Tool · Insurance ops · Zywave · Oct 2026

Zywave AI review (Oct 2026): Zywave Apex, Producer Agent, Advisor Agent, and the insurance MCP server for agency growth

Zywave sells software and data to insurance agencies, brokers, carriers, and MGAs: the miEdge prospect database, Content Cloud articles, TurboRater comparative quoting, and agency management systems. On 7 July 2026 it launched Zywave Apex, an AI platform for the front office with three parts: a Producer Agent for new-business prospecting, an Advisor Agent for renewals (still early access), and an MCP server that puts Zywave data inside Claude, ChatGPT, Copilot, or Gemini. This page is about what each part does, how available it is, and who signs off. Each section has a Try this.

Pair with HubSpot AI (general CRM outreach) · Also new: Duck Creek AI · Sixfold AI · Oct 2026 · zywave.com

What this Zywave page covers

This page is for agency owners, sales leaders, producers, and account executives at independent agencies and brokers, especially those already using Zywave data or quoting tools. The question it answers: what Zywave Apex’s agents and MCP connection do for prospecting and renewals, which parts are live and which are early access, and who must approve what goes to a prospect or client.

Sources are Zywave’s own pages, read on 18 October 2026: the Apex launch release (7 July 2026), the Summer 2026 release (30 July 2026), and the Apex, Producer Agent, Advisor Agent, and MCP pages. Zywave first announced insurance AI agents on 16 December 2025 and said then they would be generally available in Q1 2026; we could not confirm the date that happened, but the Apex page now labels Producer Agent Live.

Ground rules: the producer approves every outbound email before it sends; the account executive checks every renewal on an Advisor Agent list against the client file before anyone calls; a licensed person reviews any coverage-gap note before it reaches a client; and Zywave’s engagement figures are its own claims, not our measurement. This page explains features. It is not coverage or legal advice.

Simple Zywave AI brand mark diagram
A simple mark we drew for this page. Where we couldn’t get a clean product screenshot, we used simple diagrams of the workflow instead.

Major features (with how someone uses them)

Examples at a glance: each feature has a concrete example from a real job plus a short Try this you can run this week.

1. Advisor Agent for renewals (primary, early access)

What it is: The Advisor Agent watches in-force accounts as they enter a 90-day renewal window, ranks them by churn risk, coverage-gap size, and revenue impact, pulls policy and carrier appetite details into a renewal brief, and hands producers a prioritised action plan. It connects to TurboRater for the quoting handoff. It is labelled Early Access: Zywave’s early adopter program opened on June 23, 2026, and the agent page lists Agency Matrix and Partner Platform as the supported agency systems, with others assessed case by case.

A ranked renewal list decides who your team calls first. If the list misreads an account, the client who was about to shop gets no call, or a happy client gets a nervous one. The client file and the account executive’s knowledge come first.

Concrete example (account executive at a mid-size commercial agency) An account executive in the early access program asks Advisor Agent for the five renewals most at risk next quarter. She checks each against the client file. Three make sense: a large premium increase, a carrier leaving the class, and a coverage gap after a new location. One client already re-signed a multi-year deal, and one is flagged for a gap that an umbrella policy at another carrier covers. She calls the three and notes the two misses for the Zywave team.
Ask for the list, check the file, note misses, then call
Simple workflow diagram: ask for the list, check the file, note misses, then call.
Try this

Primary workflow:

Ask Advisor Agent for the five renewals most at risk next quarter. Have the account executive check each one against the client file before anyone calls the client.

2. Producer Agent: finding and researching prospects

What it is: The Producer Agent (labelled Live) starts from a target audience you describe, or one it builds from your agency system, and matches it against Zywave’s miEdge data (Zywave cites 4M+ employers and 85M+ household profiles). It then looks up context such as hiring data, OSHA records, broker-switching signals, and renewal timing. Zywave lists pre-built integrations with Applied Epic, AMS360, Salesforce, and others, and tells buyers to confirm current integration availability with a rep.

Prospect data ages. A renewal date or decision-maker that was right last year can be wrong now, so the producer should check the facts on any account before using them in a conversation.

Concrete example (producer at a benefits agency) A producer asks Producer Agent for manufacturers with 100–300 employees in her county. Before any outreach, she checks the top ten against LinkedIn and the companies’ own sites. Two decision-makers have left, and one company was acquired. She removes them and keeps the rest.
Describe the audience, check the facts, remove misses, then reach out
Simple workflow diagram: describe the audience, check the facts, remove misses, then reach out.
Try this

Check before outreach:

Ask Producer Agent for twenty prospects in one niche. Have the producer check each decision-maker and renewal date against a second source before any of them gets an email.

