Tackling the challenges of sustainable growth with data and AI

Considerations in implementing AI-powered prediction, optimization and automation

By | 3 minute read | October 13, 2021

Automated irrigation


Now in its fourth year, the Call for Code Initiative has grown to more than 400,000 developers and problem solvers across 179 nations, and it has generated more than 15,000 applications. Call for Code aims to drive immediate and lasting humanitarian progress through the creation of practical applications built on open source-powered software. As part of the Call for Code initiative, we introduced and completed the AI Spot Challenge for a safer planet this summer, with the goals of exploring solutions using AI-powered prediction, optimization and automation to help mitigate both physical and cyber threats and risks.

The winning solutions and other submissions addressed a spectrum of challenges in improving worker safety, living conditions and health with applications of artificial intelligence (AI) and adjacent technologies, such as IoT and augmented reality. One of the crucial aspects of a solution was to design an end-to-end approach from ideas to AI-enabled apps even at the onset. This blog post will discuss insights on addressing unmet needs, exploring solutions, bringing teams together, learning about the enabling technologies and deploying AI projects.

Exploring unmet market needs and driving value

One of the most frequently cited challenges in building an AI solution is to identify problems that can be solved and prioritize them by value, effort and risk — with a focus on user outcomes. Enterprise Design Thinking is a framework to solve our users’ problems at the speed and scale of the modern enterprise. Maintaining a human-centered approach to AI solutions is as vital as ever. Augmenting and improving existing processes and tasks requires continuous iterations by empowered, diverse teams that can turn abstract ideas into concrete solutions rooted in user outcomes.

User value also needs to be measured by broader metrics, moving beyond the traditional metrics to include sustainability. By the end of 2021, 9 out of 10 companies surveyed say they will be working on various sustainability initiatives across the enterprise, according to IBM Institute of Business Value. Hybrid cloud and AI architecture can become part of your strategic arsenal for suitability initiatives. Turning data from disparate sources into outcomes also requires interoperability, ease of exploiting the open source ecosystems and access to relevant data across any cloud.

Bringing teams together for sustainable future

A common hurdle in building AI solutions is the investment in identifying, building, enabling and motivating teams to execute ideation, business case development, design and architecture and then operationalizing the solutions. People with diverse backgrounds and skills, when empowered, can rapidly refine solutions, find other collaborators and learn from each other to validate and increase benefits toward sustainable competitive advantage.

Developers, data scientists, analysts and subject matter experts join all other problem solvers who are eager to share ideas and learn new AI capabilities. Unifying tools, processes and talent not only helps project initiation and deployment but also value optimization. One of the key areas that warrants attention is AI testing ‚ what’s considered the “last mile of AI.” Continuous testing and delivery of models and software pipelines demand alignment and integration of AI lifecycle management with DevOps.

Deploying AI for sustainable, profitable growth

Sustainability is a strategic mandate of our time. Organizations must meet the new levels of resilience and agility to preserve our planet and generate profitable growth to reshape ways forward, fuel innovations and adapt better ways of working together. Yet, the challenges that AI deployments can create and the time it takes to get a gold model into production are still viewed as an area of improvement.

A recent ESG research uncovered that only 1% of organizations go from a trained model to production in under 10 days. 96% of respondents indicated that their organization typically takes between 11 to 30 days to go from a trained model to deploying into production. Having this in mind, ESG has completed a technical validation of IBM Watson Studio on IBM Cloud Pak® for Data and noted that “ESG can confidently recommend serious consideration of this data and AI platform if your organization is looking for an end-to-end platform that enables applied learning from production to quickly iterate optimized models while ensuring visibility across data science, application development, and business teams.” We recommend you review the assets below as you envision a sustainable, profitable path forward with open, AI-powered solutions.

Next steps