ABSTRACT: Trends in Information Technology 2024-2025
Most enterprises are actively employing or seeking to employ artificial intelligence tools such as OpenAI’s ChatGPT, and freely mix commercially available services such as AWS and Azure with their other IT infrastructures. Usage of these products and services is accelerating, a trend that is set to continue into 2025, creating opportunities and challenges.
The opportunities vary by industry and market segment, but improved customer interaction and generating new sources of revenue are typically the most urgent and important, a trend further enhanced (and complicated) by the rapidly growing use of artificial intelligence. The adoption of AI is already heightening customer expectations and radically altering customer interactions, while cloud services allow far greater flexibility in deployment, especially in high-growth environments. These opportunities are exciting and offer tremendous possibilities for revenue and market-share gain but vary tremendously due to the specific market, financial and investment situation of each business, as well as their cultural predisposition to risk.
Unfortunately, the use of AI raises significant governance and regulatory concerns, which require careful, considered analysis and planning before projects commence. Even if these factors favour early adoption, the practical difficulties in rapidly implementing the appropriate systems remain. They include:
· a shortage of skilled, knowledgeable staff,
· provision of adequate, truly up-to-date training and development,
· enabling secure access to multiple systems and databases,
· ensuring secured access to multiple locations, including home offices,
· spiralling, unanticipated costs, especially from cloud service suppliers, and
· legal claims (especially relating to the use of AI) running into millions of dollars.
While the opportunities vary greatly in nature, these challenges are surprisingly consistent across all sectors, with a shortage of skilled staff, increasing internal and external customer demand, shortening deployment timescales, an increasing prevalence of price increases from near-monopoly suppliers of proprietary technologies and services, physical and cyber-security breaches, and compliance breaches being most frequently quoted. The advent of new, more capable artificial intelligence (AI) engines further complicates this landscape, with legal claims running into millions of dollars already increasing the risk profile for all businesses. Worse, these factors are all causally interlinked.
Potential solutions to these specific issues include:
Outsourcing projects or specific tasks or augmenting existing staff with contractors or consultants are the obvious choices. Each presents separate challenges, but making the correct choice will reduce timescales, provide ‘on-the-job’ and experiential training, and hopefully avoid any embedded resistance to novel approaches. The results have greatly exceeded those from external skill-based training courses, where training tends to lag technology trends, provided that an over-reliance on a single, large provider is avoided. The result of such dependence is obvious from the current Post Office/Fujitsu debacle - to the detriment of both parties. The outcomes from augmenting or outsourcing should be (a) greatly enhanced performance and cost efficiencies, and (b) improved compliance and security. Recent reports show that skill shortages led to 43% of organisations falling behind on both compliance and security in the last year.
Enterprises are increasingly dependent upon their prowess in managing and securing both physical and digital perimeters. Increasing system complexities and physical sprawl increase the potential impact of their interaction, with potentially frightening results. Employing an external security specialist to review all aspects of security operations – human, location, and network - is the optimal way to ensure they are effectively managed. Such a review should include detailed consideration of human limitations, poor system design or misconfiguration, and the use of new and unfamiliar techniques and architectures, including AI, as well as the more prosaic aspects of physical and personal security such as password management, access control, and so on.
It is widely acknowledged that virtually all AI projects have failed to deliver the anticipated benefits. Worse still, its use has already led to $multi-billion legal claims. While these have to date been mainly restricted to the unauthorized use of proprietary data in training large models, many other categories of claims will likely emerge. An external review comparing the overall goals and benefits of AI usage with the ethical and reputational risks from factors such as inherent bias, inappropriate usage of data, or poorly controlled access to data and results is vital. Employing a project manager with practical experience in implementing a successful AI/machine-learning project is crucial both for setting realistic expectations and for ensuring the target benefits are achieved. These persons are notably rare and thus expensive. Reaching out to the largest consultancies is no help: they often use client projects as training grounds for their less experienced staff, with minimal supervision from over-stretched partners. This situation is doubly difficult for technology firms, creating pressure on payroll and increasing demand on their existing infrastructure.
Over the past year, price increases have averaged between 9-15%. Even where longer-term contracts have been agreed, there still exists considerable risk whenever requirements change or when these contracts expire. The behaviour of cloud infrastructure providers can be unpredictable, resulting in volatile price movements and unpredictable service levels. In addition, system complexity and opacity mixed with a lack of internal cloud supervisory skills means costs could rapidly escalate as usage grows and provide fertile ground for security breaches and regulatory non-compliance. Most cloud vendors cherish their proprietary technology, making interoperability difficult and sometimes impossible to maintain. Our recommendation is always (a) to employ an experienced, independent professional to review and realign plans before commencing any project, and (b) to obtain quotations from a wider range of suppliers rather than rely upon the beneficence of a preferred supplier.
CONCLUSION: DETERMINANTS OF SUCCESS
The ultimate winners will be agile firms that:
(a) Work with those skilled, cost-effective partners who can help them manage, develop, augment, and retain the skillset of your internal teams. Using such partners is consistent with a strategy of retaining ‘in-house’ all core-business expertise and IP, where the partners play a secondary, supportive role. It also allows non-core effort and one-off project work to be readily outsourced to a competent and knowledgeable expert resource.
(b) Select knowledgeable, experienced advisors to review all aspects of security, and then create and implement (often against entrenched opinion) a comprehensive plan to address all foreseeable risks.
(c) Choose an experienced partner who will help your enterprise create a measured and considered approach to the usage of AI. One who will be excited to participate in your adoption journey but who will help you be aware of and include due consideration of the associated, evolving legal and compliance risks.
(d) Adopt a mixed-cloud environment to remain independent of any one provider. This will avoid predatory pricing and be capable of rapidly adapting to changes in the market, or in the way services are provided (even if they are discontinued), thus significantly reducing unforeseen risk. In addition, these businesses will be better placed to take early advantage of evolutions in markets, products, demand levels, or supply constraints.
A subsequent blog will further break down the components of this outline approach, using specific examples of how and where external suppliers are used to support, augment, outsource or replace varying aspects of client systems. For example, augmenting existing teams is most frequently used in the following areas:
o Supporting customers or specific aspects of core applications
o Cyber security & incident response outside of business hours
o Monitoring, alerts & routine maintenance
o Infrastructure change management.
By contrast, outsourcing is most frequently used for developing and maintaining complete, non-core systems.
More than ever before resulting from system failures, the lifetimes of major corporations and the average tenure of individual CEOs have both fallen to historic lows. The stakes are high, and the risk of failure has never been greater.
If you would like to contact us directly to discuss any of the above, or if you would like to receive further information or discuss your situation, please click here: