NewSIInE — New Standard In Insurance Economy

Schedule A — The thesis

The cost base of European P&C is architectural. We don't have the architecture.

European personal-lines and SME carriers run administrative expense at 12–18% of premium — barely moved in two decades, because it lives in legacy systems and layered back offices that cannot be restructured without breaking the company. An entrant built after 2024, where AI agents perform underwriting triage, policy administration and claims, operates the same book at a fraction of that. The gap is margin, price, or commission — dosed by market.

Incumbent composite carriers

12–18%

Digital-era insurers

9–12%

NewSIInE at scale

4.8%

Administrative expense as % of GWP, excluding acquisition commission. NewSIInE figure is the plan's 2034 base case at ~€537m GWP and ~95 FTE — roughly €5.7m premium per employee, ~6× even well-run incumbents (Warta, Poland's #2 insurer: €2.7bn revenue with ~3,000 staff ≈ €0.9m/FTE).

GWP per employee — with and without the exclusive sales network.

employees as reported including the exclusive / tied sales network

UNIQA Group

€0.6m

€0.4m

VIG (Vienna Insurance Group)

€0.5m

€0.5m*

Generali Group

€1.1m

€0.4m

Allianz Group

€1.1m

n/d

Warta (PL #2 insurer)

€0.9m

Talanx Group (incl. Hannover Re)

€1.7m

NewSIInE plan (2029 · 2031 · 2034)

'29 €0.9m

'31 €2.8m

'34 €5.7m

How to read this — and why the second bar matters. Every insurer's "GWP per employee" excludes the self-employed people who actually sell the policies. Count the full human machine and the picture changes: Generali discloses 88,000 employees and 163,000 agents — its €1.1m collapses to ~€0.4m; UNIQA reports "21,000 employees and exclusive sales partners" — ~€0.6m on employees becomes ~€0.4m all-in. *VIG's ~30,000 already includes its employed sales force (self-employed agents excluded, count undisclosed) — its true all-in figure is below €0.5m. Warta and Talanx barely move: they have no exclusive retail network to hide. NewSIInE's figure is identical on both countings — by design (distribution is independent brokers and multiagents, paid transparently by commission, never inside the FTE line), so its bars show the plan trajectory instead: €0.9m in 2029, the second selling year — past UNIQA, VIG and at Warta's level; €2.8m in breakeven-year 2031 — past every incumbent including Talanx; €5.7m in 2034. With networks counted, the retail incumbents converge at €0.4–0.5m per person — the honest gap at plan-end is not 6×, it is 12–14×.

And the gap is widening, not closing — incumbents’ own AI push raises their costs first.

+ CAPEX

European incumbents are committing hundreds of millions of euros — at market level — to AI transformation programs: new investment that lands on the cost base years before any benefit does.

+ PEREX (temporary)

Transformation means running old and new processes in parallel: change teams, retraining, consultants — personnel expense rises first, and incumbent behavior shows headcount reduction is the step deferred longest.

+ WRITE-OFFS

The previous wave of IT investment — core migrations, digital front-ends — is still inside its depreciation period, stacking amortization on top of the new AI spend.

Net effect: best case, incumbent expense ratios stay flat and fall only later. Mismanaged — including unwisely designed AI runtime (token/inference) costs that scale with volume instead of falling — they can even rise. Either way, the gap holds or widens precisely during the window in which NewSIInE builds and scales with no legacy, no parallel run, and no depreciation overhang.

Schedule B — Operating model · this is Lucy's section Lucy

Lucy runs the flow. Licensed humans own the judgment.

Every core process is designed as agent + human exception desk — not a human process with AI features. The claims agent descends from a system our founding CEO already shipped in production inside a major European insurance group: this is a rebuild with full ownership, not a first attempt.

Process	Lucy does	Humans keep
Underwriting	Reads broker and multiagent submissions, registries and geodata; prices standard risk instantly; triages referrals with a written rationale, same day	Appetite, tariff governance, genuine edge cases
Claims	FNOL in any channel, coverage check, photo-based damage estimation, fraud scoring, straight-through settlement for standard claims in minutes — rated 4.9/5 by claimants in NIQA production, above the 4.5 human ceiling	Complex, bodily-injury and litigated claims; sampling audit of every Lucy decision
Policy admin	Endorsements, renewals, documents, reconciliation — fully straight-through	Exception queue only
Partner service	Multilingual broker & multiagent assistant: product answers, quotes, claim status, commission queries, 24/7	Relationships, pools, key accounts
Reporting	Real-time bordereaux via API; IFRS 17 / Solvency II data pipeline designed in from day one	Sign-off and regulatory dialogue

Governance by design: every AI decision is logged, explainable and reversible, with a human appeal path — built to high-risk-grade standards under the EU AI Act and DORA even where not formally required. For supervisors and capacity partners, the audit trail is the product.

