If you are a founder screening 50 or more CVs for one role, the fastest realistic workflow is: (1) write down the three must-haves you actually need, (2) skim each CV for 30 seconds against those must-haves, (3) deep-review the survivors on skills, experience, and trajectory, (4) shortlist five to ten candidates, (5) reject the rest with a proper email today. Manual, the whole cycle takes three to four hours per 50 CVs. With an AI CV screening tool the same process takes about 15 minutes end to end. Both approaches are covered below, honestly costed, with the trade-offs you will actually run into.

Small-business hiring is not just enterprise hiring scaled down. The stakes per role are higher, the tooling is thinner, and your time is the constraint. This is what a workable process looks like when you do not have an HR team behind you.

Why small-business hiring is different

Every hire in a 20-person company is 5% of headcount. In a 200-person company it is 0.5%. Getting the first ten hires wrong is often fatal for a small business. That single fact changes every downstream decision.

You have no HR gatekeeper filtering the pile. You probably do not have an ATS budget or the time to learn a full recruiting suite. You also do not have a dedicated recruiter, so the person doing the screening is usually the same person who will interview, hire, and manage the successful candidate. Your time is the constraint, not applicant volume.

That combination means process discipline matters more, not less. But the process has to fit in the time you actually have.

Step 1: Nail the three must-haves before you open a CV

Write down three things on paper. One line each. Not eight. Three.

Example for a software developer role: "3+ years React, has shipped a live product they can point to, works well async." Anything more granular is a nice-to-have and gets scored separately, not filtered on.

The common mistake at small-business scale is a wishlist of eight "must-haves" that filter out every candidate who could actually do the job. Cut the list to three and treat them as pass/fail. Everything else becomes a scoring factor.

Step 2: The 30-second first pass

Skim, do not read. Aim to cut the pile by 60 to 70% in this step.

Look for the three must-haves in the top third of the CV. Reject on: obvious formatting failures (image-of-text CVs, unreadable layouts), wrong seniority band (senior title but junior responsibilities described, or vice versa), wrong location or right-to-work status.

Timebox it: 30 seconds per CV. Fifty CVs takes 25 minutes. If you find yourself spending two minutes on a CV in this step, you are already doing the deep review; move on and come back.

Step 3: Deep review of the survivors

Typically 15 to 20 CVs make it here. Now spend 5 to 8 minutes each.

Score against a simple rubric: skills (40 points) / experience relevance (30) / trajectory (20) / culture-fit signal (10). Write two-line notes per candidate: highlights and one question to ask at interview. That is recruiter-ready output your future self will thank you for.

Flag red flags without auto-rejecting on them:

Job-hopping (more than three roles in three years without progression)
Skill inflation (senior title, junior responsibilities in the bullets)
Copy-paste generic CV with no tailoring to your role
Chronology that does not add up

Time budget: 6 minutes times 20 CVs is 2 hours. Plus the 25 minutes for step 2, plus rejection emails, this is a full afternoon per 50-CV role.

Step 4: Shortlist five to ten and reject the rest

Sort by score, pick the top 5 for a common role or top 10 for a competitive one. Rejection emails go out the same day, not "when you have time." Silence is the worst answer; it kills your employer brand and small businesses cannot afford that.

Template rejection: three lines, warm, no critique. "Thanks for applying. We are moving forward with candidates whose experience is a closer fit for what we need right now. We appreciate your interest in [company] and wish you well." Send it and move on.

Step 5: The AI shortcut

Everything above can be run in about 15 minutes with an AI CV screening tool. Upload the JD, drop in all 50 CVs, and get a ranked shortlist with reasoning per candidate. Read the top ten, sanity-check the bottom five for false negatives, send rejections. Same defensible process. Different unit-time cost.

Free plans exist. SwiftShortlist's free tier covers three jobs and 50 CVs per job with no credit card, which is enough to test the workflow on one real role.

Where AI wins for a small-business context: consistency (every CV judged the same way, no fatigue drift), speed (2 minutes for the batch instead of 3 hours), and career-changer detection (semantic matching catches "led a squad" as management experience where keyword search misses it).

Where you still do the work: the three must-haves have to be in your JD or the AI cannot score against them. The final interview and hiring decision is yours. And under GDPR Article 22 you have to have a human review every reject decision, even if the AI made the initial recommendation. See the GDPR guide for what that means in practice at small-business scale, and the cost-per-hire calculator to model the saving on your own numbers.

Rough math: manual = 3.5 hrs times 3 hires per year = 10.5 hrs per year. AI-assisted = 15 mins times 3 hires = 45 mins per year. That is 10 recovered hours you can spend on interviewing or shipping the product.

What to avoid

"Free ATS" tools built for enterprise workflows. Zoho Recruit, SmartRecruiters, Wellfound are excellent products, but the setup + learning curve is measured in hours you do not have. If your alternative is 50 emails and a spreadsheet, they will overshoot.
Trusting a keyword-only screener. These filter out your best career-changer candidate. Semantic AI (SwiftShortlist, Klearskill, others) reads context; keyword filters do not.
Skipping the human review on rejects. GDPR Article 22 applies at any scale. Every reject decision needs a human sign-off, even if the AI queued it up.
Ghosting rejected candidates. Small businesses run on word of mouth. A rejected candidate who got a proper email will still say something civil about you; a ghosted one will not.

The takeaway

The three-hour manual process works. The 15-minute AI-assisted version works too. Both hit the same 15 evaluation dimensions and produce a defensible audit trail. Pick the one that matches your time budget and your comfort with tooling.

If you want to try the AI path, SwiftShortlist has a free plan sized exactly for small-business hiring: three jobs, 50 CVs per job, no credit card, no ATS to learn. For the bigger picture on what AI screening is and how it fits into hiring, read the complete guide to AI CV screening.

How to Screen Resumes for a Small Business (Without an HR Team)