The adage “garbage in, garbage out” is a crucial reminder that data management is foundational to artificial intelligence (AI).
“AI is only as good as the data you feed it with. So, data quality, data governance and data lifecycle management [which oversees data from its point of origin to its eventual deletion] are relevant and important to every company that wants to deploy and use AI,” says Tianyi Jiang (TJ), CEO and co-founder of AvePoint. Listed on Nasdaq, AvePoint is a data management and data governance solutions provider.
Last September, the company, having already listed on the Nasdaq, attracted additional investment from Temasek-backed 65 Equity Partners, which took up a stake of around 9%. With this investment, AvePoint is likely to have a listing in another exchange. "We believe that an eventual dual listing on the Singapore Exchange S68 will garner additional investor support in the region,” says TJ.
Meanwhile, the company is focusing on growing its business. TJ shares that there are three phases to any AI deployment. The first is ensuring data quality, which involves data acquisition and data preparation. “Data analysts and scientists spend 80% of their time cleaning, standardising and organising data for their AI or machine learning projects. This is because data comes from various systems that serve different business applications but you need to ensure data is consistent [for AI’s use],” he explains.
Secondly, organisations must continuously train their AI/machine learning models to address concept drift, wherein a predictive model’s performance declines over time due to changes in its environment.
Finally, there is a need to ensure the quality of AI’s output as there have been instances of AI hallucinating or generating false or misleading information but presenting it confidently as facts.
See also: Alibaba anoints new chief in revamp of stalling commerce arm
“Gartner expects 10% of all the data in 2025 to be produced by generative AI. If that data is the result of AI hallucination and is fed into further AI/machine learning models, it will pollute the quality of the end result. If this vicious cycle continues, we could end up with deviations from the reality of the world that could lead to real-world dangers like not having the real view of the market or polarisation of views,” warns TJ.
Picks and shovels for AI
AvePoint’s AI and Information Management 2024 Report reveals that organisations were concerned about data privacy and security (71%) as well as quality and categorisation of internal data (61%) before implementing AI. Internal data quality remains a challenge for the majority of businesses when they implement AI.
See also: Break up Google? What’s at stake in antitrust action
Recognising this, AvePoint offers solutions that help organisations build a resilient data infrastructure so that they can maximise the value of AI in their operations. “We think of your AI model as a high-performance engine that needs to be fed with high-quality data. [Our solutions] help cleanse data input and output for your AI models. We provide the picks and shovels for the AI revolution,” shares TJ.
For instance, AvePoint Opus ensures organisations can manage all stages of their data lifecycle. It features AvePoint Maestro, which uses AI models powered by Azure Machine Learning to analyse content and metadata and assign appropriate policies to documents.
With AvePoint Opus, organisations can better manage information and ensure compliance. By using AI to automatically classify data, businesses can reduce the risks associated with information over-retention or accidental data deletion to meet compliance standards.
AvePoint Opus can also help optimise cloud storage. Specific rules ensure organisations meet retention and disposal requirements to reduce excess cloud storage costs and maintain greater control over their IT budgets.
Besides that, the solution offers automatic records management. This allows businesses to focus on higher-value projects and speed time-to-value for data-driven insights.
“To truly unlock the power of AI and machine learning, organisations need a comprehensive data strategy that will accurately analyse, govern and classify their data. AvePoint Opus provides a solution that is automated and capable of learning over time, allowing organisations to manage the troves of data they produce today to build that data foundation, maintain compliance and reduce storage costs,” says TJ.
Cosmic: A real-world use case
To stay ahead of the latest tech trends, click here for DigitalEdge Section
AvePoint is no stranger to the Asia Pacific market. The region contributed 20% to the company’s overall revenue as at Dec 31, 2023. Regionally, it has presence in Australia, China, Japan, Singapore, the Philippines, South Korea and Vietnam. The Australian Government Department of Health and the Monetary Authority of Singapore (MAS) are among its key customers in Asia Pacific.
AvePoint worked with MAS and six major commercial banks in Singapore to develop the recently launched Collaborative Sharing of Money Laundering/Terrorism Financing Information and Cases (Cosmic) digital platform. The banks are DBS, OCBC, UOB, Citibank, HSBC and Standard Chartered Bank.
Those banks can use Cosmic to share information with one another if a customer’s profile or behaviour displays certain objectively-defined indicators of suspicion, or “red flags”. However, they must have in place policies and operational safeguards to protect the confidentiality of information shared. According to MAS, information sharing is currently voluntary and focused on three key financial crime risks in commercial banking: misuse of legal persons; misuse of trade finance for illicit purposes; and proliferation financing.
Cosmic uses AvePoint’s solutions which enable it to use machine learning and AI algorithms that can detect certain anonymous patterns, TJ shares in an interview with the Singapore Business Review. He adds that AvePoint plans to work with other government agencies and regulators in and out of Singapore on similar projects “once they see the efficacy of Cosmic”.
AI for internal innovation
To drive faster innovation internally, AvePoint is integrating AI across its entire operations. It is also opening an AI Industry Lab in Singapore, in partnership with the Economic Development Board of Singapore and other institutes of higher learning.
Focusing on talent development, TJ shares that AI Lab will “hire PhD holders in Singapore as well as bring in foreign researchers”. AvePoint plans to hire up to 50 people for its AI lab in Singapore over the next three years.
The lab will also send local/Singaporean researchers to AvePoint’s offices worldwide so that they can gain first-hand experience of the challenges faced by other markets and verticals.
“We want to expose our employees to different countries, regions and verticals to promote cross-pollination or sharing of knowledge and ideas to enrich one another. This can also help us collect signals [of opportunities and challenges] from all over the world to enable us to move and innovate faster,” says TJ.
Pursuing inorganic opportunities
In February, AvePoint announced that it will be anchoring a new growth equity fund, A3Ventures, as part of its effort to pursue inorganic opportunities.
“A3Ventures will invest in B2B software companies that are ready for the global stage, including those focused on accelerating generative AI for the digital workplace around the globe. In addition to creating a thriving ecosystem for AvePoint cloud customers and partners, our prominent role will provide line of sight to attractive assets, enhance our engagement with our channel network, expand our influence with strategic partners, and increase our total addressable market,” says TJ in the 4Q2023 earnings call on Feb 29.
On May 9, AvePoint reported that its total revenue for 1Q2024 was US$74.5 million ($100.55 million), up 25% from the same period in 2023. The company expects total revenues of US$73.8 million to $75.8 million in 2Q2024, or y-o-y growth of 15% at the midpoint.
TJ comments: “Our first quarter was a very strong start to the year. Our performance was again driven by the robust capabilities of our platform, as well as the growing recognition among customers and partners of the need — now more important than ever — for a strong data foundation, which in turn is critical to deploying a successful AI strategy. We are laser-focused on steady execution as we continue capitalising on the massive opportunity ahead of us.”
