Artificial intelligence has risen to be a top-of-mind topic for many business and IT executives, and we can expect even more interest focused around it this year. But the hype and the reality of AI do not always match. Here are a few trends that we can expect to see in the areas of analytics and AI.
AI will go deep into the “trough of disillusionment”
Gartner’s Hype Cycle describes the journey that many new technologies will experience, looking at maturity and adoption of technologies and applications and their relevance to solving real business problems and uncovering new opportunities – through five phases from initial invention to long-term adoption.
Artificial intelligence, like many new technologies, will enter the “trough of disillusionment” - where real-world deployments and testing projects encounter problems, and where the current capabilities of existing tools do not align with described benefits.
While the most prominent users of these technologies, like Facebook and Google, will continue to forge on, making AI look easy, the rest of the world will find it difficult to develop more than the most basic machine learning projects. As such, the companies without the deep experience of AI and the talent to support it, simply won’t see the same results as the larger players.
The impact of this could lead to a wide polarization around AI, where the leaders will generate massive amounts of value for themselves while those unable to capitalize on the benefits of AI, will be left struggling – making the role of the open source community around machine learning libraries critical.
Stream processing will become “business as usual”
Using application components with limited instruction sets to act on data as it is produced, or Stream Processing, will become more commonplace – over more traditional batch processing and analytics. This near real-time processing will become more and more important for application development teams, as stream processing offers that flexibility and faster response on data. Open source projects like Apache Kafka and Apache Spark will continue to grow.
We’ll see stream processing get built into standard backend and operational databases, making it easier to implement it at scale, while still supporting the storage and management of data over time.
The fear of lock-in will be more endemic
Even as concerns around data security of cloud platforms reduce, accidents and misconfigurations will continue to occur – but these are more user error than the fault of the cloud vendors themselves.
Instead, as more companies move to the cloud, the fear around cloud will shift more so to IT infrastructure and architecture strategies becoming reliant on these big public cloud providers. CIOs will see cost increases from their providers, and must enter into this with the understanding that any commitment may lead to lock-in.
In response to this, 2018 will be the year that CIOs fully embrace multi-cloud strategies. 2017 saw many start that journey and this is just a logical expansion of that thinking. Running across multiple cloud vendors will deliver the scale and flexibility benefits that companies are looking for, without being tied to the vision and costs of just one company. If a vendor starts charging more, then data can be transferred to other providers or back to on-premises, internal IT, depending on the most appropriate economic models.
CIOs will look to keep their providers honest as competition heats up by employing a multi-cloud strategy rather than being tied to single cloud platforms. Balancing the mix of Service Level Agreements, ongoing bills and data management and distribution across multiple cloud platforms, while still serving the same applications will be an Enterprise Architect’s dream come true.
Graph will mature and expand at scale
Graph databases will continue to grow and mature in 2018, from a niche set of specific implementations around data interrogation and relationship modelling, through to powering significant and web-scale applications.
Businesses can use graph to better know their customers and the relationships between them. By modelling family relationships within accounts, companies can prevent slips that otherwise affect customer experience. But this ability to track customer data and records over time should be even more useful as we approach the date for compliance with the General Data Protection Regulation (GDPR) in May 2018.
Digital transformation will be less about cloud, more about competition
In the 1980s, US towns feared the Walmart Effect, where the area’s existing small businesses were decimated by the arrival of cheap, big-box supermarkets. Today, there is a similar fear growing around the digital equivalent, the Amazon Effect.
With this in mind, businesses recognize and will need to be more cognizant of how important company data is when it comes to digital transformation and competition. Rather than simply moving to cloud platforms - cloud infrastructure run by the very companies that could pivot or extend their market reach by becoming competition - traditional enterprises will need to and are looking at how to maintain "Data Autonomy."
This fear of multi-faceted competition will become the main driver for large digital transformation projects in 2018, as those businesses want to roll out more modern applications while still keeping control over their data. As an example, in 2017 Walmart restricted anyone doing business in thier online marketplace from using anything in Amazon Web Services. Target made the decision to migrate out out AWS citing competition. This will continue.
In 2018, more and more brands will pursue data autonomy through a multi-cloud strategy in order to compete and stay ahead of those big public cloud vendors – and the need and urgency to challenge these big cloud players head on with data-driven applications will intensify.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access