An illustration is the single document a buyer remembers. It is also the document the carrier, the producer, and the state insurance department will litigate over twenty years later. For indexed universal life (IUL), variable annuities (VA), registered index-linked annuities (RILA), and guaranteed lifetime withdrawal benefit (GLWB) products, the illustration is no longer a static 30-page PDF — it is a live projection engine that must reconcile actuarial inputs, regulatory caps, hedge costs, and agent assumptions in under two seconds. Carriers still running WinFlex desktop installs or batch overnight illustration servers are losing cases to competitors whose agents can re-quote at the kitchen table on an iPad.
The economics are material. LIMRA's 2024 retail life sales tracking shows IUL accounting for roughly 24% of individual life premium and indexed/RILA annuities exceeding $120B in 2024 sales. Each of those sales requires at least one compliant illustration, frequently three to five iterations per case. A carrier writing 80,000 IUL policies annually is generating well over 300,000 illustration runs, every one of which is a potential market-conduct exhibit.
The Regulatory Architecture Illustrations Must Honor
The substrate is NAIC Model Regulation 582 (Life Insurance Illustrations Model Regulation), adopted in some form by 48 states. It requires a basic illustration, a numeric summary, signed acknowledgments, and a clear separation between guaranteed and non-guaranteed elements. The illustration actuary must certify annually that the disciplined current scale being illustrated is supportable — meaning carriers cannot show a 6.5% IUL credit if their hedge budget only supports 5.8%.
For IUL specifically, Actuarial Guideline 49 (2015), AG 49-A (November 2020), and AG 49-B (effective May 1, 2023) progressively tightened how indexed crediting and policy loan arbitrage can be illustrated. AG 49-B caps the illustrated benefit from index multipliers and bonuses at 0.50% above the option-budget-derived maximum rate, and prohibits illustrating positive loan arbitrage that exceeds 50 basis points. Engines built before May 2023 that still need software switches to enforce these caps are a compliance liability.
| Constraint | AG 49 (2015) | AG 49-A (2020) | AG 49-B (2023) |
|---|---|---|---|
| Max illustrated index rate | Based on 25-yr historical lookback of benchmark index | Same, but tighter benchmark definition | Same — annual certification required |
| Multiplier/bonus benefit | Not directly limited | Limited to non-multiplier max rate | Capped at 50 bps above option-budget max |
| Loan arbitrage spread | Up to 100 bps | Capped at 50 bps | Capped at 50 bps, applies to all loan types |
| Alternative scale | Not required | Required at lower of policy loan rate | Required and prominently displayed |
Variable annuity illustrations operate under different regimes. SEC Rule 498A (the variable annuity summary prospectus, effective July 2022) and FINRA Rule 2211 govern projections; Monte Carlo or hypothetical scenario illustrations must disclose assumptions. For RILA products, the SEC's May 2024 final rule on registered index-linked annuities (adopting Form N-4 for RILAs by May 1, 2026) mandates standardized illustration of crediting strategies including buffers, floors, caps, and participation rates under specified return scenarios. Carriers selling RILA today on shared VA chassis must rebuild illustration logic before the 2026 compliance date.
Anatomy of a Modern Illustration Engine
A current-generation illustration platform separates four concerns that legacy WinFlex-style monoliths fused together: product configuration, projection calculation, compliance enforcement, and presentation. Each runs as a discrete service with its own deployment cadence.
The product configuration service holds the policy form: charges (cost of insurance, premium loads, M&E, administrative), credited rate scales, cap/floor/participation tables, rider mechanics, and loan rules. Carriers like Pacific Life and Lincoln Financial have moved this from MG-ALFA or Prophet exports into a JSON-schema product catalog refreshed nightly, allowing actuarial to publish a new IUL cap structure to production illustrations in 48 hours rather than a six-week IT release cycle.
The projection service is the actuarial math: monthly or annual policy-year roll-forward of account value, surrender value, death benefit, and rider values across guaranteed, current, and midpoint scales. For VA and RILA, this expands to 1,000–10,000 Monte Carlo paths with per-path crediting logic. Performance targets are non-trivial: an agent re-running an IUL illustration on an iPad over a coffee-shop LTE connection expects a full 121-year projection (issue age 0 to age 121) in under 1.5 seconds. Carriers hitting this number have moved from synchronous SOAP calls to gRPC or REST with computed-column caching and have rewritten projection cores in Rust, Go, or vectorized C++.