Meet Lucy — the insurance machine, in person

Lucy, the AI agent

“Chief Operating Officer”

One AI agent, everywhere at once — with the whole company's knowledge behind her: every product and wording, every process, every country's regulation, the live portfolio. The operating rule that makes this real: nothing goes live until Lucy knows it — every new product, market and process is onboarded by teaching Lucy first, so her knowledge can never drift from production, because it is production.

Three faces, one core. For colleagues, Lucy is the operating manual that answers and executes — every internal workflow starts as a conversation with her. For brokers and multiagents, she is the underwriting and service desk that never sleeps: quote, clarify, bind, endorse, FNOL — in the intermediary's language, finished on the spot. For customers, she is the empathy the numbers already proved — the 4.9/5 claims experience, personified.

And Lucy is the brand: incumbents show org charts, NewSIInE shows Lucy — the visible, nameable proof of a new-generation AI-native company. Customers won't say “I contacted my insurer” — they'll say “Lucy sorted it.” The guardrail behind the face: Lucy acts within appetite, corridors and full POG/IDD governance, licensed humans own the judgment calls, and every decision is logged with rationale and appeal path.

Schedule B2 — The experience moat (CX) · Lucy again Lucy

Finished on the spot. The CX moat is Lucy.

Human-run client service has a structural ceiling: opening hours, queues, capacity peaks, handoffs, “someone will call you back.” A well-designed agent has none of these — and the biggest difference is not availability, it is completion. Experience is the harder half of the differentiation to copy: an incumbent can cut cost by cutting people, but it cannot give a human team 24/7 on-the-spot resolution.

Moment	The human-run industry	Lucy
Claims	FNOL is a report that starts a waiting loop: file, wait, ping-pong for documents, wait for a decision, wait for a call	FNOL is the claim management: Lucy clarifies circumstances in conversation, documents and photos are uploaded and verified in the same session, and if everything clears — decision and payout initiated on the spot. With a well-designed AI agent this level of CX is reachable in production — resolution in hours, satisfaction above human-run service ceilings
Client service	Policy modification means a form, a queue and a callback window	Address change, cover adjustment, endorsement, documents — done in-session, 24/7, any channel, any language
Sales	Quote today, documents tomorrow, bind next week	Quote → clarify → bind in one sitting — for the customer and for the broker's front office alike

Why it compounds: in retail P&C the claim is the product moment. Satisfied claimants renew, renewals are the cheapest premium in insurance — so on-the-spot completion converts directly into loyalty and lower churn, a better-seasoned book, and a falling blended acquisition cost. The experience moat funds the cost moat.

Schedule C — Distribution & products

Brokers move European P&C. We built the carrier brokers actually want.

No direct-to-consumer acquisition spend — the intermediary owns the customer and the acquisition cost; NewSIInE owns product, risk and a radically cheaper machine. Brokers and multiagents don't get a portal — they get Lucy: same-day quotes, in-session endorsements and claims decisions their clients rate 4.9/5. Launch products are chosen for one property: short-tail, high-frequency, AI-automatable.

Travel

The flywheel starter

Document-based claims — 80–90% straight-through automatable, settled while the customer is still abroad. Low-stakes first product for a broker to try; fast volume, fast data.

SME package

The loyalty lock-in

Same-day bindable quotes for trades, offices and retail where incumbents take weeks. The single strongest wedge in broker & multiagent distribution — live from day one on a disciplined trade whitelist.

Standalone casco

The premium engine

Own-damage motor without MTPL: pure property risk, photo-based AI estimation, repair-network steering. Fits CEE's unbundled buying culture — a wedge incumbents can't copy without cannibalizing bundles.

Plus household, liability and accident at launch; cyber and PI follow. Motor liability launches in 2029, once the machine is proven — ring-fenced on dedicated capacity, price-follower, with bodily-injury claims retained by the risk carrier and pre-agreed kill-criteria. The switching engine of CEE aggregator markets, entered on our terms.

Schedule D — Roadmap

MGA for the whole plan. Capital-light by design. Control preserved throughout.

