CEO and Co-Founder at Trinetix.
Generative AI is a deeply complex field, with many players entering the scene to harness the innovation. With such significant figures as PwC and KPMG implementing generative AI tools and the impact of artificial intelligence (AI) on productivity estimated to be around trillions of dollars, many enterprises have started treating the technology as the ultimate problem-solver, a cure to their common pain points. However, are they right? Or is it hype talking?
My personal policy is always to be hype-averse. As CEO of a company that specializes in enhancing workflows with AI-powered capabilities, I’ve been studying and harnessing this innovation for 12 years. I know that there are no silver bullets, no one-for-all solutions—only innovations that later transition into everyday tools offering value if applied correctly.
Generative AI is no different. To understand its potential fully, we must consider its current progress, how it’s being applied within companies and what it lacks to unleash its full potential.
Common Use Cases Of Generative AI
Currently, generative AI technology is commonly used by enterprises to develop the following directions:
• Sales: Generative AI helps sales teams and reps reduce repetitive activities such as generating sales call transcripts, follow-ups and styling content of pitch materials to fit the new customer segment.
• Customer support: AI-powered platforms become powerful and versatile support assistants as they interact with customers, processing their requests, gathering feedback and supplying enterprises with data for personalizing client experience.
• Marketing: Marketing teams use AI-powered tools to accelerate and optimize content generation, for instance, creating outbound email copies, brand-consistent newsletters and blog posts.
• HR automation: Aside from client interactions, generative AI can be used for employee well-being check-ups and organizational process automation (scheduling days off, gathering feedback, etc.).
• Research and documentation: Generative AI is handy with document organization, exploring legal questions and compliance checks, which can create a safer, more controlled environment for enterprises.
Among these top use cases, we can easily notice a pattern of AI being used to manage the bulk of monotonous tasks (with the potential to impact around 60%-70% of employee working time), speed up such enterprise processes as accounting and knowledge gathering and increase the volume of customer engagement content—from marketing to sales engagement. Now, let’s look at how these use cases are implemented across enterprises.
Generative AI For Enterprises: Deployment Models
According to Gartner, enterprises currently use three generative AI deployment models: out-of-the-box foundational models, prompt engineering and custom models. Let’s take a closer look at each model.
1. Foundational Model
I think this deployment model is currently the most popular one, defining the stage of generative AI adoption across enterprises. Within this model, businesses use available generative AI services (like ChatGPT or Bard) for solving their tasks and closing goals. It offers plenty of opportunities for experimenting with AI capabilities and letting employees gain the experience necessary for interacting with AI-powered services.
The downside of such a model is limited control and versatility, which becomes particularly evident when companies want to provide a more in-depth and customizable approach but find out that the service’s features aren’t tuned to their specific needs.
Moreover, with some platforms pushing back against AI-generated content, the need for injecting exclusivity and flexibility when optimizing content generation becomes more and more apparent.
2. Prompt Engineering
The logical next stage is when companies start expanding the basic model boundaries by using tools for customizing prompt inputs and outputs within a foundational AI model. Doing so can yield more accurate results that carry specific value to business plans and vision, allowing for more personalized data input. As noted by Gartner, such a model is also relatively cost-effective, which makes it a viable option for startups and small businesses.
However, it is still more complex than an off-the-shelf solution and needs to be tuned to business systems before it can deliver results, exposing the need for in-house AI engineers or tech partners’ assistance.
3. Custom Model
This model is the endgame of transitioning from “as is” to “as it should be” for businesses that wish to tap into the full potential of the generative AI and turn it from a merely useful service into leverage. However, there is a reason why we only see a few examples of such models in the business environment, aside from Google and Microsoft Copilot.
Custom models imply switching from foundation services to licensed large language models (LLM) explicitly tailored to the needs, goals and data of the company’s enterprise. Such a change requires considerable budget, time and human resources, which prompts the question: are LLM-based tools worth the investment? The short answer is yes.
The long answer is based on my professional experience: At least nine in 10 business process transformation plans are at risk of failure because of internal data management issues, such as legacy documentation and having to navigate between multiple information sources.
LLM-powered systems alleviate most of such issues by enabling an in-depth analysis of a wide range of documents (paper invoices, digital tax reports and images). They also facilitate interaction between multiple enterprise documents by storing data in vector format and categorizing data via embeddings (transformed data bits representing a text group through a semantically connected keyword).
As a result, enterprises can implement a more personalized approach to all the directions generative AI can be applied in: from marketing to sales to HR automation and data analytics.
Only As Efficient As It’s Taught To Be
So, is generative AI a universal cure? Certainly not. AI isn’t an artificial mind—it offers solutions based on the approach and data quality. However, it can be taught and trained to solve problems, delivering exclusive benefits to any enterprise interested in going beyond foundational models.
As digital adoption accelerates and the AI solutions market continues to grow, companies seek to gain leverage through empowering their processes with custom-built solutions. Therefore, foundational models will likely no longer suffice and be replaced with generative AI products made to fit the company’s niche, business vision and employee needs within a comprehensive enterprise AI strategy.
To advance to this new level, companies will require a narrow focus on long-term goals and pain points, and to ensure readiness for a considerable technological upgrade.
Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?
Read the full article here
