Innovation Challenges

;
Challenge Owner(s)
Green Link Digital Bank, The Good Life Healthcare, Playpoint, Prudential Plc, JL Cable Detection (JLCD), beIN Sports Asia, Goldbell Corporation, C&W Services Singapore, Ministry of Education (MOE), NParks
Organiser(s) Infocomm Media Development Authority (IMDA)
Industry Type(s)
Digital/ICT, Education Services, Environmental Services, Financial Services
Opportunities and Support Over S$500,000 prize monies
Application Start Date 10 March 2025
Application End Date 21 April 2025
Website Click here to learn more

About Challenge

Open Innovation Call 25 

Our latest Open Innovation Platform Call is now live! Transform challenges from beIN Sports Asia, C&W Services, Prudential and more into innovative solutions. Submit your proposals in key tech areas, including AI, Data Management and Fintech.

Learn More

Challenge Owner(s)Ministry of Education (MOE)
Industry Types(s)
Digital/ICT, Education Services

ENHANCING FLOOR CLEANING OPERATIONS IN SCHOOLS

Schools in Singapore are required to adhere to the environmental sanitation regime set by the
National Environment Agency (NEA) to ensure a safe and healthy learning environment for students

and staff.

What We Are Looking For

• Automated wet/dry cleaning of varied floor surfaces. Able to autonomously clean a variety of
floor types, including tiles, vinyl, concrete and pebble wash, and adapt to both wet and dry-
cleaning requirements with minimal manual intervention.

• Situational awareness for obstacle detection and avoidance. Able to detect and navigate
around obstacles such as people and objects to avoid collisions, and to adjust movement
pathways dynamically based on real-time environmental conditions.

• Navigation of infrastructure challenges. Able to traverse infrastructure challenges in the
ventilated areas (e.g. corridors, walkways, school canteen) of a typical school building,
including curbs, drains, gutters, gaps and uneven surfaces, ensuring smooth movement and
operations around the school compound. Ability to move between different floors of a multi-
storey building without lifts (e.g. able to navigate stairs) will be an advantage.

• Automated task scheduling and alerts. Able to automatically schedule cleaning tasks including
during night-time, with real-time notifications and alerts for status updates or when human
intervention is required.

• Manual mode. Allow manual overrides and inputs to cleaning modes/scheduled tasks to

address specific cleaning requirements.

Learn More

 

Challenge Owner(s)NParks
Industry Types(s)
Digital/ICT, Environmental Services

SMART DETECTION AND DETERRENCE OF LONG-TAILEDMACAQUES

Macaque sightings are relatively common encounters in residential estates fringing nature reserves
such as MacRitchie, Central Catchment and Bukit Timah. The macaques are forest fringe animals
belonging to the long-tailed macaque species that move between forest patches in search of food and

territory.

What We Are Looking For

  • Precise and accurate detection of long-tailed macaques. Ability to accurately detect and recognise the correct species of long-tailed macaques which can be triggered by visual, sounds, movements, behaviours, body heat but not limited to these.
  • Automated recording of macaque movements. Ability to record long-tailed macaques and their movements by location, without compromising the privacy of residents in the estate. These recordings can be in the form of photos and videos with the precise location that they are detected at tagged.
  • Functional in low-light and unfavourable weather conditions. Ability to capture, process and identify presence of long-tailed macaques in low light environment and unfavourable weather conditions including strong sunlight, strong winds, heavy rain, haze or fog.
  • Automated real-time alerts. Upon detecting of long-tailed macaques, automatically push alerts to the relevant NParks officers in real-time so that they can immediately plan for the next steps. Transmission of data can be via LTE-M or other methods without involving additional infrastructure set-up costs.
  • Guide and re-direct long-tailed macaques back to the wildlife. Include a non-intrusive solution to guide and re direct macaques that have ventured into resident estates back to the forest (e.g. deterrence drones, robotic guard dogs, smell). The solution can utilise digital technology or a combination of both digital technology and non-technology related.

Learn More

Challenge Owner(s)C&W Services Singapore
Industry Types(s)
Digital/ICT, Infrastructure

CENTRALISED DATA MANAGEMENT PLATFORM FOR FACILITY MANAGEMENT

C&W Services Singapore manages complex facility operations across various government and private

sector properties.