The compliance service enforces AG 49-B caps, Model Reg 582 disclosures, state-specific variations (New York Reg 187 best interest language, California Insurance Code 10509.910 for senior buyer disclosures), and producer authority checks (state appointment, product training, AML, Reg BI for variable). When an agent in Florida tries to illustrate a non-approved alternate index strategy for a 78-year-old buyer, this service blocks the run before PDF generation rather than relying on home-office review to catch it later.
The presentation layer renders PDFs, interactive web views, and increasingly conversational summaries. Ensight and Covr have demonstrated that replacing a 47-page PDF with an interactive scenario explorer reduces case-not-taken rates by 18-25% in pilot books, because clients comprehend the trade-off between premium and income at age 70 visually rather than scanning columns.
Dynamic Projections — Where the Math Gets Hard
Static cash-value tables are the easy case. Modern illustrations must handle four dynamic projection problems that legacy tools handle poorly or not at all.
First, GLWB and GMWB rider mechanics on deferred annuities. The benefit base typically rolls up at 5-8% simple or compound for a deferral period, ratchets to high anniversary values, and converts at issue-age-based withdrawal percentages (e.g., 5.0% at age 65, 5.5% at age 70). Illustrating lifetime income requires projecting account value depletion separately from the protected benefit base — and showing what happens when account value reaches zero but lifetime payments continue. Engines must handle bonus base versions, joint-life reductions (typically 50 bps on the withdrawal percentage), and post-2024 buffer-style RILA income riders introduced by Equitable, Allianz, and Brighthouse.
Second, IUL distribution illustrations using policy loans. Clients buying IUL for supplemental retirement income want to see tax-free loan streams from age 65 to 90. The engine must solve for the maximum sustainable loan that keeps the policy in force to age 121 under both the current scale and the alternative scale required by AG 49-B. This is a non-linear solver — typically a bisection or Newton-Raphson over the loan amount — and must converge on every keystroke when the agent adjusts premium or solve year.
Third, Monte Carlo projection for variable products. A typical VA illustration shows 10th, 50th, and 90th percentile outcomes plus a probability-of-success metric for income riders. Generating 5,000 paths × 30 years × 12 months × dozens of sub-accounts in under two seconds requires GPU offload or extensive pre-computation. Equisoft's WealthElements and Zinnia's Annuity Net have published p95 latencies of 1.1–1.7 seconds for 5,000-path VA illustrations on standard cloud SKUs.
Fourth, dynamic in-force re-illustration. When a policyholder calls to inquire about a loan or premium reduction, the engine must pull current account value from the policy admin system covered in Article 2, recompute future projections under the proposed change, and produce a re-projection that complies with the same AG 49-B caps as new business. This is where in-force management discussed in Article 5 intersects with illustration systems most painfully — most carriers cannot do this without an overnight batch.
Integration: Why Standalone Illustrations Are Dying
An illustration that requires re-keying into the eApp wastes 12-18 minutes per case and introduces a 3-5% NIGO rate from data mismatches (face amount typed as $250,000 in illustration, $25,000 in eApp). Modern stacks treat the illustration as the authoritative quote artifact: the signed PDF carries a quote ID that the eApp consumes, the underwriting workbench references when applying the predictive mortality models discussed in Article 1, and the policy administration system uses to instantiate the contract at issue.
iPipeline's iGO + Resonant pairing, FireLight from Insurance Technologies, and Ensight all expose illustration-to-application APIs that pass 80-150 fields directly into the eApp. Carriers integrating this way (Nationwide, Prudential, Symetra) report NIGO reductions of 30-45% on the illustration-derived fields. The integration also enables suitability pre-checks — running the illustration through the suitability rules engine before the agent ever generates the PDF, surfacing fixed/indexed annuity replacement issues, senior-buyer concerns, or product-state-license mismatches at quote time rather than home-office review.
The Vendor Landscape
The market segments into three groups. Established illustration specialists — iPipeline (acquired by Roper in 2019), Insurance Technologies (FireLight, ForeSight), and InsMark — dominate the agent-facing illustration tooling, with iPipeline's Resonant claiming over 100 carriers and FireLight reporting more than 350,000 monthly users across carrier and distributor deployments. These vendors have invested heavily in cloud rebuilds since 2021 but still carry legacy code paths.
