The Challenges of IoT and Cognitive Projects
Although experience of IoT and cognitive tools is expanding, there are significant obstacles in both development and implementation. A development methodology must respect the dynamics and the nature of the technological challenge. Cognitive projects use technologies that may be less well known, such as network databases and programming languages with specific kinds of function libraries, meaning that resources with the right skills may be harder to find. Deployment and implementation is more likely to face the ‘So What’ and ‘Ah but’ challenges of business stakeholders to avoid becoming seen as a ‘hammer looking for a nail’ with unpredictable business benefits.
Feedback from early projects indicates:
- difficulty in integrating cognitive projects into existing systems and processes
- perceptions of high costs of the technologies and expertise
- poor understanding of the technologies and their capability by business managers
- a lack of skilled resources available
- the technologies remain immature
- overselling and hype from vendors is rampant
A Framework for Integrating Cognitive Technologies
Whilst the technologies and their deployment may be relatively new, the IT industry has developed in its approaches and techniques to manage uncertainty. We recommend a four stage approach, embodied in the diagram below:
- Understand the technologies. Not simply a case of understanding what they do, companies should seek to understand which technologies are most likely to apply to their business issues, and from which they can derive ROI. Your business case will likely arise in those areas where you have business issues or poor performance, so look then to appropriate technologies and the skills required to use them. IT functions need to maintain a constant programme of research and education, and may need to boost their capability in key functions such as data scientists with statistical understanding
- Create a portfolio of projects, based on a systematic evaluation of needs and capabilities. Identify the opportunity areas – areas of bottlenecks in processes caused by poor availability of data, challenges in managing processes at scale, inadequate levels of resource to handle workload – where cognitive projects might help. Pick specific use cases, and identify individual technologies required.
- Use pilot projects. Proof of concept pilots are ideally suited where the business opportunity has a strong ROI, but the technology capability is a bit of an unknown. Consider using a Centre of Excellence or similar structure for cognitive projects, to pool or optimise scarce resource across business units or functions.
- Consider Scale up as an exercise in itself. Post pilot, deployment of new functions into the business must integrate into existing systems and processes or risk being ignored. The change management challenge is higher than for example large-scale ERP implementation, especially with the often unspoken threat of machines ‘replacing’ humans.
A Proven and Practical Development Methodology
Working with our partners in IoT Advisory Consulting, we use a proven framework from over 40 engagements to ensure benefits are realised. This is based in IoT projects, and takes advantage of concepts like Design Thinking, MVP (Minimum Viable Product) and Agile delivery.
Augmented by governance, programme management and a continuous improvement process to rapidly build from MVP to fully functional system.