Data Assistant
Monday, 8:40am. Who wrote the most premium last month?
Most owners need three spreadsheets and most of a day to answer that. In Vora you ask in plain English, and the answer comes back with its chart, drawn from the same numbers as your dashboard.
The same question, two different Mondays
Before Vora
- 01Export production from one tab, persistency from another, last month's carrier report from email.
- 02VLOOKUP until the names line up. Ask ops to check the totals.
- 03Answer the question on Wednesday, after the meeting it was for.
Answered Wednesday, 2:15pm
With Vora
Who wrote the most premium last month?
Keisha Monroe wrote $9.9k in submitted annual premium, ahead of Vanessa Ortiz at $6.4k.
- 01Type the question the way you would ask a person.
- 02The assistant calls a read-only lookup on the same numbers your dashboard uses.
- 03The answer and the chart arrive together, before your coffee cools.
Answered Monday, 8:40am
Example agents and figures.
Ask it yourself
Admins ask questions like who wrote the most premium this month or how persistency compares by carrier. The assistant uses read-only lookups across production, book mix, persistency, commissions, debt and the scoreboard, then draws the chart.
Pick a question. The assistant chooses one of 13 read-only lookups, runs it on your agency only, and draws the chart from what came back.
Ask about your agency
Read-only lookups. No raw personal data is sent, and nothing is stored.
top_producers(period: mtd, limit: 5)
Keisha Monroe leads with $9.9k in submitted annual premium, followed by Vanessa Ortiz and Hannah Lowe.
Example answers. The lookup shown is the one the assistant called.
Useful because it is fenced in
It cannot write SQL or take actions, never sees raw personal data and does not keep conversation history. Requests run through a zero data retention gateway and are never used for training.
It cannot
Write SQL
It can only call the 13 lookups. There is no query box behind the curtain.
It cannot
Change anything
Every lookup is read-only. It cannot post, edit, delete or send.
It cannot
See raw personal data
Lookups return masked labels, and your question is scrubbed before it leaves.
It cannot
Remember the chat
Conversation history is not kept, and requests are never used for training.
Scrubbed before it leaves
People paste phone numbers and birth dates into questions. Vora replaces anything shaped like an SSN, a date of birth, an email or a phone number before the request goes out, through a zero data retention gateway, never used for training.
What the model receives
1 detail removedHow many policies does the client at [redacted phone] have with us?
The same patterns Vora runs before every request. Lookups return masked names too.
The only 13 things it can look up
Each one read-only, each one scoped to your agency and your role.
top_producers
Top agents by submitted production and policy count.
production_overview
Submitted, issued and in-force production for a period.
production_trend
A production metric over time, by day or month.
book_breakdown
How the book splits by carrier, product or status.
persistency
The 9-month ratio and the 1, 2, 4, 7 and 9-month checkpoints.
active_counts
Producing agents, active policies and clients.
commissions_summary
Expected commissions, payout buckets and a monthly series.
top_producers
Top agents by submitted production and policy count.
production_overview
Submitted, issued and in-force production for a period.
production_trend
A production metric over time, by day or month.
book_breakdown
How the book splits by carrier, product or status.
persistency
The 9-month ratio and the 1, 2, 4, 7 and 9-month checkpoints.
active_counts
Producing agents, active policies and clients.
commissions_summary
Expected commissions, payout buckets and a monthly series.
top_producers
Top agents by submitted production and policy count.
production_overview
Submitted, issued and in-force production for a period.
production_trend
A production metric over time, by day or month.
book_breakdown
How the book splits by carrier, product or status.
persistency
The 9-month ratio and the 1, 2, 4, 7 and 9-month checkpoints.
active_counts
Producing agents, active policies and clients.
commissions_summary
Expected commissions, payout buckets and a monthly series.
top_producers
Top agents by submitted production and policy count.
production_overview
Submitted, issued and in-force production for a period.
production_trend
A production metric over time, by day or month.
book_breakdown
How the book splits by carrier, product or status.
persistency
The 9-month ratio and the 1, 2, 4, 7 and 9-month checkpoints.
active_counts
Producing agents, active policies and clients.
commissions_summary
Expected commissions, payout buckets and a monthly series.
agent_debt_summary
Agents with the largest carrier debt balances.
carrier_debt
Open debt and credit by carrier.
scoreboard
The scoreboard, as your privacy settings allow.
find_agent
Agents by part of a name. Names only.
find_client
A count of matching clients, masked to first name and last initial.
sms_activity
Text counts by period. Counts only, never message text.
agent_debt_summary
Agents with the largest carrier debt balances.
carrier_debt
Open debt and credit by carrier.
scoreboard
The scoreboard, as your privacy settings allow.
find_agent
Agents by part of a name. Names only.
find_client
A count of matching clients, masked to first name and last initial.
sms_activity
Text counts by period. Counts only, never message text.
agent_debt_summary
Agents with the largest carrier debt balances.
carrier_debt
Open debt and credit by carrier.
scoreboard
The scoreboard, as your privacy settings allow.
find_agent
Agents by part of a name. Names only.
find_client
A count of matching clients, masked to first name and last initial.
sms_activity
Text counts by period. Counts only, never message text.
agent_debt_summary
Agents with the largest carrier debt balances.
carrier_debt
Open debt and credit by carrier.
scoreboard
The scoreboard, as your privacy settings allow.
find_agent
Agents by part of a name. Names only.
find_client
A count of matching clients, masked to first name and last initial.
sms_activity
Text counts by period. Counts only, never message text.
AI where the busywork used to be
Statement mapping
When a carrier spreadsheet does not match a known format, AI proposes a column mapping from the headers and masked cell shapes. An admin reviews and approves it before anything is applied.
unknown_carrier_stmt_oct.xlsx
Awaiting approval- Pol NbrSC6****73Policy number
- Ins NameR. J*****Client name
- Comm Amt$###.##Commission amount
- Pd Thru##/##/####Paid through date
Call summaries
Recorded calls can be transcribed and summarised, with the summary placed on the lead's timeline so the next conversation starts where the last one ended.
Call with Richard J.
6m 42sSummary
- Wants $15k final expense, budget near $60 a month.
- Waiting on his daughter to review. Call back Thursday.
Ask AI across the help center
About 160 help articles live inside the app. Ask AI answers from the article you are reading and the closest matches, with no access to your agency's data.
Limits you set
Daily usage limits per user and per agency. AI requests are metered so admins can see usage, and they are not billed separately.
- Plain-English questions, charted answers
- 13 read-only data lookups
- No raw PII sent to the model
- Daily usage limits per user and agency
- AI-assisted statement mapping
- Call transcription and summaries
- Ask AI across the help center
Works with
All features
Ask your first question about your own book
We load your carriers, grids and history during setup, so the assistant has something real to answer from on day one.