A second cohort focuses on the modernization of the buyer experience: Ensight (acquired Proformex in 2023, raised a Series C in 2022), Covr Financial Technologies, and Bestow's enterprise platform reframe the illustration as an interactive scenario explorer rather than a PDF. Ensight reports producer adoption metrics showing 2.4x more illustrations run per case versus legacy PDF workflows, because agents experiment with funding scenarios rather than committing to one.
A third group — Equisoft, Zinnia (formed from the Sammons Financial/Eldridge combination of SE2 and Policygenius B2B in 2022), and Andesa — provides integrated stacks where illustration sits alongside policy administration, claims, and reinsurance, the last covered in Article 6. The trade-off is real: integrated platforms reduce data-integration cost by 40-60% versus best-of-breed, but illustration-specific innovation tends to lag specialist vendors by 12-24 months.
If your illustration cannot recompute on a tablet in under two seconds and flow its quote into the eApp without a single re-keyed field, you are subsidizing your competitors' producers.
— Head of Distribution Technology, top-10 US life carrier
An Implementation Playbook
Modernization programs that succeed share a sequence. Carriers that try to replace the engine, the UI, and the integrations simultaneously routinely overrun by 18-30 months and $15-40M.
Externalize product configuration from the illustration tool into a governed catalog (typically a versioned Git-backed JSON/YAML repository with actuarial sign-off workflow). Eliminate Excel-based rate tables. Establish the disciplined current scale data model that supports AG 49-B certification.
Rebuild the projection core as an API-first service. Vectorize the math, instrument every calculation for audit, expose deterministic seeds for Monte Carlo. Run shadow-mode against the legacy engine for 90 days, target <0.01% variance on identical inputs.
Implement AG 49-B caps, state rule packs, producer authority checks, and the signed quote-ID artifact. Integrate with eApp, suitability, and underwriting via the quote ID, eliminating re-keyed fields.
Replace static PDF-first UI with interactive scenario explorer. Add tablet/mobile-optimized flows. Launch in-force re-illustration capability tied to live PAS data.
Sunset legacy WinFlex/desktop installs. Add Monte Carlo GPU offload if VA/RILA volumes warrant. Publish performance and quality SLOs externally to distribution partners.
Two cost benchmarks help size the program. For a mid-sized carrier ($500M-$2B annual premium across life and annuity), full modernization runs $8-18M including vendor licenses and internal implementation, with payback typically in 24-36 months from NIGO reduction, faster cycle time, and avoided compliance findings. For top-10 carriers with proprietary engines, the range is $25-60M, dominated by the cost of decommissioning bespoke COBOL or APL projection cores that have decades of accumulated product variants encoded.
Metrics That Tell You The Program Is Working
CIOs and COOs should track six numbers monthly. First, illustration latency at p50, p95, and p99 — a p99 above 4 seconds on agent tablets predicts case abandonment. Second, illustrations-per-case — values rising from 1.2 toward 3-4 indicate agents are using scenario exploration rather than producing one PDF and hoping. Third, illustration-to-application conversion rate — best-in-class is 35-45% for life, 50-60% for annuities. Fourth, NIGO rate on illustration-derived fields, target sub-1%. Fifth, time from actuarial scale change to live in production — best-in-class is 24-72 hours, legacy environments average 6-10 weeks. Sixth, illustration-related complaints per 10,000 policies — a leading indicator of market-conduct exposure, typically 0.4-1.2 for well-run carriers.
These metrics also feed back into the data warehouse described in Article 10. Comparing illustrated lapse and persistency assumptions against actual experience, by product, distribution channel, and producer cohort, is one of the highest-value uses of integrated illustration and policy data — it informs both pricing and the next round of disciplined current scale certifications.
The illustration is no longer a sales aid bolted onto the side of operations. It is the contract-shaped artifact that ties underwriting, suitability, administration, and in-force servicing together. Carriers that treat it as such — investing in API-first projection services, governed actuarial catalogs, and integrated producer experiences — will outsell those who treat illustrations as a PDF generation problem, and they will do so with materially lower compliance risk.