Rating Agency & Exam Response: every questionnaire drafted from last year’s answer and today’s filed figures, checked against what each agency was told
Each request drafted from last year's answer and today's numbers, owner-signed.
See one case, screen by screen ↓A response solution for insurers that drafts rating agency and exam answers from past answers. Six agents split each questionnaire or request list into its questions, reuse the last approved answer with this year’s figures refreshed from the filed statements, check every figure against what each audience was already told and route each answer to its owner. Owners approve paragraph by paragraph and the CFO signs.
How to choose AI solutions for insurance →
Updated
Rating agency questionnaires, regulator exam requests and auditor support lists ask much the same questions every year — capital, reserves, pricing, reinsurance, ERM — in new formats and with new numbers. The liaison opens last year’s Word file, chases contributors by email and rebuilds each answer by hand, while the statutory reporting team pulls the same figures out of the Annual Statement, Schedule P and the RBC report one more time.
The harder problem is consistency. The same metric on two bases — renewal rate change with or without an exited program, reserve development before and after the year-end true-up — reads to an analyst as two stories. Nobody can easily check a new draft against what AM Best was told in May, what Moody’s heard in July and what the examiners and the auditor have on file. Each mismatch that slips through becomes a follow-up question, a management-meeting slot and some credibility.
Estimated split for one rating agency questionnaire, in working days, by hand and with the solution. The owners’ review and the CFO’s signature stay with people.
Six specialist agents draft every answer and check it against what each agency was told; owners approve and the CFO signs.
Calderwood Mutual Insurance Group, a regional, independent-agency writer in eight Midwest states in the working solution, receives S&P Global Ratings’ 2026 annual review questionnaire on 28 September: 48 questions in 9 sections, due Friday 30 October. Here is what happens next, screen by screen.
Claire sees the year at a glance: AM Best’s Supplemental Rating Questionnaire sent 14 May, Moody’s mid-year update sent 22 July, the Ashby Lorne LLP interim audit at 120 of 141 items closed, the S&P review in drafting and the Ohio Department of Insurance full-scope exam with its first batch due 2 November. Eight requests and meetings, 559 questions and items in 2026.
The agents split the S&P questionnaire into its 48 questions, keeping S&P’s numbering, and matched 43 to an approved answer already sent to AM Best, Moody’s or last year’s S&P. Every number in the reused language was replaced from its filed source. Question 3.5 is drafted from Moody’s mid-year update Q5, 87% the same question: the first-half 2026 combined ratio now reads 99.4% against a plan of 98.0%.
Hover any figure: “99.4% · Quarterly statement, 30 Jun 2026, p.4 — Last answer said 101.2%.”
Question 4.3 asks for commercial auto rate adequacy. The draft says renewal rate increases of 13.8% in 2025, taken from the final FY2025 rate monitor — all commercial auto, including the Harbor Transport program. In May the company told AM Best 14.2%, excluding that program, signed by Victor Salgado, and told Moody’s the same in July. The guard quotes both statements side by side and blocks approval of the paragraph until someone decides.
“Told AM Best 14.2% in May — this draft says 13.8% for the same line and year.”
The guard offers three ways out: use 14.2% and the same basis note, keep 13.8% and explain the basis, or ask Gwen Abernathy, who owns commercial auto answers. Claire takes the recommended fix. The paragraph now says 14.2% “excluding the Harbor Transport program exited in 2025” and goes back to Gwen to approve.
“4.3 updated · the figure ledger now shows one number for every audience.”
Question 5.2 says $21.0M of favourable prior-year development in calendar 2025, refreshed from the Q4 2025 board reserve paper — written before the year-end true-up. Schedule P and the Moody’s update say $18.4M. The guard recommends the filed figure. The other ways out are to use $18.4M and mention the board figure, or to ask Tom Lindqvist, who owns reserve answers.
“Use $18.4M from Schedule P — Recommended — the filed figure.”
The consistency report lays 12 key figures across six audiences in 18 months — AM Best, Moody’s, S&P, the Ohio exam, the auditor and the board — after comparing 236 sent answers. Net written premium of $2,412M and an RBC ratio of 412% agree everywhere. The property catastrophe retention is not a contradiction: $40M before the 1 July renewal, $50M now. The guard asks for one sentence saying so.
