Wibmo AURA

AURA

Autonomous Retrieval Agent

A conversational way to work with Wibmo Products

The AURA proposition

Make IVS easier to work with.

A proposed conversational layer for internal teams and bank customers to request reports, explore performance and manage routine updates, within existing IVS permissions.

01
Generate reports
  • Request the data you need
  • View and download the response
02
Investigate issues
  • Break down failures
  • Explore possible causes
03
Compare performance
  • Compare periods and authorised banks
  • Track trends and recurring patterns
04
Set up alerts
  • Define what needs attention
  • Receive notifications when conditions are met
05
Schedule reports
  • Prepare recurring bank reports
  • Reduce repeated preparation
06
Share outputs
  • Export reports and charts
  • Share findings with supporting evidence

Start with trusted reporting. Expand capabilities as their value and controls are proven.

Today

How teams deliver reports and insights today

01
Receive the request. Banks, product and analytics teams need reports, performance comparisons, issue analysis and regular updates.
02
Gather and check the data. Teams pull data from dashboards, logs and spreadsheets, then check completeness and consistency.
03
Prepare the output. Analysts build reports, compare trends, investigate changes and summarise findings.
04
Review, share and follow up. Teams validate the output, share it with stakeholders and repeat parts of the work for follow-up requests.

Recurring requests add to the manual workload. Data preparation, checks and coordination can stretch delivery across several days.

What AURA returns

One prompt. The right response format.

AURA presents the data as numbers, tables, charts or explanations, based on what you need.

01
Numbers
  • Volumes, success rates and failure counts
  • Period and scope stated
02
Tables and reports
  • Detailed breakdowns
  • Sort, filter and download
03
Charts
  • Trends, comparisons and distributions
  • Changes made easier to see
04
Explanations
  • Findings supported by data
  • Gaps clearly stated

Not just a written response. An output you can review, explore and share.

Working with AURA

Describe what you need. AURA brings it into view.

01Track daily performance

Bank / Internal

“Show ICICI Bank’s daily authentication volumes and success rates for the last 30 days.”

Response Daily table · Trend chart

02Compare banks

Authorised internal users

“Compare FlashPay authentication volumes and success rates for Axis and SBIC over the last 30 days.”

Response Comparison table · Trend chart

03Review merchant performance

Bank / Internal

“For Kotak debit cards, group the top 10 merchants by authentication volume into success-rate ranges for the last 10 days.”

Response Grouped merchant table · Distribution chart

04Schedule a report

Bank / Internal

“Prepare our bank’s authentication report every Monday, comparing the previous two weeks.”

Response Scheduled report · Download

Bank users see only their own bank. Internal users see only authorised banks.

The AURA experience

One prompt. A visual response.

User prompt

Press Play to enter a prompt.

Follow-up prompt: “Break this down by authentication method.”

The bigger opportunity

AURA starts with reporting. The opportunity goes much further.

Evolve IVS from a menu-driven portal into a conversational platform for internal teams and bank customers.

01

Start with insights

Generate reports, compare performance and explore failures. Reduce repetitive work and dependence on support teams.

02

Expand into assisted operations

Support investigations, propose configuration changes and guide workflows. Keep execution subject to permissions and explicit approvals.

03

Explore cross-product intelligence

Connect patterns across authentication products, flag emerging issues and recommend improvements for human review.

Give teams time back. Help banks do more. Create room for growth.

Potential commercial model: Basic capabilities included within usage limits; paid tiers for advanced insights and assisted operations.

Data and access · Proposed controls

AURA adds intelligence. IVS keeps control.

Only the IVS Data Layer reads bank records. Model inputs are limited to approved information.

LIMITED INPUTS

What the AI model may receive

  • The prompt, with sensitive details removed
  • Approved field names and metric definitions
  • Only policy-approved summary figures

Summary figures are still bank-derived data and need explicit approval.

PROTECTED RECORDS

What must stay out of model inputs

  • Raw transaction records
  • Cardholder and card data
  • Credentials or direct database access

The IVS Data Layer reads records. Authorised views remain inside IVS.

IVS enforces access

Bank users: own bank only. Internal users: authorised banks only.

Read-only at the start

No configuration changes. Future actions need separate permissions and approvals.

Keep an audit trail

Record requests, access decisions and response references. Mask sensitive details.

Before launch: Approve model hosting, permitted inputs, retention and no-training terms.

How we get there

Where the AI works, and where it does not.

AURA
01
Understand the user’s intent
02
Map to approved definitions
03
Structure the request
07
Build the applicable view
IVS
Backend
04
Validate and authorise
05
Execute structured request
06
Retrieve the structured response

AURA opens possibilities. IVS Backend grounds them in deterministic data. AURA presents it through the user's lens.

Delivery · Five phases

From trusted data to bank access.

01Start here
Pilot and Data Cleanup
  • One bank; map sources, tables and columns.
  • Clean data and verify connections.
Done whenPilot data is clean and complete
02
Meaning and Controls
  • Approve field meanings, metrics, calculations and filters.
  • Enforce read-only, bank-scoped execution with traceability and limits.
Done whenDefinitions approved; access tests pass.
03
Prove the Full Flow
  • Run prompt → request → IVS → view against verified responses.
  • Validate with analysts; measure response quality, Time and Cost.
Done whenResponse and Access tests pass; Analysts accept results.
04
Expand internal use
  • Extend reports and trends to more prompts and banks.
  • Add guided investigation, briefings and alerts; track usage and benefit.
Done whenAnalysts repeatedly use it, with measured benefit.
05
Enable Bank Access
  • Pilot bank-facing views with verified data isolation.
  • Clear contractual and data-handling approvals.
Done whenAccess is approved and the bank accepts the pilot.
Architecture and flow

One prompt - How AURA works end to end?

Scroll the diagram sideways to follow the arrows.
06
Watcher
  • Scheduled checks and briefings
  • Surfaces what nobody asked
Only the IVS Data Layer reads bank records (not AURA)
Read-only, within Bank Permissions
User Prompt
01
Router
  • Decides which block handles it
  • Does not generate responses
03
Request builder
  • Turns intent into a request
  • Writes no queries of its own
04
Data layer
  • The only block touching data
  • Runs under the user's permissions
05
Presenter
  • Number, table, chart or text
  • States period and gaps
Response
02
Catalogue
  • What exists: metrics, methods, BINs
  • Field names and synonyms
07
Case memory
  • Keeps resolved cases
  • And what fixed them
08
Evaluator
  • Grades against known responses
  • Logs every step
The way forward

A focused start. A bigger opportunity.

Deevanshu · Product Owner Working with the team under Anand’s supervision.

Proposed first step
An internal pilot with one bank’s data, priority reports and read-only access.
Leadership support
Align on the pilot direction and help prioritise engineering capacity and security review.
Next review
A working pilot, validated outputs and a recommendation on what to expand.

Reporting is the starting point. The ambition is to help banks do more with Wibmo.

01 / 12
Arrow keys to move