Whether it’s a chatbot, a content generator, or a predictive model, startups and scaleups are hustling to find marketable products to capitalize on the generative AI wave.
But some are charting a different course. Instead of just rushing to productize AI, they’re exploring how this technology can be leveraged to enhance and optimize their own existing operations and performance.
Zoho Canada recently hosted a webinar on the topic of generative AI in business intelligence and how the tool can be deployed to make organizations more data-driven.
The discussion offered four key insights for startups thinking about applying generative AI within their operations, and featured Matt Aslett, director of research, data, and analytics at Information Services Group’s Ventana Research, and Zoho Canada’s Chandrashekar Lalapet Srinivas Prasanna (LSP).
The clock is ticking
Businesses are beginning to understand the power of generative AI in enhancing their data-driven decision-making.
But the rate of AI adoption remains slow. A recent report from Deloitte Canada revealed that 56 percent of Canadian businesses don’t yet use AI, though nearly 20 percent said they plan to adopt it in the next few years.
This, according to LSP, is not okay.
“The goal is to make organizations data-driven or insights-driven, and to do that you have to democratize access to data.”
“The goal is to make organizations data-driven or insights-driven, and to do that you have to democratize access to data,” he said.
Aslett agrees that AI for business intelligence will soon become table stakes. His firm predicts that by 2026, nine in 10 analytics processes will be enhanced by AI and machine learning to both streamline operations and increase the value that can be derived from data.
“We’re at the brink of a significant period of change in the [business intelligence] space,” Aslett said. “We think that by 2026, three quarters of enterprises will realize that their existing analytics capabilities are not as effective as they could be without those generative AI capabilities.”
Get the jump on the competition
In a recent ISG survey, Aslett noted that only seven percent of participating businesses said they were neither using or exploring generative AI.
“This is true across all industries,” Aslett added. “If you’re not at least exploring the potential benefits of AI in [business intelligence], you are at risk of getting left behind.”
While being data-driven is more often talked about at the enterprise level, he believes it should be prioritized by emerging startups too, where resources and funding might need to be redeployed in order to derive insights from existing data.
According to ISG data, generative AI is expected to reach one-half of AI spending this year, and the average proportion of IT spend on AI is expected to rise from two percent to almost six percent in 2025.
“The ramp up of budget allocation to generative AI has really been significant,” Aslett added, meaning startups will have to start investing in order not to get left behind.
You don’t know what you don’t know
Aslett explained that efficiency is the primary motivation for businesses adopting AI. By automating data preparation and cleansing, insights, predictions, and forecasting through intuitive natural language interfaces, AI is accelerating what are often manual and time-consuming tasks.
This could be particularly useful in remedying Canada’s long-standing productivity gap—a challenge even the country’s finance minister has called the “Achilles’ heel” of the economy.
By speeding up these processes, AI not only has the power to make organizations more productive, but it can also make them more efficient by improving the quality and accessibility of data.
“AI can really facilitate decision making by lowering barriers to accessing and understanding data for potentially all people across the organization,” Aslett added.
While natural language processing as a business intelligence tool is not new, Aslett explained that the tech previously required a good deal of upfront work in order to get internal data ready for analysis.
“Large language models solve so many of those problems,” Aslett said, noting they can automatically generate synonyms to aid with data collection, accept prompts, and allow users to set boundaries around certain queries.
During data analysis, generative AI models can be trained to identify anomalies and outliers, as well as identify hidden patterns, trends, and causations. Models can also generate charts, graphs, or even entire reports summarizing data and making it more accessible to decision-makers.
Finally, large language models can provide recommendations based on data-driven insights and simulate various scenarios based on the data to predict outcomes under different conditions.
For LSP, the last component speaks to Zoho’s core mission to help businesses obtain key insights, “…and based on those insights, get to specific actions that can drive a change in their business,” he said.
“Proceed with caution”
Of course, AI is not without risks to organizations. Implementing AI can introduce data privacy concerns, ethical considerations, and the potential for bias in decision-making processes.
There’s also the risk of over-reliance on automated systems. Relying too heavily on AI can lead organizations to overlook the importance of human judgment and intuition in critical decision-making scenarios. It’s why Aslett advises companies to “proceed, but proceed with caution.”
“Organizations really need to be aware of accepting the output of those models and services as the gospel truth,” he said. “You need to be aware that those large language models generate content that is grammatically valid and appears absolutely valid and in some cases is, but it’s not necessarily guaranteed to be factually accurate.”
To learn more about Zoho’s AI solutions, please visit: www.zoho.com/zia/
Feature image courtesy Unsplash.