People & HR

How to manage benefits enrollment with AI

People & HR3 AI tools7 steps6 friction points

Want an agent to handle this on your live data? See the Starch version →

Benefits enrollment is the annual (or ongoing) process of collecting employee elections, verifying eligibility, communicating plan options, chasing down missing forms, and getting everything into your HR or benefits administration system on time. For small and mid-sized teams, this usually falls on a single operator or HR generalist who has to coordinate across carriers, employees, and payroll — all while the rest of the business keeps running. The window is tight, the stakes are real, and the margin for error is low.

The workflow feels like an AI problem because so much of it is structured, repetitive communication: writing plan comparison summaries, drafting enrollment reminder emails, answering the same five questions from employees about deductibles and HSA limits, and producing checklists that don't change much year to year. If you can describe the task clearly, an LLM can generate a draft — and for operators who are doing this alone, that feels like getting a capable assistant on short notice.

ChatGPT, Claude, and Gemini can genuinely help here. They're strong at drafting employee-facing communications, summarizing plan differences in plain English, building enrollment checklists, and writing FAQ documents from the raw plan details you paste in. They won't connect to your HRIS, send emails, or track who has and hasn't submitted their elections — but as writing and thinking partners for the content layer of enrollment, they're useful today.

Skip the manual work → See the Starch version

People & HR3 AI tools7 steps6 friction points

AI walkthrough

How to do it with AI today

A practical walkthrough using ChatGPT, Claude, and other off-the-shelf LLMs — what they're good at, what you'll have to do by hand.

Tools that work for this

ChatGPTClaudeGemini

Step-by-step

1 Gather your raw materials: export or copy the plan summaries from your benefits broker or carrier portal — premiums, deductibles, out-of-pocket maxes, network type — and paste them into Claude or ChatGPT as your starting context.

2 Ask the LLM to produce a plain-English plan comparison table your employees can actually read, formatted for a company all-hands email or a shared Google Doc. Specify that it should avoid insurance jargon and flag the key trade-offs (low premium vs. low deductible).

3 Prompt the LLM to draft a full enrollment kickoff email: dates, deadlines, where to enroll, who to contact with questions, and a one-paragraph summary of what's changed from last year. Paste in last year's email if you have it for continuity.

4 Use the LLM to generate a benefits FAQ document. Paste in the plan PDFs or summary of benefits and ask it to anticipate the 10-15 most common employee questions and write clear answers. This becomes your employee-facing resource.

5 Ask for a enrollment project checklist — broken out by role (HR, employees, payroll) and deadline — covering every step from open enrollment announcement through carrier confirmation and payroll deduction setup.

6 Draft reminder email sequences (day-of-open, one-week warning, final-48-hours) by prompting the LLM with the kickoff email as context. Ask it to vary the tone and add urgency progressively without being alarmist.

7 After enrollment closes, paste your list of employees and their submission status into the LLM and ask it to identify who still hasn't responded and draft a personalized follow-up message for that group.

Prompts you can copy

Here are three health plan options from our carrier. Write a plain-English comparison table for employees, highlighting monthly premium cost, deductible, out-of-pocket max, and the single most important trade-off between each plan. Avoid insurance jargon.

Draft an open enrollment kickoff email for a 25-person company. Enrollment opens November 1 and closes November 15. We're on Gusto. New this year: we added a dental plan. Tone should be clear and direct, not corporate.

Based on these plan details, write a benefits FAQ document answering the 12 most common questions employees ask during open enrollment. Format it as Q&A. Keep answers under 60 words each.

Write three enrollment reminder emails — 7 days before deadline, 3 days before, and day-of — that escalate in urgency. Start friendly, end firm. Each should be under 100 words.

I have a list of 22 employees. 14 have completed enrollment, 8 have not. Draft a short, direct follow-up email to the 8 who haven't submitted, reminding them the deadline is Friday and explaining what happens if they miss it (auto-enrolled in current plan).

Reality check

Where this gets hard

The walkthrough above works — until your numbers change, the LLM hallucinates, or you have to re-paste everything next month.

No connection to your HRIS, Paylocity, ADP, or Gusto — you manually copy employee lists and status data into the chat every single session, with no guarantee the list is current.

Nothing persists between sessions. The FAQ doc you built last Tuesday, the reminder sequence you refined, the checklist — all of it lives in a browser tab. Close it and you start over next year.

LLMs can't track who has responded. You still need a separate spreadsheet or manual process to monitor enrollment status and decide who gets the follow-up nudge.

Reminder emails don't send themselves. Every draft the LLM produces has to be manually copied into your email client or HRIS — there's no connection between the output and the system that actually delivers the message.

Outputs drift. The enrollment FAQ format you carefully dialed in during last year's open enrollment isn't reproducible without saving and re-pasting the exact prompt — and even then, results vary.

Large plan documents hit context limits fast. Paste in three full Summary of Benefits and Coverage PDFs and you're already pushing the window, with the LLM starting to drop or paraphrase detail from earlier sections.

Tired of the friction?

Starch runs the whole workflow on live data — no copy-paste, no hallucinated numbers, no re-prompting next month.

See the Starch version →

Starch alternative

The same workflow on Starch

Starch is an agentic operating system — the layer where an agent builds and runs the actual software your benefits enrollment workflow depends on. Instead of one-off prompts you re-run manually, Starch builds persistent apps and automations connected to your live HR data in Paylocity or ADP.

Starch syncs your Paylocity or ADP data on a schedule — employee roster, enrollment status, benefit elections — so every surface you build reflects who's actually enrolled today, not a CSV you exported last week.

Describe the enrollment tracker you need in plain English — 'build me a dashboard showing each employee's enrollment status, the plan they selected, and a flag for anyone who hasn't submitted by the deadline' — and Starch's agent builds it.

Connect Gmail or Outlook from Starch's integration catalog and set up an automation that sends enrollment reminders on a schedule, targets only employees who haven't completed their elections, and logs each send — no manual copy-paste into your email client.

Use the Knowledge Management starter app to publish your benefits FAQ, plan comparison docs, and enrollment guides in one searchable place — AI finds answers across all your docs instantly so employees stop emailing you the same questions.

The Email Agent starter app triages enrollment-related questions hitting your inbox, summarizes threads, and drafts replies — so you're not personally answering 'what's the difference between the PPO and HDHP' for the fifteenth time during the enrollment window.

Task Manager (currently in beta) lets you capture every enrollment to-do — carrier confirmations, payroll deduction setup, late submitters — with deadlines and priority levels, so nothing slips in the two-week crunch.

Get closed-beta access →

Toolkit