How AI can reduce administrative work for HR teams

About this session: Work smarter with AI

This blog is based on a live Pazcare webinar, "Work Smarter with AI," hosted in partnership with Tarun, founder and CEO of Wanderfly, an employee wellbeing company, and a certified AI practitioner from Stanford University. Tarun has built AI systems for his own operations, taught AI to students and organizations, and brings a rare combination of hands-on practitioner experience and structured pedagogy to the subject.

As Tarun opened the session: "If you are not using AI on a regular basis, we are wasting somewhere around four to five hours on a weekly basis. At the same time, our brain has a capacity of focus for like four to five hours in a day, and in that four to five hours if you are able to use AI, we can become 10x productive." That is the premise of everything that follows.

Want to watch the full session? Catch the complete "Work Smarter with AI" webinar on YouTube and see every tool, demo, and framework covered live by Tarun. 

What is AI in HR?

Understanding artificial intelligence in HR

AI in HR refers to the application of artificial intelligence technologies, primarily large language models (LLMs) and automation tools, to streamline, accelerate, and improve the quality of human resources processes. It is not about replacing HR professionals. It is about removing the administrative friction that prevents HR teams from doing the strategic, human-centered work that only they can do.

As Tarun framed it in the webinar: "You don't build cars, you drive them. You don't build Excel, you use formulas. At the same time, you don't build AI, you prompt it."

This distinction is important. The majority of HR teams adopting AI are not building models or writing code. They are using already-built LLMs through simple conversational interfaces to accomplish tasks that used to take hours.

Automation vs. AI

Automation and AI are related but distinct. Automation executes a fixed, predefined sequence of steps without variation, a scheduled email, a rule-based data export, a recurring report. AI, particularly in the form of LLMs, understands context, follows nuanced instructions, generates original content, and adapts its output based on the specifics of each request. When an HR professional asks an LLM to write an onboarding document personalized to a specific employee's role and department, that is AI doing something automation cannot.

The real power in modern HR workflows comes from combining both: using AI to generate and draft, and automation tools to route, store, and act on the outputs without manual intervention.

How LLMs support HR teams

Large language models, ChatGPT, Claude, Gemini, Perplexity, and others, are trained on billions of words and documents and can understand context, follow complex instructions, and produce human-quality text on demand. In Tarun's words: "LLMs are trained on billions of words, billions of context, that understands the context, follows the instructions and generates the human quality text."

For HR teams, this translates directly into faster first drafts, better-quality policy documents, more consistent employee communications, and the ability to process and summarize large volumes of information in seconds.

Why HR is one of the fastest-growing functions adopting AI

India's AI adoption is accelerating rapidly. According to the NASSCOM AI Adoption Index 2024, 87% of Indian companies are in active AI adoption stages, and the Indian AI market is projected to grow at 25 to 35% CAGR through 2027. HR functions, which are responsible for some of the most document-intensive and process-heavy work in any organization, are among the most immediate beneficiaries of this shift.

Why HR teams are struggling with administrative overload

HR teams are expected to manage more people, more compliance requirements, more documentation, and more employee touchpoints than ever before. The formal organized workforce in India has grown significantly, according to the Ministry of MSME, 20.39 crore people were employed across registered MSME enterprises as of July 2024, and every hire in that workforce generates onboarding paperwork, policy communication, benefits enrollment, and performance tracking requirements for an HR team somewhere.

Common administrative tasks that consume HR time

The day-to-day administrative burden on HR teams is significant and relentless. The most time-consuming categories include:

• Drafting and updating HR policies every time regulations change or the organization evolves. 
• Creating offer letters, onboarding documents, and employee handbooks personalized to each new joiner's role. 
• Managing employee communication across departments, hierarchies, and geographies. 
• Coordinating recruitment: screening CVs, scheduling interviews, sending status updates. 
• Preparing compliance documentation under India's evolving labor law framework. 
• Administering performance reviews, setting documentation standards, and managing feedback cycles. 
• Generating HR reports and analytics for leadership and board presentations.

