A retirement advisor that answers pension questions across four national systems for a third of a cent. The model plans the question and writes the answer. It never does the maths.
Financial services — pension & retirement funds
Databricks Agent Framework
Claude Opus 4.1 · Claude Sonnet 4
Agentic advisory with deterministic tools
Reckonsys built an agentic pension advisor on the Databricks Agent Framework that:
Chapter 01
Claude Opus 4.1 reads the member's question, works out what is being asked and which calculations it needs, then writes the answer.
What it does not do is arithmetic. Every tax calculation, withdrawal limit and benefit projection runs as a governed Unity Catalog function — eighteen of them, country by country. The model orchestrates; the code computes.
Member
Intelligence layer
Tool layer — code, not memory
Answer
A member turns 58 and wants to know when she can access her pension. It is a simple question with an answer that depends on her balance, her contribution history, her retirement age under the relevant national scheme, and the tax treatment of each withdrawal option available to her.
Today that question goes to a helpline. She waits, gets a callback scheduled days out, and eventually reaches an advisor who reads from a brochure that does not account for her situation. A simple question takes a week. Six months later, if a dispute arises, there is no record of what she was told.
At fifty thousand members and ten to twelve dollars per call interaction, the routine questions alone run past six million dollars a year — and advisors charging a hundred and fifty an hour spend their time on queries a calculator could answer.
The load-bearing decision. A model that remembers roughly what the concessional contribution cap was is useless in a regulated advisory setting. Put the rule in a function and the number is either right or it is a bug someone can find.
“A hallucinated withdrawal limit is not a wrong answer. It is a regulatory event.”
Privacy
Member names are anonymised before anything reaches the model and restored afterwards for the personalised response. The model reasons about a case; it does not need to know whose case it is.
Coverage
Each geography carries its own functions implementing that country's actual regulatory logic — not a generic calculator with a currency symbol swapped.
Australia
Superannuation
ATO tax calculator, Age Pension impact, balance projection
United States
401(k) & Social Security
IRS 401(k) tax, Social Security benefits, RMD calculator
United Kingdom
Workplace & State
HMRC tax calculator, State Pension check, pension drawdown
India
EPF & NPS
EPF tax, NPS benefits, EPS pension, corpus projection
Chapter 02
Before any answer reaches a member, Claude Sonnet 4 reads it against the retrieved data and the regulatory context, checking accuracy and adherence rather than fluency.
It is a different model from the one that wrote the answer, which is the point — a model reviewing its own reasoning tends to agree with itself.
Layer 01
Intelligence
Claude Opus 4.1 — privacy protection, query planning, tool orchestration, synthesis
Layer 02
Tools
18 country-specific Unity Catalog functions for tax, benefits and projections
Layer 03
Validation
Claude Sonnet 4 as judge — accuracy, quality and regulatory adherence
Layer 04
Governance
MLflow tracing and Unity Catalog — audit trail, cost tracking, citation registry
In regulated advice, generating the answer is half the job. Being able to show that it was checked, cited and recorded is the other half.
Validation runs in one of three modes: LLM judge, deterministic rule checks, or a hybrid of both. The choice is a configuration decision, so a fund with a stricter compliance posture can dial toward deterministic checking without a rebuild.
The agentic pipeline is not a single call. It is eight phases, and each is instrumented individually for duration and cost, which is what makes a slow or expensive query diagnosable rather than mysterious.
Economics
Cost is tracked per query and split between the main model and the judge, in real time — a number attached to each answer, not an estimate at the end of the month.
Human advisory: $10–12 per call interaction, with advisors at $150/hour and a week's turnaround on a simple question. At 50,000 members, over $6M a year on routine queries.
Agentic advisory: $0.003 per query, main model plus judge, answered instantly and available around the clock. Volume spikes cost what volume costs, with no added headcount.
A three-thousand-fold cost difference does not just save money — it changes which questions are worth answering. At ten dollars a call, a fund discourages members from asking. At a third of a cent, a member can ask six follow-up questions at two in the morning and the economics do not care. That is a different service, not a cheaper one.
Audit
MLflow traces the whole pipeline; Unity Catalog governs the functions, the member profiles and the citation registry. Regulatory references are attached to every response automatically.
The dispute six months later now has the question, the data retrieved, the functions called, the rules cited, the validation verdict and the cost.
That is the difference between advice you gave and advice you can prove you gave.
Nothing here is specific to pensions. Any regulated advisory domain has the same shape — a question that needs understanding, a calculation that must not be improvised, and a regulator who will ask what rule you applied. Wealth management, insurance comparison, loan eligibility and investment research all fit the pattern by changing only the functions in the tool layer.
$0.003 a query against $10–12 a call, with routine questions deflected from advisors charging $150 an hour.
Answers computed against the member's own balance, age and contribution history under their own national scheme.
Every interaction logged with its citations, validation verdict and the functions it called. Reconstructable months later.
Instant responses around the clock, with volume spikes absorbed at marginal cost rather than through overtime.
This case study documents a reference architecture for agentic advisory in financial services, implemented on Databricks Agent Framework with Claude Opus 4.1 and Claude Sonnet 4. Figures for cost per query, function count and pipeline structure are drawn from the working implementation. Available for walkthrough on request.
Let's collaborate to turn your business challenges into AI-powered success stories.
Get Started
