Use Case Intelligence RSS Search use cases Search
IBM Software / SaaS Japan

Unionthink Co., Ltd.

Document Processing

Unionthink's customer center was manually categorizing rising volumes of product inquiries, making it hard to improve FAQs, keep up with demand and sustain customer satisfaction with a limited workforce.

Read the original story Explore all use cases
>90% prediction accuracy for classifications

Reported impact

Business outcome

Unionthink reached high classification accuracy and established a foundation for broader AI use in its customer center, reducing manual work while improving the speed and consistency of inquiry analysis.

  • Customer Experience
  • Productivity Improvement

Solution

Unionthink worked with IBM to turn inquiry classification into an AI-assisted workflow using IBM watsonx.ai and IBM Watson Natural Language Processing Library for Embed. The company first trained a custom language model in watsonx.ai on manually categorized customer inquiries, then compared the accuracy of multiple IBM-provided models to select the best option. That model was refined through technical verification and tuning of the training data before being used to support the customer center's daily classification tasks. A parallel proof of concept explored keyword extraction from inquiries so the company could better analyze recurring topics and improve its published FAQ content. The resulting setup gives Unionthink a repeatable language-understanding pipeline for inquiry triage, helping a small software vendor replace manual categorization work with AI-based classification that scales more effectively as demand rises.

Data flow

Manually categorized inquiries are used to train a custom model in watsonx.ai, new customer-center inquiries are classified with IBM NLP services, and the resulting categories and keywords feed FAQ analysis and response-improvement workflows.

Solution architecture

2 components · 2 layers
  1. Compute
    • IBM watsonx.ai Hosts the custom language model trained on historically categorized customer inquiries.
  2. Serving
    • IBM Watson Natural Language Processing Library for Embed Provides the NLP capabilities used to classify inquiries and support downstream keyword-extraction use cases.

Architecture clues

  • The proof of concept also explored keyword extraction beyond pure classification.
  • The team trained on previously manually categorized inquiry data rather than starting from generic prompts alone.
  • Unionthink compared several IBM models before selecting the most effective one.

Evidence from the source

>90% prediction accuracy for classifications
IBM and Unionthink created a custom language model using IBM watsonx.ai
explored tasks like keyword extraction from inquiries
improving category classification for manual inquiries
Transparency Evidence & methodology 85/100 reference value · 89% 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 value85/100
  • 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 impact15/25
    1 metric describe measurable impact.
  • Deployment scale5/10
    Rewards evidence of a live or scaled deployment and meaningful integration scope.
Evidence confidence89%
  • 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 specificity10/10
    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. IBM

    Unionthink Co., Ltd. customer story

    First-party
    https://www.ibm.com/case-studies/unionthink
    Source type
    Vendor-published customer story
    Published
    Jul 2, 2024
    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

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
IBM Oct 15, 2025 IBM Tax Organization Software / SaaS Global IBM's tax organization needed to automate global compliance work because inaccessible and incomplete data made tax-return preparation and document validation too manual across many countries and source systems. 3 use cases on track to save tens of thousands of hours annually Document ProcessingSecurity and Compliance EY.ai for tax · IBM watsonx.ai · IBM watsonx.data IBM Oct 7, 2025 Dynamiq Software / SaaS United States Dynamiq needed a way to help legal teams process large volumes of contracts, compliance reports, and regulatory documents faster, while reducing manual review work and improving the consistency of compliance decisions. 50% reduction in contract review, down from 1.5 hours to 45 minutes Document ProcessingEnterprise Knowledge Search IBM Granite · IBM watsonx Orchestrate · IBM watsonx.data IBM Jun 10, 2025 Barré Technologies Software / SaaS Hungary Barré Technologies wanted to help customers handle labor-intensive document and process-management work such as invoice processing, repository navigation, and data lookups without relying on scarce specialist staff or… 30% productivity increase Document ProcessingEnterprise Knowledge Search IBM Cloud Pak for Data · IBM FileNet · IBM watsonx