None of these tasks require strategic HR judgment. All of them require significant time and attention that could be directed elsewhere.

The cost of manual processes

The cost of manual administrative work in HR is not just measured in hours. It is measured in the quality of strategic work that does not happen because the administrative queue never empties. An HR leader who spends four hours drafting a compliance document is an HR leader who does not have four hours to design a retention program, build a learning and development framework, or analyze attrition trends. Tarun was direct about this in the webinar: "If you are not using AI on a regular basis, we are wasting somewhere around four to five hours on a weekly basis."

The role of AI in HR

AI in HR functions best as a productivity partner, not a replacement. It handles the volume work, drafting, summarizing, formatting, researching, and structuring, so that HR professionals can focus on the judgment work: interpreting context, building relationships, making decisions, and designing programs.

Tarun made this point clearly: "It can't replace your judgment. If you are going to ask any LLM out there, it will always say something what you tell unless and until you would have given a specific context. It is not able to replace relationships, it is very, very important."

Key AI capabilities relevant to HR include: writing and drafting original documents from prompts; summarizing long reports, legal documents, and policy updates; researching regulatory changes and benchmarking data; analyzing data and generating chart-ready summaries; translating communication across languages; planning and structuring complex workflows; and explaining technical or legal content in plain language.

What AI cannot do: guarantee accuracy without human review, access private data without explicit sharing, replace the HR professional's judgment in sensitive or interpersonal situations, or take responsibility for decisions.

8 ways AI can reduce administrative work for HR teams

1. Automating HR policy creation

Writing and updating HR policies is one of the most time-consuming routine tasks in HR. Every regulatory update, every new joining, every organizational change potentially triggers a policy revision. AI handles this work with speed and consistency that manual drafting cannot match.

In the webinar, Tarun demonstrated this live using Claude. He created a project in Claude, uploaded company documents, and set instructions for the model: "Make sure whenever I am requesting for an HR policy to be given for a new employee, I will just provide you their name and their designation. You need to dynamically change the HR onboarding document and provide me the output in PDF format."

The result: a fully formatted onboarding policy and employee handbook, personalized to the specific employee, generated in minutes. As Tarun showed the output live: "Welcome on board, dear Tarun, it understands what are the compensations, what is the equipment and asset allocations, all the things. Just give and go ahead. Otherwise I think everybody can relate how much time these things are going to take on a regular basis if not with AI."

How to apply it: Create a Claude project. Upload your existing HR policy documents, offer letter templates, and company guidelines as context. Set instructions specifying how documents should be personalized. Then simply provide a new joiner's name and designation and let the model generate the document.

2. Simplifying recruitment processes

Recruitment is one of the highest-volume, most repetitive administrative functions in HR. Screening CVs, writing job descriptions, drafting outreach messages, preparing interview question sets, and communicating with candidates at each stage all consume significant HR bandwidth.

AI tools allow HR teams to generate first drafts of job descriptions, screening criteria, candidate communication templates, and interview frameworks in seconds. With the right context and instructions, an LLM can produce a JD for a senior role that reflects the company's culture, compensation range, and specific skill requirements in a fraction of the time it would take to write from scratch.

For teams using Gemini, native Google Workspace integration makes this particularly seamless: as Tarun noted, "since it's a Google product it can integrate with Gmail, Google Docs, Drive natively, making the complete funnel very very easy."

How to apply it: Use Claude or ChatGPT with the CRAFT framework (covered below) to generate JDs, candidate communication templates, and screening rubrics. Use Gemini for teams deeply embedded in Google Workspace to streamline the full recruitment communication flow.

3. Improving employee onboarding

Onboarding is where first impressions are made and where administrative complexity peaks. New joiners need offer letters, welcome communications, policy handbooks, IT setup guides, benefits enrollment information, and introductions to team processes, all personalized, all accurate, all on time.

