Once a niche technology, generative AI is rapidly becoming a mainstream tool, offering unprecedented opportunities for innovation. Rooted in technology dating back to the 1950s, it has the hallmarks of a solid foundation, and companies are seizing the opportunity while it is still advantageous.[1]
At Avasant’s Empowering Beyond Summit, Chandrika Dutt, Associate Research Director at Avasant, and Swapnil Bhatnagar, Partner at Avasant, delivered a presentation titled “Research Insights – Enterprise AI Trends.” They highlighted the lasting importance of generative AI, the enterprise shift towards revenue-impacting projects, and the key considerations for enterprises when leveraging generative AI.
Generative AI is Here to Stay
Generative AI’s trajectory stands out from other technological trends, which points towards its long-lasting potential. Bhatnagar identifies two main factors that differentiate generative AI from other innovations like blockchain: the rapid development of regulations around emerging technology and its democratic adoption. He argues that the regulatory environment around blockchain could never keep pace with its technological advancements. In contrast, there is growing interest in establishing a comprehensive regulatory framework for generative AI. The EU has set a strong governance baseline with its GDPR framework, although most AI development is happening in America.[2] In 2023, the US introduced the bipartisan CREATE AI Act to increase access to AI tools, along with President Biden’s Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence.[3] Though there is still progress to be made, the fact that regulations are being considered and are attempting to keep pace with AI innovations points to a widespread belief that generative AI is here to stay.
The democratic adoption of generative AI is unique. Bhatnagar explains that development in generative AI is not being driven by a few major technology companies with specific incentives. Rather, it is emerging from various sources across the board. The advent of smaller AI models that can be deployed at lower costs, with less hardware, and with the potential to run locally have paved the way for the democratization of AI.[4]
The shift toward smaller models has been driven by a decrease in hardware availability combined with the rise of models being made more accessible through their utilization of Low Rank Adaptation (LoRA) and Quantization.[5] LoRA simplifies the customization of AI models. It allows users to freeze most of a model’s parameters and insert their own trainable layers without having to update the massive number of parameters already in place. This approach “dramatically reduces the number of parameters that need to be updated, which, in turn, dramatically speeds up fine-tuning and reduces the memory needed to store model updates.”[6] Quantization decreases users’ memory needs and delivers faster results. It is like lowering a video’s quality to decrease its file size and increase speed. Reducing precision, but not significantly enough to affect the quality of the output, leads to quicker results.[7]
Revenue-Impacting Projects
So far, most of the focus on AI’s business potential has been on its ability to increase efficiency and as a byproduct, cut costs. Enterprises have begun to shift away from this and towards leveraging generative AI for revenue generation because while there is a limit to cost cutting, growing revenues provide unlimited potential.[8] Dutt identifies three ways for companies to go about this shift: monetizing their data, launching new product features as a service, and tapping into new customer bases. Where this shift has found its highest potential for impact has been in revenue operations. Requiring the collaboration between the business and technical sides of a company, generative AI solves one of the most challenging issues faced by the revenue operations wing of companies: managing data use across different company areas. [9] Once data is managed efficiently, enterprises can leverage revenue-impacting uses of generative AI, namely better customer targeting and acceleration of research and development projects. L’Oreal developed TrendSpotter, an AI tool that collects data from various sources to predict beauty trends coming in the next 6 to 18 months.[10] Accurate trend detecting then impacts current business decisions, hopefully leading to increased revenue.
Generative AI for revenue-impacting projects also creates opportunities for technology service providers. As customers can adopt generative AI to tackle workflow and automation challenges, which they previously relied upon service providers to do, adoption of generative AI will lead to a decrease in the demand for service providers’ traditional services.[11] With this new technology also comes rising demand for new services, and if service providers can meet this need and expand their service offerings, they can profit from the “emerging market for services in relation to genAI/AI which could be worth more than $200 billion by 2029.”[12] Accenture is a technology service provider that has responded to this change. Building upon a 20-year relationship with Adobe, Accenture is integrating Adobe Firefly Custom Models, Adobe’s generative AI model, into its marketing offerings.[13] By following this example, service providers can ensure that they are solidifying their share of the emerging market.
Key Considerations
Choosing to adopt generative AI raises a series of crucial questions. According to Bhatnagar, enterprise customers commonly focus on five key question areas: cost versus performance, on-premises versus cloud deployment, compute power versus sustainability, enterprise versus provider accountability, and speed versus responsibility. Whether they are asking how to ensure accountability or if revenue is being generated in a responsible manner, at their core, these questions are rooted in concern about the potential risks of generative AI. Chief among these is the risk posed to cybersecurity.
