AI Capability Cards

Exercise 1: AI Tech Radar - Map your organisation's AI opportunities

How to Use These Cards

For the exercise:

  1. Review each capability with your team
  2. Write the capability name on a Post-it note
  3. Place on the butcher paper matrix based on:
    • X-axis (horizontal): Incremental → Transformational
    • Y-axis (vertical): Process Improvement → Strategic Innovation
  4. Identify gaps: What quadrants are empty?

1. Robotic Process Automation (RPA)

Technology: Software bots that automate repetitive tasks

Example: Automated invoice processing, data entry, form filling

Typical ROI: 30-50% cost reduction in targeted processes

Implementation Time: 3-6 months

Data Requirements: Structured data, clear business rules

Questions to Consider:
  • What repetitive tasks consume significant staff time?
  • Are your processes standardized enough for automation?
  • What's the volume that justifies automation investment?

2. Predictive Analytics

Technology: Machine learning models that forecast future outcomes

Example: Demand forecasting, customer churn prediction, equipment failure prediction

Typical ROI: 15-25% improvement in planning accuracy

Implementation Time: 6-12 months

Data Requirements: Historical data (minimum 2 years), clean datasets

Questions to Consider:
  • What decisions would benefit from better forecasting?
  • Do you have sufficient historical data?
  • Can you act on the predictions once generated?

3. Natural Language Processing (NLP)

Technology: AI that understands and generates human language

Example: Chatbots, sentiment analysis, document summarization, email classification

Typical ROI: 40-60% reduction in routine customer service inquiries

Implementation Time: 3-9 months

Data Requirements: Text data corpus, conversation logs

Questions to Consider:
  • What percentage of inquiries are routine/repetitive?
  • Do you have documentation/FAQs to train the system?
  • How important is conversational quality to your brand?

4. Computer Vision

Technology: AI that interprets visual information

Example: Quality inspection, asset monitoring, safety compliance, inventory counting

Typical ROI: 70-90% reduction in inspection time, 95%+ accuracy

Implementation Time: 6-12 months

Data Requirements: Thousands of labeled images

Questions to Consider:
  • What visual inspection tasks are currently manual?
  • Can you collect sufficient training images?
  • What's the cost of errors (false positives/negatives)?

5. Recommendation Systems

Technology: AI that personalizes content and product suggestions

Example: Product recommendations, content curation, next-best-action suggestions

Typical ROI: 10-30% increase in conversion rates

Implementation Time: 6-9 months

Data Requirements: User behaviour data, product catalog, interaction history

Questions to Consider:
  • Do you have sufficient variety in offerings to recommend?
  • Can you track user behaviour/preferences?
  • What's the business impact of better matching?

6. Dynamic Pricing & Optimization

Technology: AI that adjusts prices based on demand, competition, inventory

Example: Real-time pricing, markdown optimization, promotional planning

Typical ROI: 5-15% revenue increase, 10-20% margin improvement

Implementation Time: 9-15 months

Data Requirements: Pricing history, competitor data, demand signals

Questions to Consider:
  • How price-sensitive are your customers?
  • Can you implement price changes quickly?
  • Do you have competitive intelligence capabilities?

7. Intelligent Document Processing

Technology: AI that extracts and interprets information from documents

Example: Contract analysis, invoice processing, compliance document review

Typical ROI: 60-80% reduction in document processing time

Implementation Time: 6-12 months

Data Requirements: Document corpus for training, annotation effort

Questions to Consider:
  • What documents require significant manual review?
  • Are documents standardized or highly variable?
  • What's the risk/cost of extraction errors?

8. AI-Powered Search & Discovery

Technology: Semantic search that understands intent and context

Example: Enterprise knowledge search, product discovery, customer self-service

Typical ROI: 40-60% reduction in search time, 30% increase in self-service resolution

Implementation Time: 6-9 months

Data Requirements: Document repositories, metadata, usage patterns

Questions to Consider:
  • How much time do employees spend searching for information?
  • Is knowledge scattered across multiple systems?
  • What's the cost of not finding the right information quickly?

9. Autonomous Agents

Technology: AI systems that plan and execute multi-step tasks

Example: Autonomous procurement agents, research assistants, workflow orchestration

Typical ROI: 50-70% reduction in coordination overhead

Implementation Time: 12-18 months

Data Requirements: Workflow data, decision criteria, integration with existing systems

Questions to Consider:
  • What multi-step processes require constant human coordination?
  • Can you clearly define decision criteria and boundaries?
  • What level of autonomy is appropriate for your risk tolerance?

10. Generative AI (LLMs)

Technology: Large language models that generate text, code, or creative content

Example: Content creation, code generation, report writing, design variations

Typical ROI: 30-50% increase in content production speed

Implementation Time: 3-6 months (using existing platforms), 12+ months (custom)

Data Requirements: Minimal for general use, proprietary data for specialized applications

Questions to Consider:
  • What creative/knowledge work is repetitive?
  • How important is brand voice/quality control?
  • Do you need domain-specific fine-tuning?

11. Fraud Detection & Risk Management

Technology: AI that identifies anomalies and suspicious patterns

Example: Transaction fraud, cyber threat detection, insurance claim fraud

Typical ROI: 40-70% reduction in fraud losses, 60%+ detection improvement

Implementation Time: 9-15 months

Data Requirements: Transaction history, labeled fraud cases, real-time data feeds

Questions to Consider:
  • What's the current cost of fraud/risk to your business?
  • Can you respond quickly to detected threats?
  • What's acceptable false positive rate?

12. AI-Driven Business Intelligence

Technology: Automated insights and reporting from business data

Example: Anomaly detection in KPIs, automated report generation, insight discovery

Typical ROI: 50-70% reduction in reporting time, faster decision-making

Implementation Time: 6-12 months

Data Requirements: Integrated data warehouse, clean business metrics

Questions to Consider:
  • How much time is spent creating reports vs. acting on insights?
  • Are insights discovered too late to act on?
  • Do you have data infrastructure to support this?

Add Your Own Capabilities

During the exercise, you can add industry-specific or emerging AI capabilities that aren't listed here. Write them on Post-it notes and place them on the matrix!

← Back to Exercises