AI transforms onboarding document creation from a manual, error-prone process into a fast, consistent, scalable one. With a Claude project set up with company context and policy documents, HR teams can generate a complete, personalized onboarding package by providing only the new joiner's name, designation, and department.

Tarun demonstrated this live in the webinar and the output included: "Welcome on board, dear Tarun, on behalf of the entire team we are absolutely delighted. It has generated employee ID, it has generated date of joining, everything." The full document was ready to download and share immediately.

How to apply it: Build a Claude project containing all onboarding document templates, company policies, and cultural context. Set custom instructions for how to personalize by role. Test with a few recent joiners to calibrate the output quality, then integrate into your standard onboarding workflow.

4. Managing employee communication

HR communication, announcements, policy updates, benefit reminders, performance cycle communications, and sensitive individual messages, requires both volume management and quality control. The volume is relentless. The quality standard is high.

AI accelerates the drafting of all routine employee communication without compromising tone or accuracy. Tarun specifically highlighted the value for day-to-day communication: "If you want to reply to an email, you don't need to write these days. Gemini is there, you just have to prompt it to write the email."

For more complex internal communications, leadership messages, performance review guidance, organizational change announcements, Claude's strength in long-form writing makes it the better tool. Tarun was consistent on this: "Claude is really good at understanding longer documents and it's really really good at writing."

How to apply it: Use Gemini for Gmail-integrated replies and routine communications. Use Claude for longer, more nuanced communications where tone, structure, and accuracy carry more weight.

5. Summarizing long documents and reports

HR teams regularly receive and process significant volumes of text: government labor law updates, IRDAI circulars, compliance guidelines, audit reports, board-level people analytics summaries, and vendor evaluation documents. Reading, parsing, and synthesizing these documents manually consumes hours that could be spent on higher-value work.

AI, particularly Claude which Tarun consistently positioned as the best for deep document reading, can summarize, extract key action items, flag relevant clauses, and translate complex regulatory language into plain English in seconds. "Claude is really really good at deep reading," Tarun confirmed in the webinar.

How to apply it: Drop the full document into Claude and prompt specifically: summarize this in bullet points, flag every clause that affects our current HR policy, or extract all dates and deadlines and present them in a table. The specificity of the instruction determines the quality of the output.

6. Automating reporting and HR analytics

Finance and leadership teams increasingly expect HR to present data-backed insights, headcount trends, attrition analysis, compensation benchmarking, benefit utilization rates, in visual formats that are easy to interpret. Building these reports manually from spreadsheets is time-intensive and prone to error.

Tarun highlighted ChatGPT's strength here, noting that it "can analyze the spreadsheets and generate charts. Every finance person and HR person writing bigger reports, by the end of it, investors or board meetings, they all want to see different charts, different diagrams to make it much easier for them to understand."

For teams without coding experience, CodeX (the code interpreter within ChatGPT) is a particularly powerful tool: "If you don't know anything, just if you have an Excel, take that Excel, drop it in CodeX, tell it to create like whatever you want, it does amazing work for you."

How to apply it: Export your HR data to Excel or Google Sheets. Upload to ChatGPT's code interpreter and prompt for specific visualizations: attrition by department, headcount growth by quarter, or claims utilization by age group. The model generates charts and analysis without any coding required.

7. Supporting compliance and documentation

India's labor law landscape is active and evolving. The Code on Wages, the Industrial Relations Code, the POSH Act, the DPDP Act 2023, and ESIC and PF regulations create an ongoing compliance obligation that HR teams must track, interpret, and operationalize. Missing an update or drafting non-compliant documentation carries legal risk.

Tarun specifically addressed this in the webinar: "The Indian government is changing a lot of laws. How can we be in line with it? The best thing to use is Claude." He also recommended Perplexity for staying current with regulatory updates: "Perplexity can become one of the best for them in terms of understanding what are the regular updates and benchmarking."

How to apply it: Use Perplexity's deep search to track new labor law notifications and regulatory updates on a regular cadence. When a relevant update is identified, bring it into Claude with your current policy documents and prompt: given this update, which parts of our existing policies need revision and how?