What We Are Looking For

• Standardisation of data. Clean and standardise data to improve data usability, accuracy and
reliability for data analysis.

• Seamless integration with multiple data sources. Ability to automatically pull data from
existing systems including CMMS, IoT system, vendor and client systems. The solution should
also support integration with third-party APIs for seamless data aggregation.

• Automated data processing. Ability to ingest, clean and process raw data in large scale and in
real time. The solution should also support multiple data formats including structured, semi-
structured and unstructured data. In addition, the data should be ready to be imported for
visualisation purposes in BI systems such as PowerBI.

• Real-time insights and alerts. Enable real-time monitoring and predictive analytics for
proactive maintenance and quicker actions to be taken to address potential issues.

• Generation of data reports. Ability to generate reports containing cleaned and categorised
datasets for further analysis.

• Centralised data lake. Design a data lake with suitable architecture to support the extraction,

housing and management of data.

Learn More

Challenge Owner(s)Goldbell Corporation
Industry Types(s)Digital/ICT

AI-POWERED CRM TO ENHANCE CUSTOMER ENGAGEMENT AND OPERATIONAL EFFICIENCY

Goldbell’s IT system team maintains oversight of their customer relationship management (CRM)
infrastructure, which was initially implemented to bolster the sales team's capabilities in customer
service delivery, revenue generation, and post-purchase support. This system serves as the primary
tool for managing customer interactions, tracking sales opportunities, and coordinating aftersales

activities across the organisation.

What We Are Looking For

• Seamless Data & Cloud Integration. A cloud-based centralised gateway enabling real-time
synchronisation between D365 and Maximo, with automated validation and secure API
integration. The system serves as a single source of truth, accessible across iOS and Android
platforms, with automated backups and detailed audit trails.

• User-Friendly Interface and Role-Based Personalisation. A modern, customisable dashboard
system offering three distinct interfaces: (i) sales manager dashboard for performance
tracking and resource allocation; (ii) personalised salesperson dashboard for account and
customer management; and (iii) customer dashboard providing data entry and consolidated
interaction history. Each interface features simplified navigation, digital document
management, and mobile responsiveness, supported by intuitive self-help features.

• AI-Powered Sales & Lead Management. An intelligent system combining advanced analytics
for performance tracking, automated lead potential scoring and distribution within the team,
predictive sales forecasting, and comprehensive CRM integration for existing clients’ profiles.
The platform will enable managers to dynamically assign orders based on team capacity and 
expertise while providing real-time insights into sales performance and customer engagement

metrics.

Learn More

Challenge Owner(s)beIN Sports Asia
Industry Types(s)Digital/ICT

REAL-TIME AUTOMATED GENERATION OF SPORTS VIDEO HIGHLIGHTS

The Digital Products & Innovation team at beIN Sports Asia oversees the generation of video highlights clips that are offered across multiple platforms, including linear TV channels across the Asia Pacific region, social media platforms like YouTube and Facebook, and beIN Sports CONNECT, the company’s subscription-based streaming service accessible via mobile apps, web, and smart TVs.

What We Are Looking For

● Smart video and audio analysis. Able to analyse both visual and audio input from a live football

match video feed.

- Key event recognition and prioritisation. Able to recognise key events in a football match video feed captured with different kinds of camera shots, such as goals, assists, penalties, and crowd reactions, and prioritise/rank them based on parameters to be defined with beIN.

- Off-pitch event recognition. Besides key on-pitch events/gameplay, able to identify key events happening off-pitch in a match through both visual and audio analysis, including words used, tone and volume of voice of the commentators, and audio crowd reactions.

- Live and replayed footage differentiation. Able to distinguish between live moments and replayed footage within the livestream video, ensuring that events selected for the highlights clips reflect actual gameplay and not replayed clips.

- Team differentiation. Able to distinguish between the two teams playing the match (e.g. by uniform colour, logos etc).

- Spoken language cue recognition. Able to detect the end of sentences in the commentators’ speech, to avoid abrupt cuts, and differentiate between different commentators’ voices.

