How to use Codex for family financial scenarios
Codex organized 21 statements into a balance database for family scenarios; Jesse Dhillon supplied assumptions and retained review, with key accounts missing.
linkedin.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Extract balances from these authorized financial statements into a source-linked database. Mark missing accounts and uncertain dates, and do not imply transaction-level analysis. Use my explicit schooling, property and work assumptions to prepare comparable family scenarios, clearly distinguishing assumptions from evidence for our review; do not move funds.
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- Jesse Dhillon
- Original account
- Read the source ↗
- Published
- Source link checked
- Reported result
- Codex created a balance database supporting an owner-led financial scenario document and family discussion; recommendations, completeness and financial benefit are not independently verified.
- Human involvement
- Dhillon gathered authenticated records, chose balance-level scope, supplied family assumptions and reviewed the output with his wife; no money movement was delegated.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Dhillon manually gathered 21 statements and identified incomplete account coverage.
- Codex parsed balances into a financial database rather than extracting every transaction.
- The owner refined assumptions and used the resulting scenario document for a discussion with his wife.
Jesse Dhillon used Codex to turn collected financial statements into a database for a family planning discussion. He first gathered 21 statements himself, dealing with passwords, one-time codes and access shared with his wife. Several materially significant accounts were still absent.
Codex extracted account balances rather than parsing every transaction. Dhillon describes that as a deliberate tradeoff against the cost and detail of full transaction analysis. The structured balances then supported a more extensive conversation about the family’s circumstances and goals.
He supplied context about employment, private equity, property, schooling and a possible move. After prompt refinement and further discussion, the resulting document compared assumptions around private versus public school, portfolio exits, housing affordability and multiyear allocations. Dhillon then used it for a structured conversation with his wife.
The agent-built database is the concrete tool action; the account does not establish that every later GPT exchange was an autonomous agent run. Nor does it disclose enough financial detail to verify the recommendations. The result is a prepared scenario discussion, limited by missing accounts and human-supplied assumptions, rather than an implemented investment plan or proven financial benefit.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: LinkedIn · Jesse Dhillon
Frequently asked questions
Did the workflow include all accounts?
No. The author explicitly says some materially significant accounts were missing.
Were individual transactions fully analyzed?
No. Codex’s structured extraction focused on balances as a deliberate scope tradeoff.
Did an agent implement the financial recommendations?
The account reports a family discussion, not investment execution or money movement.
More details
About this account
Dhillon manually gathered 21 statements and identified incomplete account coverage. Codex parsed balances into a financial database rather than extracting every transaction.
Primary post and main publication JSON-LD opened. Codex database work is distinct from later GPT conversation. Missing significant accounts and undisclosed financial assumptions prevent independent validation of advice. Suggested prompt written by Agents In Use; no verbatim original task prompt was supplied.
Credit: Jesse Dhillon — LinkedIn
Source published
Added to Agents In Use
Source checked
Tools and topics
Codex · Financial statements · Database · GPT conversation
balance database · family scenarios · incomplete accounts
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.