Every growth plan in specialty insurance rests on a market read. Most of the ones we see rest on a single sentence — "the market is softening" — and then spend the rest of the document on tactics. The sentence is correct and the plan is wrong, because commercial lines have not softened as one market since 2023. They have split. Capital has flooded property and workers' comp; loss costs have kept auto and umbrella rising for fifteen years straight. A program on the soft side needs to be recognized and remembered. A program on the hard side needs to be fast and findable. The tactics that follow are different, and so are the numbers you should expect.
This brief lays out the data we plan against, line by line, and the operating model we build from it. We update it quarterly as the indices publish. Where a number is an estimate or a single source, the exhibit says so.
The market in three numbers
Sources: CIAB Commercial P&C Market Index Q2 2026; NCCI State of the Line 2026; AM Best commercial auto review, Nov. 2026. All figures US.
Methodology
We read five kinds of sources and weight them differently, because each measures a different thing.
- Broker surveys (CIAB quarterly index) tell us what agents are seeing at renewal across account sizes. Fastest signal; sentiment-weighted.
- Brokered-account indices (Marsh Global Insurance Market Index) tell us where the large-account market is heading. Direction leads the middle market by two to four quarters.
- Bureau and rating-agency data (NCCI State of the Line; AM Best segment outlooks and line reviews) tell us whether pricing is earned — combined ratios, reserve positions, loss trends. Slowest signal; most durable.
- Reinsurance renewal reports (Guy Carpenter, Gallagher Re, Aon) tell us what capacity will cost carriers next year, which sets the ceiling on how soft a line can get.
- Stamping-office data (WSIA midyear report) tells us which risks the admitted market has stopped writing, in dollars.
A lane is "soft" when all three of pricing, capacity and reinsurance cost are moving down together; "hard" when pricing is rising and the line's combined ratio is above 100. Mixed signals are flagged as transitional. We do not use carrier marketing releases or vendor "state of the market" decks as sources.
Section 1 · Where rates are moving
CIAB's index recorded its first composite decline in Q1 2026 and a second in Q2. Property led the way down, with 75% of respondents reporting more capacity. Workers' comp and cyber followed. Commercial auto and umbrella moved the other way in the very same survey, and umbrella's streak of increases reached 35 quarters. Exhibit 1 shows the split four ways: by line, by account size, in surplus-lines dollars, and at the large-account end of the market.
The composite says soft. Two lines in the same survey say hard.
The bigger the account, the softer the market.
Surplus lines tell the same story in dollars.
At the top of the market the split is wider still.
Two things in the exhibit shape how we plan. First, the small-account market barely moved (−0.5%), so the agencies most of our clients sell through have not yet felt the commission squeeze the trade press describes; they will feel it over the next four to six quarters as the large-account softening works its way down. Second, surplus-lines auto liability grew almost 16% while E&S property fell almost 14%: the admitted market is taking property back and pushing auto out. If your program lives in the E&S auto lane, your submission flow is rising. If it lives in E&S property, your floor is rising with it.
Section 2 · Two stories, one headline
The reason the split persists is that the two sides are driven by different engines. Property pricing is a capital story: a record year for reinsurer capital, a quiet catastrophe season, and new entrants chasing yield. When capital is abundant, property gets cheaper no matter what any one carrier does. Auto and umbrella pricing is a loss-cost story: nuclear verdicts, litigation funding, repair and medical inflation, and a decade of liability loss ratios above 100. Abundant capital does not fix a loss ratio.
Soft lanes · capital story
Property, workers' comp, cyber, financial lines. Priced by how much capital wants in. Brokers have options and shop; recognition and relevance win.
Hard lanes · loss-cost story
Commercial auto, trucking, GL, umbrella and excess. Priced by what claims cost. Brokers are stacking towers; speed, clear appetite and a real answer win.
Exhibit 2 shows why the hard lanes stay hard: comp ran a 91 combined ratio in 2025 with a $14B reserve cushion, auto liability ran 108.6, and the casualty reinsurance that sits behind umbrella towers is, in Guy Carpenter's word, "nuanced" — available, but priced to each cedent's loss experience. First-layer towers that used to go to $25M now commonly stop at $5–10M.
Comp earns money. Auto liability loses it. Pricing follows.
The headline streak ended. The hard-lane streaks did not.
Cheap property-cat reinsurance does nothing for a trucking umbrella.
Agencies that write both, which is most of them, experience the split as a squeeze: shrinking property commissions on one side and exhausting auto and umbrella placements on the other. A program that takes a hard problem off their desk and puts commission back on their P&L is the easiest conversation in the market right now. That is the opening we plan around.
Section 3 · Why the conventional playbook breaks
The conventional growth playbook for an MGA, PEO or program is largely the same document regardless of lane: a positioning statement, a content calendar, a website rebuild, paid media, and a CRM. It is not wrong so much as under-specified. Exhibit 3 sets its working assumptions against what the data supports and what we do instead.
| Question | Conventional playbook | Two-speed playbook |
|---|---|---|
| What market are we in? | One market, softening. Plan for price competition everywhere. | Two lanes. Name the lane for each program and state; soft lanes compete on recognition, hard lanes on speed and appetite. |
| How much business will we lose? | Normal attrition, typically assumed in the low single digits. | Raise it. In soft lanes, standard markets are taking accounts that had no home two years ago; build the graduation story into the pitch and the forecast. |
| Who do we target? | Personas: 'the independent agent', 'the broker'. | Named lists built from Apollo, Sales Navigator, FMCSA data and your own submission history, matched to appetite by class and state, with explicit do-not-contact rules. Carriers are never prospected. |
| What is content for? | Top-of-funnel awareness; a blog that proves expertise. | Something a broker can forward to an insured — a mod-management guide, a seasonal loss alert, a market read — cut into posts, emails and scripts and timed to renewals. |
| When do we publish? | A monthly calendar; publish once and move on. | One cornerstone asset per quarter, many touches, peaking 90–120 days before each of the four renewal dates. |
| Where does the money go? | Ongoing website and paid-media spend. | Fix the site once, then spend on reach and follow-up: data, outreach, automation and the people who approve replies. |
| What does AI do? | Writes the content. | Drafts from a governed knowledge base of appetite, claims experience and approved language; a specialist signs anything a prospect sees. Approval gates wherever money or a relationship moves. |
Exhibit 3. The left column is a composite of growth plans we have reviewed for MGAs, PEOs and program administrators since 2024; no single firm is described.
