The big-company-benefits PEO for white-collar SMBs. Real compliance lift under co-employment, a flat but unpublished admin fee, and a health book that was repriced hard through 2025 and 2026.
Independently scored, one dimension at a time. Every dimension is sourced, and we say whether a claim is published, reported, or estimated.
IRS CPEO plus ESAC accreditation (roughly 7% and 9% of PEOs respectively), 50-state filings under TriNet EINs, and benefit plan compliance owned by the plan sponsor rather than by you.
TriNet drives onboarding and small firms report going live in about two weeks; four to six is realistic with benefits. The friction is calendar math (mid-year wage bases, enrollment windows), not software.
Sweet spot is 5 to about 500 worksite employees. The March 2026 Multiplier partnership adds hiring in 150+ countries, but most companies still graduate off a PEO as headcount and self-insurance economics mature.
Connect 360 gives 24/7 chat, specialist pools and a named relationship manager. Offsetting that: 2025 retention fell about 5 points to roughly 80 percent, and BBB complaints cite benefits enrollment and withholding errors.
Accounting System Integration offers prebuilt connectors for QuickBooks Online, NetSuite, Xero and Sage Intacct plus a custom option, with customizable general ledger mapping from setup. Capped because TriNet does not publish whether the PEO invoice posts as a balanced journal or a bill, or whether the sync is automatic or a manual export, so the mechanics a controller needs are not on the page.
An Integration Administrator generates a Company ID, Client ID and Client Secret directly in Admin, Manager, Marketplace, so a customer gets credentials for its own company without waiting on TriNet, and OAuth scopes are documented. Capped because the developer portal itself is reached through Become a Partner, TriNet publishes no rate limits, and the write scope for payroll is not stated in public material.
TriNet Assistant (announced March 2026) is in-product only. The OAuth API is real but contract-gated with no self-serve sandbox, and there is no MCP server, official or otherwise, as of July 2026.
No published prices for PEO or ASO tiers; everything is custom-quoted. The 2025 book-wide health fee reset, with some renewals up over 30 percent, is the transparency risk made real.
Written by the Audit Friendly research team. No vendor edits, no sponsored placement.
TriNet is the PEO you pick when you want enterprise-grade benefits and multi-state payroll compliance without building an HR department, and it is one of the few providers holding both IRS CPEO certification (since 2018) and ESAC accreditation (since 1995). For a controller, the model is the point: TriNet co-employs your staff, files payroll taxes under its own EINs, sponsors the benefit plans, and sends one invoice bundling a flat per-employee admin fee with at-cost premiums and employer taxes.
That flat fee is more auditable than percent-of-payroll pricing, but none of it is published, and the invoice takes real GL mapping work to reconcile. Know the 2025-2026 context: TriNet repriced its entire health book, some January 2025 renewals rose more than 30 percent, retention fell to roughly 80 percent, and worksite employees dropped about 10 percent in 2025. It also killed the former Zenefits HRIS, so the software-only escape hatch is gone; the non-PEO tier is now HR Plus, an ASO averaging about $50 per employee monthly per its own earnings call.
TriNet remains a strong, credentialed operator on a healthier-priced book. Buy it for the benefits and compliance lift, not for price stability, and model your exit before you enter.
Jobs, mapped to how finance teams actually buy.
Under co-employment, wages are paid and employment taxes filed under TriNet's EINs. Multi-state registrations, SUI accounts and year-end W-2s are TriNet's problem while you are in. They become yours again, often at new-employer SUI rates, when you leave, which is why exits are planned for January 1.
TriNet's scale (about 300,000 average worksite employees in Q1 2026, per its 10-Q) buys national-carrier plan access small firms cannot get alone. The flip side arrived in 2025: TriNet repriced its whole health book, and a slice of January 2025 renewals rose more than 30 percent before moderating at the April renewals.
Only about 7 percent of PEOs are IRS-certified and about 9 percent are ESAC-accredited; TriNet has held both for years. CPEO status matters to finance: TriNet is solely liable for federal employment taxes on wages it pays, and joining mid-year does not restart FICA and FUTA wage bases thanks to successor-employer treatment.
TriNet's Accounting System Integration (ASI) supports QuickBooks Online, NetSuite, Xero, Sage Intacct and custom formats, with customizable GL mapping configured at onboarding. Without it, the single bundled invoice (admin fee, premiums, taxes, workers comp) reconciles to the GL by hand every cycle.
Forget the 40-row feature grid. Pick a finance workflow and see how this tool and its closest peers actually do it, with our verdict on who wins for whom.
Here's what it really costs: published rates where a vendor lists them, aggregated quote data where they don't, plus a path to your own number.
Nothing published; TriNet's own pricing page illustrates the math at a $150 PEPM rate. Third-party estimates cluster at $100-150, with rich bundles reported higher. All-in cost is several times this once medical premiums are added.
The tier that replaced the former Zenefits software. You keep your EIN, SUI rates and benefit plans; TriNet runs payroll and HR admin. The ~$50 average across 39,000+ users comes from TriNet's Q4 2025 earnings call, roughly 3x the old software-only price.
