PTENES

Home / Cases / Credit union

Credit union · AIR

AI in lending without losing control: how a credit union prepared its data, people and governance

The credit union wanted to use AI in credit analysis and member service, but the regulator and the board asked for assurances it could not give.

42 → 55Overall index
CompleteFormat
820Respondents (59% response rate)
Context

Credit union with about 1,400 employees and 60 branches in smaller cities, serving farmers, small businesses and individuals. The technology team had modernized the core banking system, but relationship data was still scattered.

Challenge

A vendor proposed an AI credit scoring model. The board asked for a readiness assessment before approving it, concerned about explainability of decisions, bias against small farmers and regulatory requirements.

Instrument applied

AI Readiness (AIR)

Why this instrument. AIR Complete reads readiness by department and cross-references governance, data and people, giving the board an objective basis to decide the pace of adoption.

Format: Complete · Respondents: 820 of 1,400 (59%)

See the instrumentTake the free test

First application

Where the organization stood

Overall index
42 Band: Reactive 0100
  • Stagnant
  • Reactive
  • Emerging
  • Structured
  • Advanced
  • Benchmark
Results by pillar
Strategy and Use Cases 44 Data for AI 38 Technology and Platforms 52 People, Skills and Culture 41 Governance, Ethics and Risk 36 Operations, Adoption and Value 39
Plan executed

From results to action, in three waves

  1. 90 days

    Clear usage rules and priority use cases chosen.

  2. 6 months

    First use cases in production and teams trained.

  3. 12 months

    AI embedded in management, with measured gains and a new reading.

Reapplication

Before and after, pillar by pillar

Reapplied after 12 months: 42 (Reactive) → 55 (Emerging).

BeforeAfter Strategy and Use Cases 44 58 (+14) Data for AI 38 50 (+12) Technology and Platforms 52 60 (+8) People, Skills and Culture 41 53 (+12) Governance, Ethics and Risk 36 55 (+19) Operations, Adoption and Value 39 52 (+13)
Business outcomes

What changed in operations

  • Overall index moved from 42 (Reactive) to 55 (Emerging).
  • Duplicate member records reduced by about 70%.
  • Managers' response time to product questions halved in pilot branches.
  • Board approved the credit model pilot with documented criteria.
Lessons

What this case teaches

  • Saying "not yet" to the credit model was the decision that built the most trust with the board.
  • In regulated institutions, governance comes before use cases, not after.
  • Starting with an internal assistant delivered a quick win with no risk to members.

Want to see where your organization stands?

Start with the free quick test or talk to the team about the full application.

Take the free testTalk to the teamSee all cases

Other cases

AIR B2B software company

Many AI pilots, little value: how a software company prioritized use cases and built governance

Governance, Ethics and Risk from 38 to 61

Read the case → AIR Small e-commerce company

Everyone was already using AI, no one knew how: an e-commerce company brought usage inside the rules

Governance, Ethics and Risk from 34 to 55

Read the case → IESG Large mining company

Many programs, little integration: a mining company tied ESG to strategy before answering investors

Strategy and Integration from 38 to 53

Read the case →

Cases are presented without identifying clients. Sectors, sizes and figures are examples built to show the method.

WhatsApp