Section 4 · The operating model
An engagement runs in four phases. The first two produce a written read and a design; the second two build and then operate the machine with your team until the numbers hold on their own. Each phase has a fixed length, a deliverable, and a decision point where you can stop.
Diagnose
Lane read by line and state. Submission history, win/loss, partner map, stack audit. Deliverable: the written read and three moves.
Design
Positioning, appetite statement, named target lists, content spine, renewal calendar, metrics. Deliverable: the operating plan with owners and numbers.
Build
CRM, data, outreach sequences, automations and knowledge base stood up and tested. Deliverable: a running system and a trained team.
Run
Weekly scoreboard, monthly tuning, quarterly market read. Fractional or handed off. Deliverable: bound premium you can attribute.
Exhibit 4. Diagnose is sold on its own as the 1-hour strategy session and the half-day working session; most clients begin there.
The eight capabilities the system has to have
Read
Where your lanes sit in the cycle, by line and state. Who still needs you, who is about to stop, and what the bureaus and brokers said this quarter.
Position
One sentence your producers can repeat, an appetite statement by class and state, and the proof behind it. Written for brokers and for AI search engines.
Target
Named lists, not personas: agents and wholesalers matched to appetite and territory, with explicit do-not-contact rules.
Content
One cornerstone asset per quarter, cut into posts, emails, scripts and conversations, governed by an approved knowledge base.
Campaigns
LinkedIn, email, voice and social running in sync against the renewal calendar, with a human approving every reply a prospect sees.
Capture
Leads, submissions and partner activity in one CRM with source, owner and stage. Inbound alias routing so you know which partner sent what.
Automate
Nurture, renewal-date campaigns, re-engagement of past submissions, daily scoreboards, inbox triage — workflows that run while you sell.
Measure
Touches, conversations, submissions, quotes and binds by partner and channel, weekly. Attribution from first touch to bound premium.
Section 5 · Timed to the renewal clock
Submissions cluster around January 1, April 1, July 1 and October 1 effective dates. Brokers start marketing an account 90 to 120 days out, so the content and outreach cycle has to peak before that window opens and stay quietly present in between. Exhibit 5 is the calendar we build every client's plan on; the lane decides what goes into each push.
Exhibit 5. January: market-read content, appetite refresh and partner activation before Thanksgiving. April: mod-management and loss-control content as new mods publish. July: seasonal exposure alerts and the mid-year market read. October: re-engage past submissions and lost quotes.
Section 6 · The system, not the tools
Tools are the least interesting part of the plan and the part most plans lead with. What matters is that each stage of the system has a tool, an owner and a number, and that nothing reaches a prospect without a human signature. Exhibit 6 shows the stack we run today; the amber chips are products we built because nothing on the market did the job for insurance distribution.
Exhibit 6. Amber chips are Seven16-built. Tool choices change; the five stages and the approval rule do not. Full feature detail is on the Products and Stack pages.
Section 7 · Where AI belongs
AI accelerates research, drafting, routing and reporting, and it is very good at those. It does not know your appetite, your claims experience or your broker relationships unless you build the knowledge base that holds them — and it will state an appetite you do not have with complete confidence. So every automation we build has an approval gate wherever money, a prospect relationship or your reputation moves.
- Gate 1 · KnowledgeAI drafts only from an approved knowledge base: appetite by class and state, claims experience, language legal has cleared.
- Gate 2 · SignatureA specialist reads and signs every reply, proposal and post a prospect sees. Drafting is automated; sending is not.
- Gate 3 · Money and dataAnything that moves money, changes a record at scale, or contacts a new list requires a human click and a logged before-value.
What we ask clients to stop doing
Section 8 · Indicators we monitor
The two-speed read is a position, not a prediction. These are the indicators we watch each quarter, and what each one would mean for a client's plan if it moved.
- CIAB commercial auto turns negativeThe hard lane starts to soften; shift auto programs toward the recognition playbook before brokers have options.
- Umbrella streak breaksWatch first-layer capacity return to $25M. Trucking and habitational programs lose their scarcity advantage.
- Auto liability combined ratio falls below 100Pricing has caught loss costs; expect new capacity and new competitors within two quarters.
- Property-cat reinsurance stops fallingA heavy hurricane season or a capital pullback puts a floor under property; attrition assumptions can ease.
- Workers' comp filings turn positiveComp has been declining for a decade; a reversal changes the PEO and tough-comp value proposition.
- Small-account pricing goes negativeThe squeeze reaches independent agencies; the 'commission back on your P&L' conversation gets easier.
- E&S property premium grows againStandard markets are retreating; hard-to-place property programs regain their floor.
- AM Best moves auto, GL or D&O to StableRating-agency confirmation that a lane has turned; lagging but decisive.
- Litigation-funding or tort reform passes in key statesChanges the loss-cost story in auto and umbrella state by state; re-read lanes for those states.