AI not a ‘silver bullet’ for companies — experts
ARTIFICIAL INTELLIGENCE (AI) is no “silver bullet” for companies and integrating tools powered by the technology into their operations should be done strategically and responsibly, experts said.
AI-driven Software as a service (SaaS) companies and data technology firms said during SaaScon PH 2024 held by Sprout Solutions last week said companies must manage their expectations about AI tools.
“We need to scale back [our] expectations because a lot of people are coming from [thinking] ChatGPT is a silver bullet. But we have to be responsible in educating about AI,” Thinking Machines Data Science Enterprise Solutions Engineering and Gen AI Solutions Lead Kyle Patrick Bartido said.
Mr. Bartido said security, cost, privacy, and trust are their clients’ main concerns about using AI.
“There should be a way for a human to override an AI decision or AI’s reputation and also give it some feedback for training,” he said.
Symph Chief Executive Officer (CEO) and Founder Dave Overton likewise noted that incorporating AI models will not solve everything as these tools have limitations.
BlastAsia, Inc., Steer, & Xamun CEO and Co-Founder Arup Maity noted that AI costs can be high and come with no promise of return and even the risk of billions in losses.
“Will it even be worth it in terms of the Philippine cost setting?,” Mr. Maity said.
He said companies can build their own AI models at a fraction of the usual investment costs as start-ups mostly run on limited funding.
Mr. Maity added that firms need to identify four zones where they can take advantage of AI, namely, for cost reduction, revenue increase, a new business model, or simply experimenting and learning.
Senti AI CEO Ralph Vincent Regalado said companies should educate themselves about the kind of investment required to integrate AI, as these tools can drive revenues, help manage operating costs and affect their bottom line if leveraged properly.
“For some of our customers, we usually start with a very small proof of concept to prove a point that this would help them before scaling the actual implementation,” Mr. Regalado said. — A.R.A. Inosante