Dedicated payroll expert and dedicated HR manager add-ons for the ASO tier, priced on quote.
SaaS-only HRIS sunset beginning in 2025; the base was converted to HR Plus or PEO. If you are evaluating TriNet expecting Zenefits-style cheap software, that product no longer exists.
The biggest line on the invoice. After 2025's comprehensive health fee reset, some January 2025 renewals rose over 30 percent; TriNet says the over-30-percent share halved by April renewals. Model a bad-case renewal before committing.
Joining a CPEO mid-year carries federal FICA and FUTA wage bases forward under successor rules, so entry timing is forgiving. Leaving mid-year is not: wage bases restart under your own EIN, SUTA resets in most states, and employees get two W-2s that year. Plan any exit for January 1.
Medical, dental and vision plans are TriNet-sponsored, so your rate history and deductible credits need explicit handling on the way in. The 2025 book-wide health fee reset, with some renewals up over 30 percent and pricing the top cited reason customers left, shows renewal risk is real. Negotiate renewal visibility in writing.
Each invoice mixes the admin fee (opex), benefits premiums (pass-through) and employer payroll taxes plus workers comp. Booking it correctly requires the itemized fee split and department-level ASI mapping from day one; otherwise month-end reconciliation is manual spreadsheet surgery and headcount-driven variances are invisible.
On exit you re-establish state unemployment accounts at new-employer rates, stand up replacement benefits, run COBRA transitions, and spin your 401(k) assets out of TriNet's plan into your own. Export payroll registers, W-2 history and PTO balances while you still have admin access; data portability is on you.
Scored on what an outside agent can actually reach: the documented API, how it authenticates, and whether the vendor ships an MCP server.
TriNet's 2026 AI story is TriNet Assistant, announced March 24, 2026 at its Transform conference: a conversational gateway built on TriNet's data and organizational knowledge that answers questions, completes tasks and hands anything strategic to human HR specialists. It shipped alongside a partner-heavy platform push (TriNet Global powered by Multiplier for 150+ countries, TriNet IT powered by Electric AI, retirement connectors starting with Voya and Vestwell). As of July 2026 the assistant is an in-product experience with no published GA date and no external agent access.
The API surface is real and verifiable at developers.trinet.com: a v1 REST API at api.trinet.com with five families (Employee, Company, Payroll, Money and Time-Off), OAuth 2.0 client-credentials auth, read and write scopes assignable per endpoint and per company, and roughly one-hour access tokens. But access is not self-serve: TriNet provisions credentials only after you submit an intake form and an integration specialist follows up, and API access to your own company data requires a signed contract addendum (officer or director signature) per TriNet's integration center. TriNet also reserves the right to impose call quotas. There is no public sandbox, and the docs show little movement since 2023-2024; no webhooks or events are publicly documented.
On Model Context Protocol: TriNet has no first-party MCP server, and as of July 2026 none appears in the official MCP registry or major directories. The pages you will find (Guru, fastn and similar) are third-party speculation or wrapper platforms, not TriNet offerings. Unified-API vendors (Finch, Merge, Apideck) do cover TriNet for normalized, read-heavy HRIS and payroll data, which is currently the fastest legitimate path for an agent to see TriNet data without custom paperwork.
What that means in practice for a finance team: once credentials exist, an external agent can read rosters, compensation, payroll register data and time-off balances, and could automate GL journal preparation by combining API reads with ASI exports. No agent can run or approve a payroll, initiate money movement, or touch the service layer (cases, benefits questions), and the co-employment workflows that make a PEO valuable live entirely inside TriNet's platform and people. The score reflects a functional but gated API and zero agent-native surface.
Who wins for whom: A PEO that carries the filings under its own EINs: TriNet. Maximum compliance cover and a penalty guarantee: ADP. A simple US payroll stood up by a specialist in days: Paychex.
A PEO that carries the filings. Files in all 50 states under TriNet's own EINs, as an IRS certified and ESAC accredited PEO.
Where it loses: Plan ownership and renewal power pass to TriNet, so budget the off-boarding before you onboard.
Source: AF score rationales: Compliance depth 88, Implementation 78; TriNet watch-outs.
Watch: COCPA Preferred Partner Demo: TriNet (2024) 27:43 · Demo for a CPA society · Colorado Society of CPAs on YouTube
Deepest compliance cover. Files and remits in every US jurisdiction, and publishes a guarantee to pay penalties caused by its own filing errors.
Where it loses: RUN goes live in days, but Workforce Now takes 4 to 14 weeks, and moving up a tier means a fresh implementation.
Source: AF score rationales: Compliance depth 96, Implementation 62; ADP watch-outs.
Watch: RUN Payroll Demo 5:40 · Independent walkthrough · Middle Class Climber on YouTube
Specialist-led fast setup. A specialist can stand up a simple US payroll in about 48 hours, with automated filing in all 50 states and year-end W-2s.
Where it loses: You register the states, not Paychex, and that can block the first run. It is domestic only, with no employer of record.