Generative AI presents both a challenge and an opportunity for cybersecurity. As algorithms become more sophisticated, so do the threats, with incidents like a deepfake scam that defrauded a multinational firm’s Hong Kong office of $25 million.[14] The number of threats will only increase as open-sourced AI models become more accessible. Abandoning generative AI is not the answer as its threat potential is already out there. CrowdStrike, a leading company in using AI to counter cybersecurity threats, believes that the most significant advantage of generative AI is in the shift from reactive to proactive cybersecurity.[15]
Traditional cybersecurity tools rely on signature detection, where systems analyze previous attacks to develop signatures that identify similar threats.[16] When threats emerge, systems seek to match signatures, with the signature determining how the threat should be addressed. This approach works well until it must deal with advanced threats built with generative AI. In these cases, signature detection fails “to detect the transforming and fast-evolving AI-powered cyber exploits,” particularly because it fails to address real-time issues that lack a common signature.[17] This is where generative AI comes into play. Rather than relying on signature-based detection, AI-based tools can create a baseline detection system which establishes normal trend baselines. If any anomalies are identified, appropriate measures are taken. This way, AI tools can identify and defend against new, previously unseen zero-day threats.[18] When generative AI’s ability to forecast, address issues, and automate threat detection are added to its baseline detection, it can meet the challenge posed by generative AI-based cybersecurity threats.[19]
The adoption of generative AI in enterprise operations marks a pivotal transformation bringing with it both opportunities and challenges. Unlike previous technological trends, generative AI has demonstrated unique staying power and potential, as noted by Dutt and Bhatnagar. While initially leveraged for cost reduction and efficiency gains, businesses are now exploring AI’s vast capabilities for revenue generation. As generative AI evolves, it simultaneously amplifies cybersecurity risks while also providing innovative solutions to combat them. The technology has demonstrated itself to be a game-changing tool, likely to stick around in the near future.
References
[1] History of AI: Timeline and the future. Maryville University Online. (2023, October 27). https://online.maryville.edu/blog/history-of-ai/#:~:text=However%2C%20the%20possibility%20came%20to,the%20Turing%20Test%20in%201950.
[2] Pandey, H., Kaka, N., Jain, P., Muthiah, S., Daga, V., & Dasgupta, R. (2024, June 12). Tech services and Generative AI: Plotting the necessary reinvention. McKinsey & Company. https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/tech-services-and-generative-ai-plotting-the-necessary-reinvention
[3] Lynch, S. (n.d.). What to expect in AI in 2024. Stanford HAI. https://hai.stanford.edu/news/what-expect-ai-2024
[4] Bergmann, D. (2024, July 2). The Top Artificial Intelligence Trends. IBM. https://www.ibm.com/think/insights/artificial-intelligence-trends
[5] Ibid.
[6] Ibid.
[7] Ibid.
[8] Korolov, M. (2024, April 24). How cios align with cfos to build RevOps. CIO. https://www.cio.com/article/2094438/how-cios-align-with-cfos-to-build-revops.html#:~:text=The%20first%20use%20of%20generative,can%20see%20an%20unlimited%20upside.
[9]Hoskatti, T. (2024, July 12). Revolutionizing revenue operations with Generative AI. Ambit Software. https://www.ambitsoftware.com/revolutionizing-revenue-operations-with-generative-ai/
[10] Singh, G. (2024, August 7). How big companies use generative AI to grow in 2024. Webspero Solutions. https://www.webspero.com/blog/how-big-companies-use-generative-ai-to-grow/
[11] Pandey et al. (n 1).
[12] Ibid.
[13] Accenture and adobe to co-develop industry-specific generative AI solutions to accelerate marketing transformation. Newsroom. (n.d.). https://newsroom.accenture.com/news/2024/accenture-and-adobe-to-co-develop-industry-specific-generative-ai-solutions-to-accelerate-marketing-transformation
[14] Cybersecurity in the Gen Ai Era. (n.d.). https://www.accenture.com/content/dam/accenture/final/accenture-com/document-2/Accenture-Cybersecurity-In-The-Generative-AI-Era.pdf
[15] Stanham, L. (2024, February 21). Generative AI (genai) and its impact in cybersecurity – crowdstrike. crowdstrike.com. https://www.crowdstrike.com/cybersecurity-101/secops/generative-ai/
[16] Muraleedhara, P. (n.d.). The need for AI-powered cybersecurity to tackle AI-driven cyberattacks. ISACA. https://www.isaca.org/resources/news-and-trends/isaca-now-blog/2024/the-need-for-ai-powered-cybersecurity-to-tackle-ai-driven-cyberattacks
[17] Ibid.
[18] Ibid.
[19] Ibid.
By Adriana Guzman, Associate Consultant