● Smart catch-up video clip generation live on demand. Able to autonomously generate 30- second highlights video clips from a livestream video feed, which users can view on demand.

- Real-time dynamic generation. Generate highlights clips in real-time (i.e. within seconds) from the part of the match that the user has missed (e.g. first 10 minutes if the user comes in 10 minutes late) while the live match is ongoing, and dynamically update the clips as the match progresses such that the user will only see the latest clip.

- Accuracy and relevance. As the highlights clip is generated during a live match where manual editing would not be possible, the clip should be accurate in terms of key events identified and contextually relevant even in the absence of key events (e.g. goals).

- Chronological order and audio synchronisation. The events in the highlights clip should be arranged in chronological order, and audio should be automatically synchronised without commentary being cut off, maintaining continuity across clips.

- Seamless transitions. Automatically add in transitions between events (e.g. crossfades, fade-to-black) in the highlights clip.

● Self-learning capability and feedback loop. Able to learn from past events and feedback input by the beIN team to improve its accuracy of event recognition, and over time, to be able to generate highlights clips without reliance on Opta’s data feeds.

● Varied duration clip generation for post-match highlights. After the completion of a match, able to automatically generate a range of highlights clips options based on a predefined list of durations (e.g. 2 minutes, 5 minutes, 12 minutes etc), allowing users to select the length of clip that caters to their individual viewing preference.

● Customisable preference-based clip generation. Beyond standard match-based events, able to generate highlights clips based on users’ personalised preferences, such as player-specific highlights which track a particular player’s performances throughout a match, or team-centric highlights which include highlights of a particular team across multiple events. 

Learn More


Challenge Owner(s)JL Cable Detection (JLCD)
Industry Types(s)Digital/ICT

AUTOMATING DATA MANAGEMENT AND INFORMATION SHARING ACROSS TEAMS

JLCD operates in the dynamic and critical field of underground utility location within Singapore's
construction sector. The company manages a diverse portfolio of projects, ranging from large-scale
infrastructure developments such as the North-South Corridor with LTA and Changi Airport Terminal

5 with CAG to smaller projects like signage installations for local businesses.

What We Are Looking For

• Centralised Data Management. The system should provide a unified platform beyond their current Google sheets and Sharepoint documents to consolidate all project-related information. This includes customer details, quotations, invoices, utility plans, and project updates. The centralised database should allow easy access, updating, and management of data, eliminating the current fragmentation across various formats and storage systems, and significantly reduce time spent searching for information and minimise the risk of data inconsistencies.

• Real-time Project Tracking. A key feature of the solution should be a customer-facing online web application. This portal should enable clients to track their project progress in real-time, reducing the need for frequent status inquiries. The system should provide automated notifications at key project milestones such as their permit applications from the respective authorities and allow clients to access and download the reports. This transparency will not only improve customer satisfaction but also reduce the administrative burden on JLCD's staff.

Learn More

Challenge Owner(s)Prudential Plc
Industry Types(s)
Digital/ICT, Financial Services

INTELLIGENT PROBLEM STATEMENT DEVELOPMENT AND CLUSTERING PLATFORM FOR AI PROJECTS

Prudential AI Lab, established in 2024, plays a crucial role in bridging its internal business challenges with AI solutions through collaboration with startups, technology partners, institutes of higher learning (IHLs), and government agencies. Since its inception, the AI Lab has received over 150 problem statements from employees across international markets through its cohort-based innovation programme, running two to three cycles annually. 

What We Are Looking For

• Automated Problem Statement Enhancement. The solution should leverage AI/NLP technologies to guide users through problem statement development and refinement, ensuring submissions are complete, clear, and suitable for AI intervention. It should analyse submissions for quality, prompt users with specific questions (e.g. help users identify data prerequisites, technical considerations, and success metrics) to guide articulation of their problem statement and enable iterative refinement through intelligent feedback.

• Clustering and Categorisation. The solution should employ intelligent pattern recognition to automatically group similar problem statements and identify opportunities with potential synergies across different business units. It should categorise submissions based on multiple dimensions (e.g., theme, type, data requirements, complexity) and AI applicability, flag potential duplicates, and generate insights on common challenge patterns.