8. Building HR workflows with AI automation

The highest-leverage application of AI in HR is not using individual tools for individual tasks, it is connecting those tools into automated workflows that run without human intervention at each step.

Tarun demonstrated this with Gumloop and Make: "Gumloop does, considering there's an email like what you have received and that email falls under customer success. You have already labeled it customer success, then followed by an Excel sheet, a person updates information, and then from there it has to update in the Excel sheet. Everything can be done end to end with Gumloop."

He described his own agent built using Telegram, Claude API, Google Sheets, and Gamma: "I will be updating regularly, that should be updating to my Google Sheets. And at the same time the AI, the Claude API in between, should be able to assess it, understand it. Once the presentation is done, it'll send me a message telling the presentation deck is done, you can go and download." The entire workflow from input to finished presentation happened automatically.

For HR teams, this model applies directly: a new hire notification in Slack triggers onboarding document generation in Claude, which routes to a Gamma presentation for the joining kit, which sends via email to the new joiner, all without a single manual step.

How to apply it: Start with Gumloop or Make and one specific, high-frequency HR workflow. Map the current manual steps, identify which ones can be automated, and use the tool's drag-and-drop interface to connect them. Tarun's advice: "once you connect it, it automatically picks out the complete information and it forms an agent and that agent will stay with you forever."

Benefits of AI in HR

The aggregate benefit of AI in HR is straightforward: HR professionals get their time back. But the specific benefits compound across the organization.

• Speed: Tasks that took two to four hours, policy drafts, onboarding documents, compliance summaries, take minutes. As Tarun demonstrated live, a full onboarding policy for a named employee was generated in under three minutes.
• Consistency: AI-generated documents follow the same structure, tone, and policy framework every time, eliminating the variation that comes from manual drafting by different team members.
• Accuracy: When AI is given specific context and well-structured prompts, its output is precise and can be reviewed and approved faster than a document written from scratch.
• Scale: A team of two HR professionals managing 200 employees can produce the documentation and communication quality previously requiring a team of five, without compromising on standards.
• Strategic capacity: Every hour reclaimed from administrative work is an hour available for culture-building, talent strategy, retention design, and the human judgment work that AI cannot replicate.

Best AI tools for human resources in 2026

As Tarun summarized: "Every LLM is good at a few things. No LLM is best at everything." Knowing which tool to use for which HR task is the practical skill that turns AI from a novelty into a productivity infrastructure.

Best practices for using AI in HR: the CRAFT framework

The quality of AI output is directly determined by the quality of the prompt. Generic inputs produce generic outputs. Structured, context-rich prompts produce outputs that are actually usable. The CRAFT framework, which Tarun taught in the webinar, is the most practical prompt-building structure available for HR teams.

• C, Context. Who are you, what does your organization do, and what is the situation you are addressing? The more specific the context, the more relevant the output. Example: "I am the HR head at a 150-person B2B technology company in Bengaluru. We are onboarding a new head of marketing starting on the first of next month."
• R, Role. Assign the AI a specific role that matches the expertise you need. Tarun's example: "You are my marketing professional having an experience of 15 years working in B2B software industry." For HR use cases: "You are a senior HR professional with 12 years of experience in Indian labor law and employee onboarding for technology companies."
• A, Action. State precisely what you want the AI to produce. "Draft a complete onboarding document including welcome message, equipment checklist, 30-day plan, and key contact list for this role." Vague actions produce vague results.
• F, Format. Specify how you want the output structured: PDF-ready document, numbered list, table, one-page summary, slide-by-slide outline, or paragraph form. This prevents the AI from choosing a format that does not fit your use case.
• T, Tone. Specify the voice: warm and welcoming, formal and precise, legally cautious, conversational. HR communication spans a wide tonal range and the default AI tone is not always right for every situation.

Tarun was emphatic on iteration: "During the first attempt you're not going to get the best result. The second attempt you're going to get a better result. Third attempt you're going to get a good result. Fourth attempt you're going to get the best result. That's called iteration."

