The HR function at a $15M company is frequently one person, or no one.
The managing director posts the job, reviews the applications, schedules the interviews, checks the references, writes the offer letter, and then hands the new hire to a team lead who is also running four active projects.
The candidate experience reflects this reality, and so does the new hire’s first 90 days — and how to use AI for recruitment and CV screening covers the earlier stage of the same workflow.
AI does not add an HR team. It does add the capacity to make the hiring and onboarding process more consistent, more thorough, and less dependent on finding the managing director three hours of focused time in a week that already has none — the same dynamic that how to build AI into onboarding for new hires covers for the internal onboarding side of the same process.
The hiring and onboarding workflows sit in the second tier of the operations deployment sequence — typically deployed in months three to four after the core communication and reporting workflows are running. For the full deployment sequence, see the ten operations workflows your company should automate with AI first. The Foundation elements — the role vocabulary, seniority standards, and evaluation criteria — are part of what AI Foundations documents should contain.
The seven AI workflows for hiring and onboarding
Workflow 1: Job description drafting
What it is: the formal job description for a new or open role: responsibilities, qualifications, experience requirements, and company culture context.
Current time: 45 to 90 minutes per job description, drafted from a rough role brief or the prior year’s version.
With AI: the hiring manager inputs the role parameters (title, function, reporting line, key responsibilities in bullet form, must-have and nice-to-have qualifications, compensation range) into the HR Project. AI produces the job description in the company’s standard format and HR communication standards. Hiring manager reviews, edits for role-specific nuance, and approves.
Time recovery: 30 to 70 minutes per open role. For a company hiring 8 to 12 roles per year: 4 to 14 hours recovered.
Foundation required:
- Staff communication standards (the employer voice, the culture signals)
- Role vocabulary guide (the language for each function and level)
- Company culture guide
Workflow 2: Job post personalisation by channel
What it is: tailored versions of the job description for different posting channels: LinkedIn (conversational, culture-forward), Indeed (keyword-optimised, responsibilities-first), industry job boards (sector-specific vocabulary, technical depth), and referral network (shorter, relationship-framed).
Current time: 15 to 25 minutes per channel version, manually adapted from the base job description. For 4 channels: 60 to 100 minutes per role.
With AI: the hiring manager inputs the approved job description into the HR Project and specifies the channel and the adaptation conventions. AI produces the channel-specific version. Hiring manager reviews for accuracy.
Time recovery: 40 to 70 minutes per role. For 8 to 12 roles per year: 5 to 14 hours recovered.
Workflow 3: Candidate screening synthesis
What it is: a structured comparison of shortlisted candidates: a summary of each candidate’s qualifications against the role requirements, a relevance rating, and a recommended interview priority order.
Current time: 60 to 90 minutes reading and comparing 8 to 15 application packages (resumes, cover letters, supplemental materials) for a shortlisted candidate pool.
With AI: the hiring manager pastes the de-identified candidate summaries into the HR Project and specifies the role’s must-have and nice-to-have criteria. AI produces a comparative candidate summary with a relevance rating and recommended interview order. Hiring manager reviews, adjusts the order based on additional context, and selects the interview slate.
Data privacy note: de-identify the candidate materials before pasting. Remove name, current employer (if identifying), contact information, and any protected characteristic information. The comparison should evaluate qualifications, not identity.
Time recovery: 40 to 65 minutes per hiring cycle. For 8 to 12 hiring cycles per year: 5 to 13 hours recovered.
Foundation required:
- Role qualifications vocabulary (the must-have and nice-to-have criteria language from the job description)
- Evaluation criteria weighting guide (the relative importance of each qualification for this role type)
Workflow 4: Interview question and guide drafting
What it is: a structured interview guide for each interviewer: role-specific behavioural questions, technical assessment questions, culture and values questions, and a scoring rubric.
Current time: 30 to 60 minutes per interview guide, drafted from a prior guide or from scratch.
With AI: the hiring manager inputs the role requirements and the specific interview stage (phone screen, technical assessment, culture fit interview) into the HR Project. AI produces the interview guide with questions calibrated to the stage and the role, plus a simple scoring rubric. Hiring manager reviews and adds any role-specific technical questions that require sector knowledge.
Time recovery: 20 to 45 minutes per interview guide. For 3 guides per role, 8 to 12 roles per year: 8 to 18 hours recovered.
Foundation required:
- Company culture guide (the values and behaviours the company looks for)
- Role vocabulary guide (the technical language for each function to calibrate technical questions)
Workflow 5: Offer letter and hiring communication templates
What it is: the offer letter, the candidate declination communication, the offer expiration reminder, and the onboarding initiation communication.
Current time: 20 to 45 minutes per communication, adapted from a prior template or drafted fresh.
With AI: the hiring manager inputs the candidate’s name, role, compensation, start date, and any specific offer terms into the HR Project. AI produces the offer letter in the company’s HR communication standards. Hiring manager reviews and confirms all terms before sending.
Legal note: offer letters should be reviewed by employment counsel at least once before the AI-assisted template is finalised — and the same applies when considering how to address AI job concerns with your team. The template review, not every individual offer, requires legal input.
Time recovery: 15 to 35 minutes per hiring cycle. For 8 to 12 cycles: 2 to 7 hours recovered.
Foundation required:
- Staff communication standards
- Company culture guide (the employer voice in the offer stage)
- Role vocabulary guide (role and level title conventions)
Workflow 6: Onboarding checklist and schedule generation
What it is: the complete onboarding plan for a new hire: the week-one schedule, the 30-day task sequence, the system access list, the required reading and training list, the introduction meeting schedule.