• Research-backed frameworks. The solution should incorporate Design Thinking, research from the web and other data sources, and/or other AI frameworks/matrixes that can support users in the AI problem development and validation process. It should be able to flag non-AI suitable challenges and provide users with more context in the decision-making process.

• Evaluation and prioritisation mechanism. The solution should rank and score problem statements submitted based on predefined criteria such as feasibility, data availability and potential business impact, and allow for human override where necessary. 

Learn More

Challenge Owner(s)Playpoint
Industry Types(s)Digital/ICT

DESIGN AUTOMATION TOOL TO GENERATE INNOVATIVE PLAYGROUND PROPOSALS

Currently, Playpoint's design team faces challenges in meeting the rising demand for innovative playground designs while maintaining high standards of efficiency and quality. The existing design process, which can extend up to two weeks of waiting time from their clients for a new proposal, often limits the company’s ability to respond swiftly to client needs and secure more projects. While expedited timelines are sometimes necessary, they risk compromising design quality. 

What We Are Looking For

● Design library integration. Incorporate Playpoint’s existing digital design library, which includes designs in PDF format, already categorised by themes (e.g., ship, corals, forest themes) to facilitate quick adaptation of design elements. The solution should analyse this existing library, identifying common features such as layout styles, play equipment types (e.g., slides, swings, climbing walls), and safety considerations (e.g., spacing, safety barriers), and use this knowledge to generate new designs by adapting existing patterns to different contexts, like user demographics (age groups), or site specific constraints (e.g., terrain, environment).

● 3D design generation and modification. Ability to produce initial 3D designs based on specified user-defined parameters (e.g. theme, layout, materials, safety standards, age-appropriate features, existing landscapes, etc) to influence the overall aesthetic, and ensure that the generated designs meet regulations and integrate well with the surrounding environment. Allow designers to define and modify input parameters, enabling the solution to make changes to the design accordingly.

● Collaborative viewing and editing. Enable multiple designers to view, edit and collaborate in real-time, facilitating faster iteration.

● Export functionality. Allow designers to export designs in formats compatible with Playpoint’s current 3D modeling design software (Rhinoceros) and other 2D design tools (PhotoShop and Indesign).

Learn More

Challenge Owner(s)The Good Life Healthcare
Industry Types(s)
Digital/ICT, Healthcare & Biomedical

AI-DRIVEN HEALTH LITERATE CHATBOT PROVIDING INFORMED MEDICAL INSIGHTS

Health literacy is a crucial aspect of preventive healthcare that extends beyond basic health screening and vaccination. It encompasses an individual's ability to access, comprehend, and utilise health information to promote and maintain good health for themselves, their families, and communities. This concept has evolved to include the skills necessary to navigate complex healthcare systems and make informed decisions.

What We Are Looking For

• Basic Small Language Model Chat Functionality: The solution should enable users to ask questions about medical knowledge and receive accurate, easy-to-understand responses, facilitating quick access to essential health information.

• User-Friendly Interface: The system should be intuitive and accessible, ensuring that users of all ages and technical abilities can easily interact with the platform to obtain the information they need.

• Trustworthy Information Sources: The solution must ensure that all information provided is sourced from credible and reliable medical databases, maintaining the integrity and accuracy of the health insights shared with users with proper references either via hyperlinking or footnotes. The preferred databases will be provided by TGLH and their partners, but the solution should be able to tap on other sources as well.

• Data Analytics: Be able to track key user engagement metrics, including login frequency and generate word clouds of common queries, providing insights into user behaviour and information needs. The dashboard should also be able to show the breakdown of the answer’s source based on the databases (i.e. databases provided by TGLH vs others).

Learn More

Challenge Owner(s)Green Link Digital Bank
Industry Types(s)
Digital/ICT, Financial Services

ENHANCED EFFICIENCY AND ACCURACY THROUGH SMART REGULATORY REPORTING

As a financial institution regulated by the Monetary Authority of Singapore (MAS), GLDB is required to submit regulatory reports on a regular basis. These reports encompass various financial figures including the bank’s assets, liabilities, and risk exposures. To ensure compliance with MAS regulation, GLDB currently dedicates a few staff members to generate these reports daily, allowing for accurate reporting and monitoring when needed.