The guard never resolves a flag itself — it quotes both statements and a person decides.
Behind the report sit the filed figures each answer is allowed to use, each tied to its statement or report, page and line. Prior-year reserve development opens on Schedule P, Part 2 summary, p.92: one-year development of (18,398) in column 11, used in 6 answers across 6 audiences.
Request list 1 from the Ohio Department of Insurance arrives from the examiner-in-charge: 212 items on 6 tabs. The agents keep the examiners’ numbering, match 141 items to a 2021 exam response, an agency answer or the board archive, mark 48 that need 2021–2025 figures and leave 23 new ones. Item 3-07 asks for the same development as S&P Q5.2, so both are linked to say $18.4M; 18 items have no topic owner yet and wait for Claire.
The owner router assigns every question by topic in Settings: Gwen for commercial auto and pricing, Tom for reserves, Rohan for filed statements and RBC, Benedict Osei for investments and capital, Maren Ostby for reinsurance and catastrophe, Leah Brandt for risk management and ORSA. Owners sign off by Fri 16 Oct, Claire reviews the whole submission on Wed 21 Oct.
The pack assembler checks the list before the CFO signs: all 48 answers approved by their owners, a clean consistency report, 124 of 124 figures tied to a filed source, 7 exhibits attached, the transmittal letter drafted and section 8 co-signed by the Chief Risk Officer. Victor signs and releases to S&P Global Ratings, and every signed answer is filed in the answer library for next year.
“Released to S&P Global Ratings by Victor Salgado. 48 answers, 7 exhibits and the consistency report.”
The same S&P review, seen by the five people who carry it — the liaison, three answer owners and the CFO — in their own words.
Six specialist agents draft every answer and check it against what each agency was told; owners approve and the CFO signs.
Splits each questionnaire, exam request list or PBC list into numbered questions, parts and topics — Excel, Word, PDF or a portal export — and records the sheet, row or page each came from.
Finds the closest approved and sent answer to each question across agencies, regulators, auditors and years, scores how much the question is the same and returns who approved it and when.
Replaces every number in reused language with today’s figure from its filed source — Annual Statement, Schedule P, RBC report, quarterly statement, reserve report, rate monitor, reinsurance summary or investment report — and marks what changed.
Compares every figure and claim in a draft with everything told to agencies, regulators and auditors in the last 18 months, and tells a contradiction apart from a change a later event explains.
Assigns each question to the owner of its topic and sets due dates back from the deadline: owner sign-off ten working days before, liaison review seven, CFO five.
Builds the submission pack — answers in the sender’s format, the exhibits cited, the consistency report and the transmittal letter — and files signed answers into the library as new versions.
Owners approve each paragraph; the CFO signs and releases. The agents propose; a named person decides.
The liaison and the owners can ask in plain words — or tell it what to change. Every answer names its source.
Why is Q4.3 flagged?
The draft says commercial auto renewal rate change in 2025 was 13.8%, from the final FY2025 rate monitor, which includes the Harbor Transport program. In May the company told AM Best 14.2%, excluding that program, signed by Victor Salgado, and Moody’s heard the same in July. Same line, same year — an analyst comparing notes would see two numbers. I recommend 14.2% with the same basis note.
Use the AM Best figure on Q4.3 and add the basis note
Done. Q4.3 now says 14.2% “excluding the Harbor Transport program exited in 2025”, the same words AM Best received. The figure ledger shows one number for commercial auto rate across AM Best, Moody’s and S&P. The paragraph is back with you to approve.
What else contradicts what we told AM Best or Moody’s?
Two more items. Q5.2 says $21.0M of favourable reserve development, from the fourth-quarter board paper; Schedule P and the Moody’s update say $18.4M. Q7.1 gives a $50M catastrophe retention against $40M told to AM Best — that one is the 1 July renewal, so it only needs a sentence.
Which Ohio exam items have no owner yet?
18 of 212 items have no topic owner — IT general controls, cybersecurity and the use of predictive models and AI in underwriting. The first is due in the 30 Nov batch. Leah Brandt owns the nearest topic; I can assign them to her when you say so.