Challenges and limitations of AI in HR

Using AI in HR effectively requires honest awareness of where its limitations create risk.

• AI cannot guarantee accuracy. As Tarun stated clearly: "It can't guarantee the accuracy because for every LLM they have mentioned at the bottom mentioning that AI can be wrong sometimes. So please check it." Every AI-generated HR document, particularly compliance-related ones, must be reviewed by a qualified professional before it is issued or relied upon.
• Do not share private data. This is non-negotiable. Employee personal data, compensation details, health information, and performance records must never be pasted into a public LLM interface. Tarun explicitly cautioned the audience: "I would request nobody to give any kind of data, personal, company level." Use anonymized or aggregated data for AI tasks wherever possible. Use private deployment options (such as Claude for Enterprise or organizational Google Workspace integrations) for tasks requiring sensitive data.
• AI cannot replace relationships or judgment. HR is fundamentally a people function. AI can draft the communication, but it cannot assess whether an employee is actually okay. It can summarize the performance review framework, but it cannot deliver the feedback. Tarun: "It can't replace relationships, very, very important. This is not something like what to replace relationships."
• First outputs require iteration. HR teams expecting perfect documents on the first prompt will be disappointed and conclude AI does not work. The correct expectation is that the first output is a strong draft requiring review and refinement, which is still significantly faster than starting from a blank page.
• Accepting the first output is a mistake. Tarun was direct: "Please don't accept the first output." Iterate with additional instructions, add missing context, and push for specificity.

How Pazcare can help

Pazcare understands that HR teams in India are navigating a rapidly evolving operational environment, more employees, more compliance, more expectation, and now a genuine imperative to integrate AI into the day-to-day. Our employee benefits platform is built to reduce the administrative load HR teams carry around insurance, wellness, and benefits management specifically, so that the time HR saves with AI tools can be directed toward what actually matters.

From digital group health insurance onboarding and e-card issuance to year-round wellness program management and claims support, Pazcare handles the operational complexity of employee benefits so HR teams are not managing it manually. Combined with the AI productivity practices covered in this blog, the result is an HR function that has both the time and the tools to lead strategically.

Talk to the Pazcare team today to understand how our platform reduces your administrative overhead and gives you the operational foundation to use AI in HR where it creates the most value.

Key takeaways

1. AI in HR is immediately accessible to every HR team in India today through free or low-cost LLM tools like Claude, ChatGPT, and Gemini, with no coding or technical expertise required, and as Tarun demonstrated live in the Pazcare webinar, a full personalized onboarding policy, a complete presentation deck, and a set of recruitment email subject lines can each be produced in under five minutes using these tools with the right prompts.
2. The CRAFT framework, Context, Role, Action, Format, Tone, is the practical prompting structure that separates generic AI outputs from genuinely usable HR documents, and applying it consistently, combined with the discipline of iterating through at least three to four prompt refinements, is what makes the difference between an HR team that tries AI once and abandons it and one that integrates it as a permanent productivity multiplier.
3. The boundaries of AI in HR are clear and must be respected: never share private employee data with public LLMs, always review AI-generated compliance documents before use, and never use AI as a substitute for human judgment in sensitive employee situations, with those boundaries firmly in place, the 87% of Indian companies in active AI adoption stages per the NASSCOM AI Adoption Index 2024 are building operational advantages that HR teams delaying AI adoption are choosing to concede.

Start using AI in HR today

The administrative backlog on your HR team's plate is not going to reduce on its own. The compliance requirements will keep growing. The documentation demands will keep compounding. The only variable you control is whether you start using AI to manage them more efficiently this week or wait until the gap between your team's capacity and the workload's demand becomes unsustainable.

Pazcare partners with HR teams to build benefits programs that reduce administrative complexity, and we can connect you with the right resources to start integrating AI into your HR workflows today.

Schedule a conversation with Pazcare and take the first step toward an HR function that works smarter.