3. Producer Agent outreach: approve before anything sends

What it is: Producer Agent writes multi-touch email sequences using Zywave’s Content Cloud articles, and, once the producer approves, sends them from the producer’s own Outlook through Office 365. The Summer 2026 release added deeper Outlook integration and meeting booking on the producer’s calendar. Zywave reports 36%+ open rates (“1.7x industry average”) from its production data; its pages give the click-rate lift as both “10x” and “20x” industry average, so treat these as vendor claims.

Emails go out under a licensed producer’s name. One sentence that sounds like a coverage promise is an E&O problem, so every sequence needs a real read, not a skim.

Concrete example (producer at a personal and commercial lines agency) A producer reviews her first sequence of five emails to restaurant owners before approving it. Four are fine. One line says the agency can “make sure you’re fully covered for food spoilage”. She rewrites it to offer a review of the current policy instead, approves the sequence, and sends herself a copy of every touch for the first month.
Draft, read every line, fix promises, then approve
Simple workflow diagram: draft, read every line, fix promises, then approve.
Try this

Read before you approve:

Have the producer read every email in one Producer Agent sequence before approving it. Remove any sentence that promises coverage or price before anything sends.

4. Zywave Apex MCP: insurance data inside Claude, ChatGPT, Copilot, or Gemini

What it is: The Apex MCP server (labelled Live) connects an AI assistant your agency already uses to Zywave’s data and tools through the Model Context Protocol. Zywave lists Claude, ChatGPT, Microsoft Copilot, Gemini, and any MCP-compatible client. Producers can ask for a prospect list, a research brief before a first meeting, a compliance or HR article for a client, or an account update, in plain language. Zywave’s launch blog says the server connected 24 live production tools at launch, and its MCP page says “Your AI confirms every action before executing.”

The assistant now acts on your book, not just answers questions. “Add this company to my book” changes a record, so producers should read the confirmation every time, and someone should review what changed.

Concrete example (account manager at a commercial lines agency) An account manager asks Claude, connected through Apex MCP, for a research brief on a restaurant group before a renewal meeting and then to add two new locations to the account. She reads the brief against the client file and finds the employee count is two years old. When the assistant asks to confirm the record update, she corrects one address first. The agency’s operations lead reviews MCP record changes weekly for the first month.
Illustrative AI assistant connected to Zywave Apex MCP showing what producers ask, what the connection can do, and what the agency keeps
Our sketch of how Zywave describes the MCP connection. Not a Zywave screenshot, and not an official product screenshot — the layout is ours.
Try this

Read every confirmation:

Connect Apex MCP for two account managers only. Have the agency operations lead review every record change the assistant made in the first two weeks before more people connect.

5. Access controls, audit trail, and permissions

What it is: Zywave says Apex connections use OAuth2 with two-layer scopes and RFC 8707 token binding, keep a full audit trail of every tool call (OpenTelemetry), strip personal data from logs automatically, and are covered by SOC 2. Access mirrors each user’s existing Zywave licences, and admins can restrict specific Apex tools by role, team, or person. Setup is approved once at the organisation level by your IT admin with Zywave’s implementation team (MCP page, Apex page).

Controls only help if someone sets them. The default should be the fewest tools for the fewest people, widened as the agency sees how they are used.

Concrete example (operations manager at a 40-person agency) An operations manager turns on research and content tools for all producers but limits record-changing tools to two account managers for the first month. She reads the audit trail every Friday, finds one producer exporting a large prospect list, and asks why before widening access.
Limit tools, read the log, ask questions, then widen
Simple workflow diagram: limit tools, read the log, ask questions, then widen.
Try this

Start narrow:

Have the operations manager list which Apex tools each role really needs. Turn on only those, and read the audit trail weekly for a month before giving anyone more.

6. Apex AI Studio prompts, research briefs, and client content

What it is: Apex AI Studio is Zywave’s collection of ready-made prompts and “skills” for common jobs: engaging a new prospect, researching before a call, sharing resources with clients, and managing your book (MCP page). Content requests pull from Zywave’s library of expert-written compliance, HR, and risk articles (Zywave cites 15,000+), matched by state and industry.

Content written by Zywave’s experts is a better start than generic AI text, but the agency still owns what it sends. A state-specific article has to match the client’s state and the current year.

Concrete example (account manager at a benefits agency) An account manager uses a Studio prompt to find a leave-law article for a client in one state. The assistant returns one for the right state, but it is dated before the state’s most recent change. She asks for the current version, checks the date, and only then sends it with her own note.
Pick a prompt, check state and date, add your note, then send
Simple workflow diagram: pick a prompt, check state and date, add your note, then send.
Try this

Check before you share:

Have the account manager run three Apex AI Studio prompts the team uses most. Check each returned article’s state and date before anything is shared with a client.