13 screens from the working solution, on its sample data. Pick one to see it large.
Who owns which topic, how far back the guard looks and when each person signs are settings, saved as a new version for the next request.
| Setting | Default | Choose from |
|---|---|---|
| Topic owners | Eight topics; seven have a named owner | Any named owner by topic |
| Look-back window for the consistency guard | 18 months | 12 · 18 · 24 months |
| Tolerance | 0.0 points for ratios · $0.1M for amounts | Your tolerance for ratios and for amounts |
| Audiences compared | AM Best · S&P · Moody’s · state exam · auditor · board | Every audience you answer to |
| Block approval of a flagged paragraph | On | On · Off |
| Owner signs off | 10 working days before the deadline | Working days before the deadline |
| Liaison reviews · CFO signs and releases | 7 · 5 working days before | Working days before the deadline |
| Co-signer for risk management answers | Chief Risk Officer | Any named co-signer |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample data · “target” = the design goal, measured in the live solution · “estimated” = our estimate, including the 4–6 weeks by hand and the split of working days · “typical” = published rule (NAIC Model Law on Examinations, #390: a full-scope state financial examination at least once every five years). Context: NAIC Financial Condition Examiners Handbook · NAIC ORSA Model Act (#505) · AM Best Supplemental Rating Questionnaire. People, companies and figures named on this page are fictional — characters and sample data in the working solution.
A response solution for insurers that drafts rating agency and exam answers from past answers. Six agents split each questionnaire or request list into its questions, reuse the last approved answer with this year’s figures refreshed from the filed statements, check every figure against what each audience was already told and route each answer to its owner. Owners approve paragraph by paragraph and the CFO signs.
The request reader splits the questionnaire into numbered questions, keeping the agency’s numbering. The answer matcher finds the closest answer that was approved and sent — to AM Best, Moody’s, S&P, the examiners or the auditor — and shows how much the question is the same. In the working solution, 43 of S&P’s 48 questions were matched overnight; the 5 with no prior answer were drafted from the company’s reports and board papers, marked “No prior answer”, for their owners to check.
Only from filed sources: Annual and quarterly statements, Schedule P, the RBC report, the reserve report, the rate monitor, the reinsurance program summary, the investment report and the ORSA summary. Every figure cites its page and line. If no filed source exists, the figure is left blank for the owner — never estimated or rounded.
The consistency guard compares every figure and claim in a draft with everything sent to agencies, regulators, the auditor and the board in the look-back window — 18 months by default. When they differ beyond tolerance it quotes both statements, with source and signer; when a later event explains the difference, such as a reinsurance renewal, it asks for a sentence saying so. A flagged paragraph cannot be approved until a person resolves it.
Yes. In the working solution a 212-item request list from the Ohio Department of Insurance is read in one pass: 141 items matched to a 2021 exam response, an agency answer or a document, 48 needing 2021–2025 figures and 23 new. Items that ask for the same figure as a rating agency answer are linked so both say the same thing. The auditor’s prepared-by-client list is tracked the same way.
Yes. The agents draft and flag; they never send. Owners approve their answers paragraph by paragraph, the liaison reviews the whole submission, the Chief Risk Officer co-signs risk management answers and only the CFO signs and releases. Edits are kept as tracked revisions and every step is on the record.
Yes. Topic owners, the guard’s look-back window, the tolerance for ratios and amounts, the audiences compared, whether a flagged paragraph blocks approval, the sign-off dates and the co-signer are all settings. Changes are saved as a new version and apply from the next request.
The Agentic Solution Engine builds and deploys it from your requirements and documents — a past questionnaire, the answers you sent and your filed statements — and it goes live once every quality gate has passed. We will walk you through it on one of your own past requests first.
Adverse development seen in weeks, not at the next quarterly review.
1-min demoActuarial Assumption ManagementAn auditor's question answered in minutes, with the original support attached.
1-min demoPremium & Claims ReconciliationBreaks found and explained on day one of close, each tied to its source records.
1-min demoWe’ll run Rating Agency & Exam Response on a past questionnaire and the answers you already sent.