What We Are Looking For

• Automated data collection and summarisation. Identify and summarise data requirements for each regulatory report (e.g. MAS Notice 610), automatically source required data from multiple sources (e.g., core banking, internet banking, credit processing, accounting systems) and handle complex computations and ratios across various data points.

• Intelligent data validation and cross-checking. Cross-validate figures across different data sources and reports. Detect anomaly by analysing historical data and identifying deviations from established trends.

• Automated data validation reporting. Generate daily reports that flag discrepancies, identify missing data sources, and highlight erroneous fields across regulatory reports.

 

• Smart recommendations and adjustment tracking. Ability to highlight potential discrepancies, providing explanations for detected errors, and suggesting corrected values.

• Review and approval workflow. Allow analysts to review system-generated recommendations and approve changes.

• Cross-validation of data. Ability to recognise relationships between fields within reports and cross-check figures across various data sources. It should retrieve relevant information (e.g., financial reports), conduct independent computations, and validate accuracy to ensure data consistency.

• Automated adjustment across all reports. Apply adjustments across all relevant reports, highlight required data points, keep track of all adjustments.

Learn More

Challenge Owner(s)Green Link Digital Bank
Industry Types(s)
Digital/ICT, Financial Services

AUTOMATED COMPLIANCE MATRIX GENERATION IN BANKING

GLDB maintains a compliance framework that maps internal policies and procedures to applicable regulations and standards, covering various areas including information technology systems, transaction monitoring, anti-money laundering policies. This typically involves creation of compliance matrices to cross-reference requirements with internal controls. Audits are usually conducted annually to assess whether processes and policies are adhered to and effective. The compliance team works closely with individual policy owners within GLDB for these audits which are usually carried out annually. 

What We Are Looking For

• Intelligent policy and regulatory mapping. Ability to crawl, ingest and analyse data from different sources including internal policies, external regulatory databases and documents on best practices. Provide alerts and recommendations for policy updates based on new or revised regulations.

• Automated policy mapping and gap analysis. Ability to automatically map internal banking
policies and procedures against relevant regulatory requirements so that potential gaps or

inconsistencies can be identified. 

• Customisable compliance matrices. The solution should enable the generation of customized compliance matrices based on different regulatory, business unit requirements and risk appetites.

• Proactive risk insights and mitigation. The solution should suggest actionable mitigation steps to address identified gaps and insights into how policy changes may potentially affect operations, regulatory compliance and the risks involved.

Learn More

Challenge Owner(s)Ministry of Education (MOE)
Industry Types(s)
Digital/ICT, Education Services

ENHANCING TOILET CLEANING OPERATIONS IN SCHOOLS

For schools operated by MOE, cleaning is mainly done by cleaning service providers. MOE’s outcome- based cleaning contracts include requirements on areas to be cleaned in schools, and the prescribed frequency of cleaning of selected areas based on NEA’s requirements. Service providers are assessed on their performance via unscheduled audits twice a week, where auditors will assess and score the cleanliness of the various areas in the school on a scorecard based on MOE’s contract requirements, with 70% as a passing score.

What We Are Looking For

• Automated cleaning of toilet bowls/urinals. Able to autonomously clean toilet bowls and urinals of different heights, and address cleaning requirements (e.g. removing stains and residues, disinfection etc.) thoroughly with minimal manual intervention.

• Situational awareness for obstacle detection and avoidance. Able to detect and navigate around obstacles such as people, toilet stall walls and other objects to avoid collisions within confined toilet spaces, and to adjust movement pathways dynamically based on real-time environmental conditions.

• Automated task scheduling and alerts. Able to automatically schedule cleaning tasks including during night-time, with real-time notifications and alerts for status updates or when human intervention is required.

• Manual mode. Allow manual overrides and inputs to cleaning modes/scheduled tasks to address specific cleaning requirements.

• Additional cleaning capabilities (optional). Solutions that cater to more areas of the toilet (e.g. mirrors, floors, sinks) will be advantageously considered.

Learn More