virtual environment met my accessibility needs.
• I felt the process reduced bias (e.g., structured questions, consistent rubrics).
• Optional comment on interview experience. [Open text]
• Candidate NPS (same 0–10 question as above)

Offer/rejection follow‑up (8–10 items; with one open text)

• The outcome (offer or rejection) was communicated within the expected timeframe.
• The rationale behind the decision was explained respectfully.
• For offers: Compensation and benefits were transparent and competitive.
• For offers: The start date and onboarding plan were clear.
• For rejections: I understand what would make me a stronger candidate in the future.
• I would consider applying here again in the future.
• I would refer others to apply here.
• The overall process felt fair.
• What is one thing we could improve next time? [Open text]
• Candidate NPS (same 0–10 question as above)

Shortcuts and context from the market

• Template marketplaces like Qualtrics, SurveyMonkey, Jotform, and Formstack provide quick starts, but you’ll still need to tailor wording to your hiring stages and job families. ⁵
• Workable’s question bank is a handy source of phrasing for fairness, preparedness, and communications items; cross‑reference when customizing. ⁶

If your application survey flags JD clarity or compensation confusion, fix the upstream inputs before the next req goes live—use our free JD generator and salary calculator to tighten the role pitch and pay banding.

AI‑driven improvements: tagging feedback, finding root causes, and acting fast

AI for surveys means more than sentiment scores; it’s about turn‑by‑turn operations. LinkedIn’s Future of Recruiting research notes only about 27% of talent pros are using or experimenting with gen‑AI, but six in ten are optimistic—suggesting room for advantage if you operationalize AI now. ⁷

What to automate with AI

• Auto‑tag themes in open text. Group comments by communication speed, JD clarity, interviewer prep, fairness, accommodations, and compensation transparency. A weekly tag cloud shows where to focus.
• Flag likely root causes and propose fixes. For example:

- “Process too long” + “multiple duplicative interviews” → Merge interviewer panels and standardize scoring rubrics; CandE research identifies time‑to‑decision as a top driver of negative outcomes. ¹

- “JD unclear” + “skills mismatch” → Rewrite the JD around outcomes and must‑have skills; test with our JD generator.

- “Comp below expectations” → Review bands with your comp partner and validate with our salary calculator.

• Track cNPS by role, geography, and stage; alert when it dips. Set thresholds (e.g., cNPS < 0 for two consecutive weeks in EMEA interview stage triggers a root‑cause review).
• Connect survey data to hiring velocity. In 2023 CandE data, enabling candidates to apply from a text increased application NPS by 42%, and offering a chatbot Q&A increased application NPS by 50%—clues that responsiveness and clarity, not just “kindness,” move the needle. ²

Where Raffi fits upstream of your ATS

Raffi is the world's first AI recruitment agency — our agents screen, interview, and rank candidates in 48 hours, 80% cheaper than traditional agencies, with zero placement fees. Plans start at $199 per job. That AI layer turns insights into action: faster updates to candidates, consistent structured interviews, and a 48‑hour shortlist so fewer candidates stall out. Link your candidate survey triggers in the ATS; Raffi absorbs the “prep and screen” steps, normalizes interviewer prompts, and closes feedback loops quickly so your cNPS stabilizes rather than whipsaws week to week. For teams comparing interview tech, see our HireVue comparison to decide where asynchronous video fits. Per LinkedIn’s research, AI that removes grunt work lets recruiters double down on “high‑touch candidate experiences.” ⁸

“After 13 years of candidate experience research, we know these differentiators deliver a higher rate of return…from referrals to revenue,” the CandE program writes—so wire AI suggestions to concrete recruiter behaviors, not dashboards alone. ¹

Benchmarks, KPIs, and response‑rate tactics

Benchmarks are guardrails, not absolutes, but they help you set targets and prioritize fixes. The 2024 CandE report pegs North America’s CandE Score at 17 with a 14% resentment rate; award winners drive much lower resentment (e.g., 44% lower in North America, 72% lower in EMEA). Use those spreads to set improvement goals by quarter. ¹