Pricing context (October 2026)

Zywave does not publish prices for Apex, Producer Agent, Advisor Agent, or the MCP server (its /pricing address returns “not found”). Pricing is quote-only as of 18 Oct 2026. Advisor Agent early adopters get “preferential” pricing set out in an early adopter agreement. Ask whether Apex needs existing Zywave products such as miEdge or Content Cloud, which tools come with each licence, and what changes when Advisor Agent leaves early access.

Zywave says Producer Agent takes about 30 days to set up and the MCP connection under an hour, with larger multi-system setups taking four to six weeks.

Uncertainty: Advisor Agent is early access and its agent page lists only Agency Matrix and Partner Platform as supported agency systems; the Applied Epic, AMS360, and Salesforce integrations Zywave lists apply to the wider platform and Producer Agent, and it tells buyers to confirm them with a rep. Zywave said in December 2025 its first agents would be generally available in Q1 2026; we could not confirm when that happened. Engagement figures are Zywave’s own, and its click-rate lift appears as both “10x” and “20x”. We could not confirm prices, and we did not test Zywave ourselves.

When it’s useful (honest take)

Useful when you are an independent agency or broker that wants more new-business conversations without more admin, you already use Zywave data or quoting tools, your producers use Claude, ChatGPT, Copilot, or Gemini at work, and every producer will read their sequences before approving them.

Less useful when you are a carrier looking for underwriting or claims AI (see Sixfold AI or Duck Creek AI), your agency system is not supported for the agent you want, or you need renewal help now and cannot join Advisor Agent’s early access group.

Paying for Zywave AI and a few other AI tools? The $9 AI toolkit spreadsheet lists every AI subscription you have, what each one costs, and how fast it pays for itself, so you can see which ones are worth the money.

Want someone to do it for you? AI toolkit audit — $19

Related options (educational)

Where agency growth and renewal AI jobs can live (Oct 2026)
NeedBetter homeWhy
Insurance prospecting, renewal lists, and an insurance MCP connectionZywave AI (this page)Quote only; Advisor Agent early access
AI inside an agency management systemApplied Epic AI or Vertafore AIQuote only; for agencies
General CRM with AI email helpHubSpot AI or Salesforce EinsteinSee their reviews
Sales sequences outside insurance-specific dataOutreach AI or ApolloSee their reviews
Drafting and research in a general assistantMicrosoft 365 Copilot or ClaudeSanitized client data only, per agency policy

FAQ

What is Zywave Apex?

Zywave’s AI platform for insurance distribution, launched 7 July 2026. It has three parts: Producer Agent for new-business prospecting and outreach, Advisor Agent for renewals, and an MCP server that connects Zywave data and tools to AI assistants such as Claude, ChatGPT, Copilot, and Gemini.

Is Zywave Advisor Agent available?

Only in early access. Zywave’s Apex page labels it Early Access, its early adopter program opened June 23, 2026, and its agent page lists Agency Matrix and Partner Platform as the supported agency systems.

Does Producer Agent send emails on its own?

Zywave says the producer approves the sequence first, and then emails send from the producer’s own Outlook through Office 365. Read every email before approving, especially any line that sounds like a coverage promise.

How much does Zywave Apex cost in October 2026?

Zywave does not publish prices. Treat it as quote-only. Advisor Agent early adopters get preferential pricing under an early adopter agreement.

Who should approve Zywave AI output?

The producer for every outbound email, the account executive for every renewal call list, a licensed person for any coverage-gap note before a client sees it, and the operations manager for who can use record-changing tools.

Optional next step (if seats overlap)

If your agency pays for a prospect database, an outreach tool, CRM AI, and chat seats that all touch the same accounts, that overlap is worth mapping. Map the overlap with the $19 AI toolkit audit or the $9 AI toolkit spreadsheet. Email Audit $19 or Toolkit spreadsheet $9.

Checkout is live via Stripe. After you pay, we email delivery (spreadsheet) or intake next steps (audit).

Related: Duck Creek AI · Sixfold AI · Applied Epic AI · Vertafore AI · Guidewire AI · Salesforce Einstein · Outreach AI · Microsoft 365 Copilot · HubSpot AI · DocuSign AI · ChatGPT for finance · Claude for legal ops · Claude · Three checks before another AI seat · How-to: Use Zywave’s Producer Agent to prep renewal outreach, and have the producer approve every outbound message · ops hub

Affiliate disclosure: Some outbound product links may be affiliate links. We may earn a commission if you subscribe. Prices and features change — always verify on the vendor site. Full policy: disclosure.

Next steps