Use Case Intelligence RSS Search use cases Search
OpenAI Financial Services Asia-Pacific & Oceania

TRUSTBANK

Marketing Personalization

Furusato Nozei donors struggled to navigate a huge catalog of thank-you gifts and municipalities, so TRUSTBANK needed a more personalized way to guide discovery without requiring users to understand a complex tax-donation marketplace.

Read the original story Explore all use cases
Roughly 760,000 thank-you gifts listed on Furusato Choice

Reported impact

Business outcome

TRUSTBANK simplified gift discovery, helped users navigate a complex donation marketplace more naturally, and increased exposure for a broader range of municipalities and products instead of concentrating only on the most popular items.

  • Customer Experience
  • Revenue Growth
  • Scalability

Solution

TRUSTBANK partnered with Recursive to build Choice AI for the Furusato Choice app using OpenAI models. The team started by bringing AI-powered search into a marketplace of roughly 760,000 thank-you gifts, then expanded the experience into a multi-agent conversational recommender that can tailor guidance based on user intent, prior status, and vague requests such as shopping for a parent. A routing model analyzes the query and delegates work to specialized agents such as Search Agent, Recommendation Agent, and Greeting Agent, each of which can call additional subagents and tools. Recursive built the RAG system and agent infrastructure while TRUSTBANK prepared the gift database, requirements, and app integration. GPT-4.1 mini is the default runtime, with model switching experiments for latency and accuracy, and controlled randomness is used to prevent over-concentration on the same municipalities or popular items so more regions and niche products can be discovered.

Data flow

User questions enter Choice AI, where a routing model determines intent and invokes search, recommendation, greeting, or supporting agents. Those agents retrieve items from the gift database and use prompting logic plus controlled randomness to personalize and diversify suggestions before results are shown in the app.

Solution architecture

5 components · 3 layers
  1. Compute
    • GPT-4.1 mini Powers the default reasoning path for Choice AI while other model sizes are tested dynamically.
  2. Serving
    • Choice AI Provides the conversational gift-discovery experience inside the Furusato Choice app.
  3. Orchestration
    • Search Agent Finds relevant thank-you gifts from the large marketplace catalog based on user intent.
    • Recommendation Agent Personalizes which gifts and municipalities to surface for each user.
    • Greeting Agent Handles initial interaction flows and user engagement within the multi-agent system.

Architecture clues

  • Controlled randomness is introduced to diversify municipalities and products shown
  • Recursive designed and implemented the conversational AI agent and RAG system
  • Routing model delegates to specialized agents and subagents
  • User paths differ dynamically for first-time users versus existing users

Evidence from the source

"Choice AI currently runs on the GPT‑4.1 series."
"Furusato Choice ... with roughly 760,000 thank-you gifts listed."
"The core of Choice AI is its multiagent architecture."
"Using the OpenAI API, the company developed the Choice AI feature"
Transparency Evidence & methodology 75/100 reference value · 87% confidence · 1 source

Review how this reference was scored and trace each normalized claim back to its documented source.

01

Scoring methodology

How scores are calculated

Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.

Customer reference value75/100
  • Problem clarity20/20
    Measures whether the customer's operational problem is specific enough to compare.
  • Implementation depth25/25
    Rewards named components, data flow, integrations, and architecture detail.
  • Outcome specificity20/20
    Measures whether the reported benefit is concrete rather than a generic claim.
  • Quantified impact0/25
    0 metrics describe measurable impact.
  • Deployment scale10/10
    Rewards evidence of a live or scaled deployment and meaningful integration scope.
Evidence confidence87%
  • Source authority20/25
    The main evidence is an official vendor-published customer story.
  • Claim traceability21/25
    10 of 12 key fields link to a named source.
  • Direct evidence20/20
    4 direct source quotes retained.
  • Date provenance10/10
    The publication date comes from the source page or publisher.
  • Source transparency8/10
    1 source reference document publisher, URL, authority, retrieval, and supported fields.
  • Implementation specificity8/10
    Rewards source-backed components, data flow, quantified impact, and a specific outcome.
02

Traceability

Sources & provenance

Field coverage shows which facts each source supports. Dates are shown only when documented by the source.

  1. OpenAI

    TRUSTBANK customer story

    First-party
    https://openai.com/index/trustbank/
    Source type
    Vendor-published customer story
    Published
    Jan 27, 2026
    Supports
    • Customer
    • Business problem
    • Solution
    • Products
    • Solution components
    • Data flow
    • Integrations
    • technical area
    • use case category
    • Business outcome
    • outcome category
    • Metrics
    • Architecture
    • evidence quotes
    • Publication date

Date evidence

  • Published on source page Jan 27, 2026 · OpenAI · First-party
    Open date source
Not separately sourced Industry, Region, Company size. Treat these normalized attributes as lower-confidence until a dedicated source is added.

Related

Similar use cases

See all like this
OpenAI Nov 5, 2025 Chime Financial Services United States Chime wanted to turn AI into a marketing operating system that could personalize campaigns, accelerate creative work, and improve customer insight without fragmenting teams into disconnected point tools. ChatGPT is used across the entire organization in every department Marketing Personalization ChatGPT · Chime Content GPT · Custom GPT OpenAI Mar 28, 2024 Klarna Financial Services Global Klarna needed to improve online shopping discovery, automate high-volume customer service, and increase employee productivity while preserving quality across global fintech operations. $40 million USD estimated profit improvement in 2024 Customer Service AutomationMarketing Personalization AI assistant · ChatGPT Enterprise · ChatGPT plugin OpenAI Aug 10, 2026 Zapier Software / SaaS North America At Zapier, the enterprise marketing team is using ChatGPT Work to build automated funnel optimization systems, enabling them to deliver more value to their sales team with less manual effort. ChatGPT Work operates… Now, the enterprise marketing team can perform automatic QA/QC on thousands of leads every month. The… IT Operations AutomationMarketing Personalization ChatGPT Work