Editor’s note: The following is a guest post from Doug Ross, US head of generative AI at Capgemini Americas.
Generative AI solutions made a splash in the technology world, with more than half of global consumers tracking the latest generative AI trends and even testing tools themselves. Rising consumer interest accelerated generative AI strategies and use cases for enterprises.
Though the technology is still in its nascent stages, Capgemini research shows 80% of organizations have increased their investment in generative AI since 2023.
The report also found nearly one-quarter of organizations are now integrating generative AI into some or most of their locations or functions and 54% of companies are allowing their employees to use generative AI tools but with certain restrictions and principles in place.
Despite somewhat conservative projections during the early stages of generative AI, businesses should actually be motivated to begin their plans while the technology is in its infancy. Leaders who wait even six months may be markedly behind early adopters that were able to safely mitigate risks and increase their ROI.
Before kickstarting implementations, enterprises across industries should explore four pivotal generative AI guardrails during planning discussions and determine how to incorporate them into their strategies accordingly:
1. Regulations:
It is important for businesses to assess the current, and potential, regulatory landscape across the geographies in which they operate during their initial planning stages.
Although there are few generative AI-specific regulations right now, the public sector is preparing to introduce policies that will regulate the broader AI and ML landscape. Colorado’s SB21-169 law and the EU’s AI Act serve as precursors to the future of generative AI regulations.
Businesses should partner with informed stakeholders to ensure solutions and strategies meet the concerns of internal, external and legislative audits.
2. Talent:
As new generative AI solutions enter the market, companies will require new skill sets and roles, according to 69% of business leaders. This is particularly true across administrative, business and design teams. Employees may need to be reskilled and upskilled accordingly. Leaders need to plan for this huge shift in talent with actionable tactics. Centers of excellence and change management processes can help manage the shift effectively.
3. Models:
Generative AI models aren’t the only models to contemplate when developing strategies. There are also benefits to outlining new corporate operating and reporting models to safely scale generative AI solutions.
Like any other tool or development within an enterprise’s overarching digital transformation agenda, generative AI — and the corresponding tactics — must align with companywide processes and practices.
4. Ethical and explainable AI
In contrast to some AI or ML solutions, generative AI will require organizations to utilize both ethical and explainable AI tools and frameworks. This will be essential as companies offer explainable and de-biased generative AI models.
Ethical AI frameworks will involve introducing accessibility best practices as well as model testing, monitoring, validation and tracking to meet ethical AI requirements.
Explainable AI approaches will require tools and practices to determine differential privacy, AI organizational maturity assessments and pre-processing optimization. Leaders will need to set precedence and introduce a blend of ethical and explainable AI processes and protocols to securely introduce generative AI solutions to the market.
Initial tactics
Once companies examine these guardrails, they’ll have a more complete understanding of their business readiness and an appreciation of the complexities that surround generative AI.
These considerations should shape their initial plans and continuously guide their strategies as the market develops. After leaders have thoroughly assessed generative AI guardrails, they can begin to scale implementations by:
- Understanding where generative AI solutions can thrive and where certain models are less applicable
- Mapping use cases to the firm’s key business differentiators
- Determining ROI for scaling and streamlining the predetermined use cases and functions
- Folding applicable regulations or guidance into risk and governance operating models
Generative AI isn’t a passing fad. While leaders should carefully weigh risks against rewards during the nascent stages, that doesn’t mean they have to take a passive approach to generative AI implementations. Now is the time for enterprises to actively form their strategies, using guardrails to safely shape their plans.