Source: AF score rationales: Compliance depth 82, Implementation 72; Paychex watch-outs.
Watch: How It Works: Paychex Flex Demo 2:01 · Official demo · Paychex on YouTube (official)
Who wins for whom: A PEO invoice into QuickBooks, NetSuite, Xero or Intacct: TriNet, once you confirm how it posts. On ADP RUN with QuickBooks Online: ADP. Balanced entries each run, included from the Pro tier: Paychex.
Connectors, unclear mechanics. Prebuilt connectors for QuickBooks Online, NetSuite, Xero and Sage Intacct, with general ledger mapping you can customize from setup.
Where it loses: It does not publish whether the PEO invoice posts as a balanced journal or a bill, or whether the sync is automatic.
Source: AF score rationale: GL / ERP fit 65.
No public demo of this task. Ask for a sandbox run into your own system before you sign.
Clean on RUN, patchy above it. RUN posts journal entries straight into QuickBooks Online, and its NetSuite connector posts one balanced entry per run for each subsidiary.
Where it loses: Workforce Now has no native QuickBooks Online connector, NetSuite and Sage Intacct mean a paid third-party app, and RUN still needs manual entries for classes and SUI rate changes.
Source: AF score rationale: GL / ERP fit 62.
Watch: ADP Accountant Connect + QuickBooks Online sync 14:45 · Independent setup guide · QCoach, Simplify Your Books on YouTube
Four real ledger connectors. Posts balanced entries each run into QuickBooks Online, NetSuite and Sage Intacct, which also receives locations, classes and departments.
Where it loses: The GL service is a paid add-on below the Pro tier, it only pushes one way, and there is no Dynamics or Business Central.
Source: AF score rationale: GL / ERP fit 76; Paychex watch-outs.
Watch: Entering Paychex payroll into QuickBooks Online manually 16:11 · Independent walkthrough, manual entry · QuickBooksQBO on YouTube
At its Transform conference TriNet unveiled TriNet Assistant (a conversational AI gateway backed by human HR specialists), TriNet Global powered by Multiplier for compliant hiring in 150+ countries, TriNet IT powered by Electric AI for device and access management inside onboarding, and retirement plan connectors starting with Voya and Vestwell. No GA dates were given.
Q4 2025 results closed the repricing chapter: full-year revenue of $5.0B (down 1 percent), roughly 323,000 worksite employees at year end (down about 10 percent), retention around 80 percent with pricing the top cited reason for leaving, and 39,000+ ASO users averaging about $50 per month. 2026 guidance: $4.75B to $4.90B revenue with an insurance cost ratio of 89.25 to 90.75 percent.
Alongside Q4 2024 results, TriNet confirmed it would discontinue the SaaS-only HRIS (acquired as Zenefits in 2022, renamed TriNet HR Platform in late 2023) starting in 2025, converting customers to the HR Plus ASO tier or full PEO at roughly three times the old software price. Management later said conversion exceeded expectations.
TriNet publishes no list prices; its pricing page only illustrates the math at $150 PEPM. Vendr and buyer reports put PEO admin fees around $100-150 per employee per month, billed as a flat fee rather than a percent of payroll. On top of that you pay benefits premiums, employer taxes and workers comp at cost, so the all-in invoice typically runs several times the admin fee. The non-PEO HR Plus (ASO) tier averaged about $50 per user monthly per TriNet's Q4 2025 earnings call. Demand an itemized quote separating admin fee from pass-throughs; that split is your only real comparison tool.
Contractually you are not: reports cite no long-term commitment and a 30-day written cancellation notice, though fees can apply in some cases. Operationally, leaving is a project. A mid-year exit restarts FICA, FUTA and most state unemployment wage bases under your own EIN (double taxation on the reset portion), gives employees two W-2s, and forfeits TriNet-sponsored benefit plans immediately, so you need replacement plans, COBRA handling and a 401(k) spin-out ready. Nearly everyone exits January 1. Export payroll registers and history before your admin access ends.
TriNet's implementation team drives it. Small, single-state teams report going live in about two weeks; two to six weeks is the realistic range once benefits enrollment windows, prior-payroll data loads and state account transfers are involved. The finance-side work is front-loading the Accounting System Integration: mapping the bundled invoice to your chart of accounts (QuickBooks Online, NetSuite, Xero, Sage Intacct or custom) and by department, so the first invoice books cleanly instead of becoming a recurring manual allocation exercise.
Not out of the box. TriNet has no MCP server (nothing first-party, nothing in the official registry as of July 2026). It does have a real v1 REST API at api.trinet.com covering Employee, Company, Payroll, Money and Time-Off data with OAuth 2.0, but credentials are provisioned only after an intake form, an integration specialist conversation and a signed contract addendum; there is no self-serve signup or public sandbox. Aggregators like Finch and Merge offer normalized read access. TriNet Assistant, announced March 2026, is an in-platform AI and offers no external agent hooks.
Answered only from our own published research on this tool, never from general internet noise. If we cannot answer it well, our research agents will dig in and publish a sourced answer.