Current time: 45 to 90 minutes per new hire, customised from a prior plan or drafted fresh.
With AI: the HR manager inputs the role, the start date, the relevant department, and the standard onboarding sequence parameters into the HR Project. AI produces the complete onboarding plan in the company’s standard format, with week-one schedule, task sequence, and checklist. HR manager reviews and adds role-specific elements (the specific equipment, the specific system access, the specific team introductions).
Time recovery: 30 to 65 minutes per hire. For 8 to 12 hires per year: 4 to 13 hours recovered.
Foundation required:
- Onboarding format guide (the standard sections and sequence)
- Department vocabulary guide (the system names, team names, and role-specific elements for each department)
Workflow 7: 30-60-90-day check-in communications
What it is: the structured check-in communications sent to new hires at 30, 60, and 90 days: the questions the manager asks, the feedback the company provides, and the development conversation agenda.
Current time: 20 to 40 minutes per check-in communication set, frequently skipped when the hiring manager is busy with other priorities.
With AI: the HR manager inputs the new hire’s name, role, manager, and the specific check-in stage into the HR Project. AI produces the check-in communication and the structured conversation agenda. Manager reviews and personalises for the specific hire’s situation.
The highest return from this workflow is not time recovery — it is consistency. The company that runs structured 30-60-90 check-ins for every hire retains new hires at a higher rate than the one that runs them only when the manager has time. AI makes the structured check-in happen regardless of manager schedule.
Foundation required:
- Staff communication standards
- Company culture guide (the values and expectations the check-in reinforces)
Combined time recovery for a single hiring cycle
| Workflow | Time recovery per occurrence |
|---|---|
| Job description drafting | 30 to 70 min |
| Job post personalisation (4 channels) | 40 to 70 min |
| Candidate screening synthesis | 40 to 65 min |
| Interview guides (3 stages) | 60 to 135 min |
| Offer letter and communications | 15 to 35 min |
| Onboarding checklist and schedule | 30 to 65 min |
| 30-60-90 check-ins (3 sets) | 42 to 90 min |
| Total per hire | 4.5 to 8.8 hrs |
For a company making 10 hires per year: 45 to 88 hours recovered. At $70/hour average HR and management cost: $3,150 to $6,160 per year in recovered hiring and onboarding capacity.
The Foundation build for HR
Three documents, buildable in 90 minutes:
Staff communication standards (45 minutes): the employer voice: how the company communicates with candidates and new hires. Tone (professional and direct vs warm and conversational vs formal), vocabulary to use and avoid (the company’s culture language vs corporate jargon), and the specific phrases the company uses for its culture and values.
Role vocabulary guide (25 minutes): the specific language for each function, level, and expectation at this company. The difference between a “Senior Account Manager” and an “Account Lead” in this company’s structure. The specific technical vocabulary for each function’s roles.
Company culture guide (20 minutes): what the company believes, how it works, and what it looks for in people.
Not a copy of the website’s “About” page: the operational description of what it is actually like to work here.
What the company expects from its people, what it gives in return, and the specific values that determine who succeeds and who does not.
Common questions on AI in hiring and onboarding
”What about diversity and bias concerns with AI in hiring?”
The candidate screening synthesis workflow (Workflow 3) is where bias risk is most concentrated. Three protections:
De-identification before processing: removing name, photo, and demographic signals before AI synthesis reduces the likelihood that AI output reflects irrelevant characteristics.
Human decision authority: AI produces a structured comparison. The hiring manager makes every selection decision. The AI output is a synthesis aid, not a screening decision.
Criteria specificity: the more specific the must-have and nice-to-have criteria in the Foundation (specific years of experience in a specific function, specific certification, specific technical skill), the more the AI synthesis focuses on relevant qualifications rather than inferring fit from background signals.
The workflow as described is a synthesis and comparison aid, not an automated screening decision. It surfaces information for human review. The human decides.
”Can AI help with performance reviews?”
Yes, with the same screen/room distinction. AI drafts the structured elements of a performance review (achievements against goals, skill development areas, next-period objectives in the company’s review format).
The manager provides the human judgment: the overall performance rating, the specific feedback on the team member’s growth trajectory, the development conversation.
The draft-and-review model works for performance reviews as it does for every other HR communication: AI handles the structure and the compilation. The manager provides the judgment and the relationship content.
”What about using AI to screen resumes before the hiring manager sees them?”
The workflow described above processes a shortlisted candidate pool: the 8 to 15 candidates the hiring manager has already decided to consider further.
Using AI to screen the full applicant pool (determining which candidates reach the hiring manager) is a different use with a different risk profile.
The AI system is making a consequential decision about who gets considered, without a human reviewing every excluded candidate.
This use is outside the scope of this article and outside the standard Phos engagement. If the company wants to automate initial applicant screening, consult with employment counsel about the applicable legal framework before deploying.
Want the HR Project built and the seven workflows configured?
Seven AI workflows cover the highest-burden points in the hiring and onboarding cycle, recovering 4.5 to 8.8 hours per hire.
The workflows that produce the highest long-term return are not the ones with the largest time recovery — they are the onboarding and check-in workflows that produce consistent new hire experiences regardless of how busy the hiring manager is. AI makes the structured check-in happen when the manager cannot.
Path one: build the Foundation this week. Write the staff communication standards document (how your company communicates with candidates and new hires: the tone, three or four vocabulary conventions, the phrases you use and avoid). Load it into a shared HR Project. Draft the next job description using it. Compare the time and quality to your current process.
Path two: bring in a partner. Phos AI Labs builds the HR Project and configures all seven workflows, with the Foundation documents written and the first job description produced in the first session. Thirty minutes, no deck. Start here.