The capital-light MGA phase proves loss ratios, the cost machine and broker & multiagent traction on fronting capacity through the whole horizon — commission, claims-handling and profit-commission income with zero underwriting capital. A carrier conversion after the successful MGA stage (post-2034) stays a genuine option — outside this plan's numbers, kept open by the regulator-grade platform.

2027 — preparation year, zero income by design

The honest build year: platform and Lucy v1 (the claims & service agent), MGA registration and EU passporting, fronting and reinsurance agreements signed, HU/CZ/SK integrations tested. Growth starts at half-time: from H2 2027 a Chief Growth Officer and three market channel leads (HU/CZ/SK) contract brokers & multiagents into launch — the CHERRISK lesson: intermediary contracting takes 12–18 months, so the 150-producer pipeline is signed before launch. Fully funded within seed.

2028 — first premium

Hungary + Czechia + Slovakia live with the full launch set (CZ and SK from H2), plus a 5% D2C organic layer on every book from each country's second selling year, with zero direct-marketing spend — three countries, one platform, proven within the launch year; MTPL launches at year-end in HU, CZ and SK on ring-fenced capacity; 120 producing intermediaries; ~€6m GWP part-year ramp. Austria deliberately skipped — a tied-agent market is the wrong terrain for a broker-first model.

2029–2030 — two new markets every year

2029: Poland full-force via Unilink-class multiagents (MTPL from launch) + the Netherlands via the volmacht model (WA motor follows in 2031 through the same rails). 2030: Germany via broker pools + Sweden via digital brokers and affinity rails — both DE entries motor-free by design. ~€105m GWP by 2030; fronting panel extended on two years of audited loss ratios.

2031 — breakeven + Romania & Italy

Group EBITDA breakeven at ~€193m GWP; Romania (motor-free) and Italy (aggregator + RUI broker channel) enter as markets eight and nine; optional ~€10m Series B growth round with dual-class control terms — no carrier capital within this plan.

2032–2034 — nine countries compounding

Nine markets deepening simultaneously — still on ~€36m of total equity ever raised; €537m GWP, 4.8% expense ratio, ~95 people.

HU '28 CZ '28 SK '28 PL '29 NL '29 DE '30 SE '30 RO '31 IT '31

Schedule E — For risk carriers & reinsurers

The cleanest delegated book you will see this decade.

Fronting partners and reinsurers carry the risk in the MGA phase — so the proposition to capacity is engineered first, not last.

Pricing & risk selection — the region's best desk, repricing daily

The pricing organization is built on the deepest retail-pricing expertise in the region — the founder designed and rolled out the pricing-maturity methodology across every UNIQA country, and a pan-EU chief pricing actuary leads a senior desk. On top sit online, AI-driven pricing models: GLM foundations with machine-learning ensembles, real-time data enrichment at quote (vehicle, property, geo and claims signals), fraud scoring before bind, and portfolio-drift monitoring per product, per country.

The tariff engine supports daily price changes — reacting to competitor moves, claims emergence and mix shifts within carrier-agreed corridors and full POG/IDD governance — while incumbents reprice on quarterly-to-annual tariff cycles. Risk selection is codified: appetite rules, referral triggers and pre-agreed kill-criteria per segment, with the fronting carrier and reinsurer holding co-underwriting veto. Pricing governance at carrier grade, visible in real time — this is what keeps the delegated book clean.

✓

Short-tail core; MTPL ring-fenced

Travel, SME property/liability, casco, home run off in months. MTPL (from 2029) sits on dedicated capacity with separate reinsurance, capped share of book, and bodily-injury claims retained by the carrier's specialist desk — never blended into the core book.

✓

Real-time bordereaux via API

Not monthly spreadsheets: live premium, exposure and claims data, per risk, with accumulation views. You see the book as we see it.

✓

Delegated claims you can audit

Every agent decision logged with rationale, confidence and appeal path; sampling audit built for your reviewers. Claims leakage and fraud analytics protect your loss ratio, not just our fees.

✓

Aligned economics

Profit-commission structures with meaningful skin in the game; conservative wordings; pre-agreed kill-criteria per product and market. We are paid when you make money.

✓

Underwriting discipline with a track record

Appetite set by a leadership team that has run multi-country retail P&C portfolios through full pricing cycles — not learned discipline on partner capital.

