Page 20 - Logistics News - Issue 02 - 2024

AI

But collaboration and data sharing are key. While this will 7. Data democracy and the end-to-end story
appease consumers’ tendency to trust people more than data, Innovation often comes from a convergence between
generative AI will also allow them to explore freely and interact unexpected parts of the organisation. More powerful AI and
conversationally within the systems where they operate. automation capabilities can remove the need for advanced
knowledge and tools that can shift a new breed of users out of
3. Curating knowledge from unstructured data their comfort zones.
Generative AI techniques will allow organisations to unlock
the potential and value of unstructured data. In fact, Forrester By merging the roles and capabilities of data engineering,
estimates that 80 percent of the world’s data is unstructured. data science and analysis, for example, organisations can solve
Generative AI models, however, adapt to different data types and tougher problems with fewer resources. This also eliminates
do not rely on predefined rules. barriers between previously siloed functions to solve business
problems.
As such, the opportunities for combining structured and
unstructured data in a trusted way will be endless; for example, 8. From analysis to execution
using chatbots like ChatGPT for external use cases alongside Connecting generative AI to automation shifts human focus to
more private chatbots where trusted enterprise data has been action. As data can be transformed in near-real time and in the
secured. right place, new ways of using generative AI will emerge.

4. From BI to AI, and back again This reduces the burden of manual work, such as building
Generative AI is bringing about a change in the way businesses workflows, and instead allows humans to focus on decision-
analyse their information. It can, for example, take advantage of making processes. However, automate incrementally as
solutions that place analytics needs where and how people work comfort grows by monitoring progress, adjusting and then
and provide interactive analyses within workflows. automating more. And always keep humans in the loop.

Individuals might start their analytical journey in generative 9. AI customisation
AI tools, using them for simple data visualisation and business Applications of generative AI are currently generic. Over time,
projections. As a next step, they may want to tap into enterprise- however, we’ll see more customisation around industry and
grade tools for deeper analysis, bringing the benefits of business-to-business (B2B) use cases, with multi-cloud
generative AI to their trusted tools. environments providing greater efficiency and stability.

5. Data quality matters even more With less effort and hours, as well as the ability to build
While data quality and lineage mattered in the past, they have customisations on a common foundation, greater speed and
become non-negotiable in the AI world. Companies must be efficiencies can be realised to address specific industry needs.
able to create the equivalent of a ‘DNA test’ for their data to trust
its origin. 10. The data marketplace
We can apply the same principles of product management to
Solutions are emerging that offer pervasive data quality data by asking questions about what problems we are solving,
across platforms, the ability to continuously observe and quickly best use-case scenarios and by whom. This will emphasise the
quantify data quality across datasets, data transformation and importance of data quality, governance and usability for end
cleansing functions, and the ability to track data flows from users.
source to use.
Only then can we start transforming data assets into
6. AI bots must pass a ‘driver’s test’ products for reuse internally as well as externally. In this way, our
Just as data literacy has been crucial in the last few years, most valuable asset can also become a tradeable good.
we now need to turn our attention to AI literacy to improve
standards, avoid governance chaos and application glut In conclusion, our focus should always be on how well
from ‘everyday’ developers. Power has shifted into the trust, security and regulatory pressure hold up in a new data
hands of the many and, as such, organisations must paradigm. Better, more trusted data will ensure that the output
take steps to educate their workforce on the benefits and of generative AI is not compromised by a lack of data traceability
pitfalls of generative AI. and quality. •

18 NO V E MB E R/D E C E MB E R 2024 www .l o g ist i csn e w s .c o .z a