Generative AI: Technology matters, but adoption matters more
After bursting onto the scene almost two years ago, there are now whispers of a Generative AI bubble — triggered by overvaluation, hype, and the growing gap between expectations and reality. Some investors have grown impatient after realizing that adoption is lagging, and the big payoff is much further down the line than they initially thought.
Regardless, many of the titans of tech Innovation have expressed confidence about the once-in-a-lifetime transformative nature of GenAI. Mark Zuckerberg emphasized the importance of building too fast “rather than too late,” to avoid falling behind in the AI race.
Similarly, Google’s Sundar Pichai stressed to shareholders and analysts that the risk of underinvesting is dramatically greater than the risk of overinvesting. Most investments have been funneled into three areas so far:
- Hardware (AI chips in particular)
- Software (primarily in the form of Large Language Models)
- Data center space expansion
Clearly, Big Tech is taking the long view, but while these investments are necessary, they will not be the drivers of widespread Generative AI adoption.
With Gen AI, the best results will not be achieved by the most powerful platform. It takes a complete approach that considers context, goals and constraints — enabling strategy and culture to evolve together.
Based on our experience working with clients on numerous GenAI initiatives, what we have observed is that these projects so far have had relatively small budgets, as large enterprises wait to see demonstrable ROI before investing further. In my view, this will not change any time soon. Here’s why:
- The AI market is still nascent, with many enterprises experimenting with pilots and proofs of concept that are naturally limited in scope and budget.
- Generative AI adoption is driven by individual use cases rather than a broad, mass-market product or service. Despite the hype, GenAI is not an iPhone.
- Projects are relatively high-touch, because they require high levels of customization to adapt the tech to customer-specific context and unique needs.
- Similar to cloud, GenAI will be gradually integrated with existing processes and workflows. However, given the deeper level of integration required with people and process, it may take even longer to become mainstream.
- There is still reluctance based on the well-documented ethical, technical, sustainability and privacy challenges — as well as the unsavory possibility of displacing human workers.
- Given today’s macroeconomic uncertainty, enterprises want “quick wins” with immediate impact and fast ROI — neither of which are certain with GenAI.
So, what to do? Invest now and take the short-term hit in hopes of an exponential return down the road? Buck the advice of the Big Tech leaders and take a “wait and see” approach?
More broadly, will GenAI deliver on any of the eye-popping projections of its influence on the global economy?
It’s hard to predict whether GenAI will live up to this hype, but there are a few key factors that could help the cause:
1. Convergence & scale
It may only be possible to reach the full potential of GenAI by the aggregation of numerous small-scale projects across different industries and use cases, to generate a significant overall market.
2. Holistic investment
Sustained investments will be required, but not just in hardware and software. Equal attention must be directed to creating adoption levers through education and skill development, tax incentives or grants for companies adopting GenAI, more VC funding for AI start-ups, or even consortiums made up of competitive GenAI players.
3. Patience
We urgently need a reset of expectations across the board, among enterprises, technology providers, analysts and investors. GenAI may feel like magic at times, but it's not. Nor is it an overnight miracle — if you believe in that sort of thing. GenAI is a tool (a powerful one), but it should not be forced. Using it properly takes work, careful planning and attention to details. As such, it’s difficult to overstate the importance of patience in this process.
At Atos, we have created numerous proofs of concept across multiple industries and use cases, and we firmly believe that GenAI is truly a game changer. We are no longer seeing just potential, but actual results. However, we firmly believe that the future isn't just about technological advancements — it's equally about ubiquitously embedding it into everyday business operations to create real and widespread impact.
When it comes to AI adoption, the old business maxim that “culture eats strategy for breakfast” still holds true. Despite the fact that GenAI is now a board topic with significant top-down push, an organization’s culture is more likely to determine the speed, success or failure of adoption. In other words, despite careful planning and due diligence, your company’s culture is more likely to make or break an AI initiative than the technology itself.
The critical adoption challenges are resistance to change, risk aversion, siloed structures and a lack of GenAI literacy that leads to a fear of the unknown. Overcoming these challenges requires a culture change which will occur gradually and at varying speeds across different organizations.
This is why we believe the best results will not be achieved with the glitziest, most powerful platform. It takes a complete ecosystem that includes solution accelerators, consultancy services, industry-specific knowledge, and behavioral change management to overcome AI adoption challenges. The right approach to delivering effective GenAI is co-creation between enterprises and their technology partners, taking into consideration the proper context, business goals and constraints — enabling strategy and culture to evolve together.
Posted on: October 11, 2024
Adil Tahiri
Head of CTO and Client Advisory GroupMember, Atos Research Community
View detailsof Adil Tahiri>