✓

A long-term partnership through the whole plan

NewSIInE stays MGA for the full horizon — the book stays on partner paper through 2034, growing to €537m; any post-2034 carrier conversion would be built with, not against, the panel; minority equity participation open to strategic capacity partners. Capacity today is a position in the platform tomorrow.

Schedule F — For investors

Insurance economics, software-shaped.

The MGA phase earns commission and profit-commission on other balance sheets — capital-light, recurring, and compounding with every broker and multiagent integration. The carrier phase converts a proven book into underwriting margin behind reinsurance.

Plan base case	2027	2028	2030	2032	2034
GWP	€0 (prep)	€6m	€105m	€295m	€537m
Countries live	0	3	7	9	9
Producing intermediaries	—	120	1,200	3,200	6,000
EBITDA	–€4.6m	–€6.6m	–€1.0m	€16.8m	€43.0m
Admin expense ratio	—	ramp	13.6%	7.1%	4.8%
GWP per FTE	—	€0.1m	€1.7m	€3.8m	€5.7m

Funding: €8m seed at €29m post (2027, funds the full preparation year on the contracted Intuitech build cost), €15–20m Series A (2028, at first-traction metrics), an optional ~€10m Series B growth round (2031) — no carrier solvency capital within this plan. Dual-class structure keeps founder control through every round — deliberately: this only works if the operator who has built it three times keeps running it. Full model, sensitivities and assumptions available in the data room.

What is the plan worth? The valuation trajectory, year by year.

In investment banking, an MGA is priced as an insurance distribution platform — the hottest category in FIG M&A: platform deals averaged 14.3× upfront EBITDA in 2025 (up to 18.4× with earnout), public brokers trade at 16–18×, and $5.6bn of equity flowed into distribution platforms in 2025 (+206% YoY). Applying that framework to the base case:

€29m

'27

€55m

'28

€85m

'29

€115m

'30

€160m

'31

€252m

'32

€425m

'33

€602m

'34

2027–2031 · Round-anchoredSeed, Series A and the optional Series B growth round price the milestones: passporting, first premium, multi-country proof, breakeven.

2032 · Revenue-multiple phasePlatform revenue multiples take over as EBITDA turns and scales.

2033–2034 · EBITDA-multiple phase14–14.5× platform-tier EBITDA — below the 16–18× of public brokers. €36m invested → ~€602m: 16.7×, with founder control intact.

Indicative triangulation on the base case, computed on the Valuation sheet of the financial model (every multiple editable); not a fairness opinion. A strategic sale or IPO could price above the M&A corridor.

Who owns what, year by year — control by design.

Three rounds, three dilution events, and nothing in between. Founder economics settle at 39.7% while the 10:1 dual-class keeps 86.8% of votes at exit — the operator stays in charge, which is the investors’ execution guarantee, not their cost.

'27

F 64.0%

'28

F 43.1%

'29–'30

F 43.1%

'31

F 39.7%

'32–'34

F 39.7%

Founder (10:1 voting → 86.8% of votes at exit) Investors — 55.3% of a ~€602m enterprise on €36m invested ESOP

Founder value at indicative EV: €16m ('27) → €50m ('30) → €52m ('31) → €239m by 2034 — rising every single year; the MGA-only structure has no dilution dip. Full mechanics on the CapTable sheet of the model.

Schedule G — Why this team & technology partner

Three proofs, already delivered at the top of the industry.

Few founders enter this market with the full stack: our founding CEO built a cross-border digital MGA from zero, shipped production claims AI, and ran multi-country retail P&L inside one of Europe's leading insurance groups. NewSIInE stands on a decade of firsts, delivered in the real industry.

Krisztián Kurtisz

Founder & CEO

ex-CEO of UNIQA Hungary · Founder-CEO of CHERRISK · creator of the NIQA claims AI · Group Head of Retail, UNIQA CEE

Proof 01

Built a cross-border digital MGA from zero

Founded and led CHERRISK, UNIQA's pioneering greenfield digital insurance venture: product, platform, team and regulatory approvals built from zero. Within four years, one platform served four countries — Hungary and Germany under the CHERRISK brand, Slovakia and Czechia white-labeled on the same technology — among the first true multi-country digital insurance operations in Europe, and living proof that the “new market as a configuration exercise” playbook works. That first-hand mastery of cross-border scaling and channel economics directly shaped NewSIInE's broker-first, multi-market strategy.

