The hype cycle for GenAI is at its peak, with predictions from organizations like McKinsey suggesting it could inject trillions into the global economy. While the excitement is palpable, it’s crucial to temper this enthusiasm with lessons from recent history to assess its true potential in heralding the next industrial revolution.
THE INTERNET BUSINESS MODEL
Reflecting on the early 2000s internet boom, investment money flowed freely until the market recognized the many “golden” companies lacked profitable models, leading to the dot-com bubble burst. But by 2004, the rise of smartphones forever changed the internet business model and gave rise to numerous profitable, sustainable companies that are global leaders today.
The catalyst was twofold—the hardware device (smartphone) allowed internet and cloud companies to gain direct access to global users, and innovations from the application industry soared.
AI SHOULD ALSO STAND FOR ALL-INCLUSIVE
The cornerstone of a sustainable business model is long-term profitability, which is currently unattainable for GenAI due to its restricted access. Only a select few companies can bear the hefty initial costs often exceeding $1 million for equipment and the ongoing expenses of training and managing data on LLMs.
While ChatGPT is fine for consumers to experiment with, it currently has little business value because it is trained on data that is not particularly relevant to individual enterprises. For true business intelligence, companies must leverage and own their unique data.
This vital information is highly sensitive, and if companies go the cloud route, they may lose control of their own data, expose it to security risks, and become locked into very expensive and variable cloud services. It is risky and cost-prohibitive for most companies. At this point in the AI revolution, inclusivity must be front and center to truly unlock the promise of AI and allow it to fully evolve. Otherwise, AI will remain an “artificial innovation” that will not be sustainable and the AI bubble will burst.
Learning from the internet business model, AI today should stand for all-inclusive as well as artificial intelligence. For the technology to really work, AI must come out of the clouds and walled gardens of the hyperscalers and directly engage with businesses right at the edge of where business happens—workstations, servers, etc.
GAME-CHANGING ACCESSIBILITY
In my role as the general manager and president of Phison North America, my job involves making mission-critical decisions, understanding customer needs, ensuring timely product launches, and fostering an innovative culture. These experiences highlight the practical challenges and opportunities in democratizing GenAI. As a design-centric organization, 75% of our global team is composed of engineers.
We recognized the need to unlock the power of GenAI for our internal use to stay competitive. However, our journey has taken us much further than we could have imagined even last year.
We faced the classic “make or buy” decision. Finding the existing solutions to be prohibitively expensive, we chose to invest in building our own. To our great delight, we soon realized that we invented a shortcut to democratize AI and brought it to market. Now we’re helping organizations of all sizes unlock the power of GenAI at a low cost while owning and training their data, keeping it on-premises, and building their own ecosystem of partners for their unique needs.
This accessibility is a game changer because most enterprises can afford it, rely upon it, and operate it without requiring additional engineers and IT staff to run it. In less than a year, over 100 enterprises have started using it and we have found 12 distinct use cases.
WAYS GEN AI CAN STREAMLINE UNIVERSAL PAIN POINTS
There are three distinct areas that are universal pain points for organizations of all sizes: onboarding new employees and ongoing professional development, keeping up with coding demands, and automating tasks generated from huge volumes of data. Let’s take a look at how GenAI can streamline these.
1. Chat knowledge management system
Many organizations struggle with employee onboarding and professional development. A chat knowledge management system feature can enhance productivity by streamlining product specification searches and providing document analysis for higher efficiency.
With robot training, the learning process is accelerated and redundant tasks are eliminated. Personal development can be enhanced through improved knowledge-search efficiency with optimal company responses. Knowledge management is continuously refined and reassessed with information ingested from documents, emails, and Jira. With this, employee learning curves can be flattened and productivity can continuously increase.
2. Helpful tools
The creation of software code is time-consuming and expensive. Helpful coding tools can automate coding, prevent spelling errors, and write test cases to clarify processes and avoid issues (full disclosure: our software tool, CodePilot, offers these services).
Coding software tools can also help enhance efficiency and readiness through refactoring and perform documentation audits with annotated files and review assistance. The result is faster, more efficient code writing and reduced labor costs.
3. Code documentation
The documentation of coding decisions is tedious and takes time away from more creative tasks like writing new code. A code documentation feature automates it so staffers can focus on innovation rather than reporting.
The metrics are startling—we have seen a reduction in manpower from 17 weeks and 20 people to two weeks and four people.
FINAL THOUGHTS
We must learn from history and demand the democratization of GenAI to ensure a sustainable and profitable future industrial revolution, making the world a better place for everyone.