Define “good” for your team

• Response rate: For email‑based external surveys, 20%–30% is a common benchmark across CX research; nudge via multi‑channel (email + SMS or in‑app) and keep surveys <7–8 minutes. ⁹
• Candidate NPS ranges (CandE): Amazing 50–100 (typical for hires), Okay 0–50 (a win for rejected candidates), Needs Work below 0. Track cNPS overall and by stage. ²
• Time to decision: CandE research repeatedly shows “the recruiting process took too long” as a top withdrawal reason; aim for 3–5 business days between final interview and decision on priority roles. ¹
• Referral leverage: Expect 20%–40% of hires via referrals in healthy programs; your candidate experience should protect, not erode, this stream. ²

Quarterly review ritual

• Week 1: Publish last quarter’s cNPS by stage and role; highlight top three detractor themes with 1–2 verbatims each.
• Week 2: Owners propose fixes (e.g., reduce interview rounds from 5 to 3; commit to 48‑hour feedback SLAs).
• Week 6: Mid‑quarter pulse survey on the highest‑risk stage only; share progress in your all‑hands.
• Week 12: Close the loop publicly with candidates and hiring managers; enroll one new role family into the survey program.

Candidates with negative experiences are more likely to share publicly than those with positive ones, across all regions—another reason to monitor hot spots and tighten timelines. ¹

Implementation checklist + tools

This is a two‑week rollout for most teams.

• Choose a survey tool and owner (Day 1–2). Pick one platform (ATS‑native or external) and assign an ops owner and an analyst back‑up. For context on defaults and triggers, see Greenhouse’s candidate survey overview. ⁴
• Map triggers by stage (Day 2–3). Configure “Applied,” “Interview Complete,” and “Offer/Reject” sends with guardrails (e.g., don’t survey candidates twice in six months). ⁴
• Load the templates (Day 3–4). Start from the blocks above; pull phrasing ideas from Workable’s library. ⁶
• Pilot and test (Day 4–6). Run 20–50 surveys across two roles; validate that completion time is <5 minutes and open‑text tagging works.
• Launch (Day 7). Turn on stage‑based automation across active reqs.
• Analyze weekly (ongoing). Auto‑tag open text and report cNPS by stage/role/geo every Friday; alert recruiters when cNPS < 0 for two weeks.
• Act on findings (within 14 days). If JD clarity or comp doubts spike, refresh the posting with our JD generator and validate pay with the salary calculator. If interview fairness/preparedness lags, standardize prompts using our interview question packs. If declines cite offer content, tighten terms with our offer letter template and set new‑hire touchpoints via the onboarding checklist.
• Re‑measure and tie to cost (Monthly/Quarterly). Track cNPS trend vs. funnel fallout and time‑to‑hire; use our cost‑per‑hire calculator to show ROI.

“Candidate experience isn’t just softer branding,” LinkedIn’s research argues—AI frees recruiters to invest in the human parts that actually move quality and acceptance. ⁸

How Raffi handles this

Raffi is the world's first AI recruitment agency — our agents screen, interview, and rank candidates in 48 hours, 80% cheaper than traditional agencies, with zero placement fees. Plans start at $199 per job.

Here’s how that turns survey insights into action:

• Voice screening and structured prompts: We standardize the first conversation with candidates, improving perceived fairness and interviewer preparedness across roles and geographies.
• Anti‑cheat scoring and consistency: We protect the integrity of assessments without adding hoops, reducing resentment drivers like unclear or duplicative tasks.
• Faster decisions: Our 48‑hour shortlist compresses “time to decision,” which CandE research ties to lower withdrawals and higher fairness ratings.
• Humans review: AI does the heavy lifting—screening, interviewing, ranking—while human recruiters review edge cases and deliver timely, clear updates.

If you want survey signals to show up as better candidate experiences within days, start a role with Raffi today: https://client.getraffi.ai/raffi/start.