Proof 02

Shipped production claims AI in a regulated environment

Built NIQA, an AI claims agent handling triage, coverage interpretation, fraud signals and customer dialogue — live, in production, in a regulated insurance environment, years before the market caught up. And it keeps compounding: the current development target is a −48% reduction in claims FTE while claim volumes grow over the coming years. That is exactly the curve NewSIInE's economics are built on — cost per claim falling as the book expands — delivered not as a projection but as the continuing trajectory of a system the CEO already shipped.

Proof 03

Ran multi-country retail P&C at scale

As CEO of UNIQA Hungary and Group Head of Retail across CEE markets: pricing cycles, portfolio steering, broker and multiagent economics, country by country, at one of the region's strongest retail franchises. Designed and implemented a pricing-maturity assessment methodology rolled out across all UNIQA countries — a group-wide standard for measuring and systematically lifting pricing capability, market by market. For NewSIInE this is the loss-ratio insurance policy: the same structured discipline governs every market entry, so the machine prices as sharply in its seventh country as in its first.

Partnering company: Intuitech — the agents are already in production.

NewSIInE does not build its agent platform alone from a blank page. Founding technology partner Intuitech builds fully autonomous AI agents for banks and insurers — and is the engineering home of the NIQA claims-agent lineage our CEO shipped. What most insurtechs promise on slides, this partnership has already run in production, in regulated finance, in multiple countries.

Máté Jendrolovics Founder & CEO, Intuitech

The engineer-founder behind the agent platform: builder of fully autonomous agents running in production at banks and insurers — from end-to-end lending operations (onboarding, KYC, credit memos) to the NIQA claims lineage. 30+ implementations, ISO 27001, EU AI Act aligned — the safe pair of hands carrying the 2027 build year.

LinkedIn →

✓

Autonomous insurance claims — live, not demo

Intuitech operates AI-native FNOL and fully autonomous claim handling for the large majority of standard claims, from first notice to payout, with contractually guaranteed extraction accuracy. NewSIInE’s claims engine is a configuration of proven capability, not a science project.

✓

Bank-grade agent experience

Their agents run end-to-end retail and commercial lending operations at banks — onboarding, KYC, income and collateral verification, credit memos, disbursement. Lending automation is harder than most insurance workflows; that discipline transfers directly to underwriting and policy admin.

✓

Scale and track record

30+ agentic operations implementations, 120+ product capabilities in production, a 260+ strong team across 15+ countries, and reference clients across CEE banking and insurance — recognized by Deloitte and Mastercard programs.

✓

Regulator-ready by construction

ISO 27001 certified, GDPR and EU AI Act aligned, with a proven ~3-month path from design to production. For fronting carriers and supervisors, this means the 2027 build year runs on an audited delivery framework — not a startup’s first attempt.

Why this matters for the plan: the preparation year’s biggest execution risk — getting the claims and policy agents to production quality before first premium — is carried by a team that has done it repeatedly, while NewSIInE keeps full ownership of insurance IP, data, and the underwriting judgment layer.

Schedule C1 — Capacity UNIQA

Our first risk-carrier partner: UNIQA.

The machine writes on the paper of an A-rated international insurance group from day one — and not a stranger's paper: the group where our founder was country CEO and Group Head of Retail CEE.

✓

A-rated security behind every policy

UNIQA Group: €8.4bn GWP, 18 million customers across CEE — brokers, multiagents and customers see a household-name carrier on the policy schedule, not a startup balance sheet.

✓

A partner that knows the founder's execution firsthand

Krisztián ran UNIQA Hungary as CEO and rolled out the group's pricing-maturity methodology across every UNIQA country. This is capacity granted on evidence, not on a pitch — the fastest possible diligence any fronting partner could do.

✓

Footprint matched to the launch map

UNIQA operates across the exact CEE core where NewSIInE launches first — HU, CZ, SK — one partner, one framework agreement, three markets live in year one, with room to extend along the nine-country plan.

✓

Carrier-grade governance from the first policy

Co-underwriting veto, appetite corridors, real-time bordereaux, delegated claims with a sampling audit built for the carrier's own reviewers — the cleanest delegated book, visible live. When the incumbent benchmark in our own exhibits chooses to back the machine, that is the strongest external validation the model can get.

Why it matters: fronting capacity is the MGA model's single existential dependency — and it is secured with the counterparty best placed in Europe to judge this exact team, on the exact home markets, before the first policy is sold.

The first real AI-native insurance machine in Europe — an expense ratio and a customer experience incumbents cannot follow.