Join our daily and weekly newspapers for exclusive content on the latest updates and industry-composure AI coverage. learn more
A Brooklyn-based startup is the target of one of the most notorious pain points in the world of artificial intelligence and data analytics: the laborious process of data preparation.
Structure Today emerged from stealth mode, announced its public launch, under the leadership of seed funding with $ 4.1 million in seed funding Ban capital venturesWith participation from 8VC, Integral venture And strategic fairy investors.
The company’s platform uses an proprietary visual language model called called Dora According to industry surveys, a process that usually consumes 80% of data scientists’ time – to automate data assembly, cleaning and structure.
In a special interview with venturebeat, Ronak Gandhi, co-founder of Structurifying, said, “The amount of information available today has exploded completely.” “We have hit a major divine point in data availability, a blessing and a curse. While we have unprecedented access to information, it is largely inaccessible because it is so difficult to convert into the right format for making meaningful business decisions.”
The approach to structure calls data experts to “bottleneck of data preparation”, it reflects the growing industry-wide focus when solved. Gartner research indicates that Inadequate data preparation Successful AI is one of the primary obstacles for implementation, with four out of four businesses lack the data foundation required to fully capitalize on the general AI.
How AI-AI-AAP is unlocking the hidden business intelligence on the change scale
At its core, the structural users allow users to make a custom dataset by specifying the data schima, selecting sources and deploying AI agents to remove that data. The platform can handle everything from SEC filing and linkedIn profiles to news articles and special industry documents.
According to Gandhi, the one who sets separately is his in-house model Dora, who navigates the web like a human being.
“It is super high quality. It navigates and interacts with the same thing, just like a person,” Gandhi explained. “So we are talking about human quality – this is the first and most important center of the principles behind Dora. It reads the Internet in a human way.”
This approach allows the structure to support a free level, which Gandhi believes will help to make access to structured data democratic.
Gandhi said, “The way you now think about data, it is a really precious thing.” “It’s really precious that you spend so much time and pass around and wrestle, and when you have it, you like, ‘Oh, if someone was to remove it, I will cry.”
The vision of structuralFi is to “reduce data” – making it something that can be easily rebuilt when lost.
From finance to construction: how business industry-specific challenges are deployed to solve the specific challenges
The company has already seen adoption in many areas. Finance teams have used it to extract information from pitch decks, construction companies convert complex ground technical documents into readable tables, and sales teams collect real-time organizational charts for their accounts.
Slater stitchPartner at Bain Capital Ventures highlighted this versatility in the announcement of funding: “I ever have a handful of data sources in every working company that is very important to work with both and a huge pain, whether it is buried in PDF, scattered in hundreds of web pages, hidden behind an enterprise soap aPI, etc.”
The variety of the early customer base of the structure reflects the universal nature of the challenges of data preparation. As Tektargate ResearchData preparations usually include a series of labor-intensive stages: collection, discovery, profiling, cleaning, structured, change and verification-all before any actual analysis begins.
Why human expertise remains important for AI accuracy: ‘quadruple verification’ system of Inside Structure
A major discrimination for structuring is its “quadruple verification” process, which combines AI with human inspection. This approach is a significant concern in AI development: ensuring accuracy.
“Whenever a user sees something that is suspicious, or we can potentially recognize some data as suspects, we can send it to an expert in terms of that specific use,” Gandhi explained. “That expert can act in the same way as (Dora), navigate on the correct piece of information, remove it, save it, and then verify whether it is correct.”
This process not only corrects data, but also training examples that improve the performance of the model over time, especially in special domains such as construction or drug research.
“Those things are very dirty,” Gandhi said. “I never thought in my life that I would have a strong understanding of geology. But we are there, and I think, there is a great strength – able to learn from these experts and put it directly into the Dora.”
As data extraction tools become more powerful, the concern of privacy is essentially generated. Structurification has implemented security measures to address these issues.
Gandhi said, “We do not do any authentication, whatever a login is required, anything that you need to go behind some meanings of information – our agent does not do so because it is a privacy concern,” Gandhi said.
The company also gives priority to transparency by providing direct sourcing information. “If you are interested in learn more about any particular information, you directly go to that material and see it, such as inheritance providers where it is this black box.”
Structure enters a competitive landscape that includes both installed players and other startups addressing various aspects of the data preparation challenge. Like companies Change, Information, MicrosoftAnd Chitra All data preparation offers the capabilities, while many experts have been acquired in recent years.
According to CEO Alex Reichenback, its structure separates, it has a combination of speed and accuracy. Recently LinkedIn Post by Reichenbach claimed that he had “10x” to his agent, cutting the cost ~ 16X through the improvement in model adaptation and infrastructure.
The company’s launch comes amid the increasing interest in AI-operated data automation. according to a Techtarget reportAutomating data preparations “is often quoted as one of the major investment areas for data and analytics teams,” is becoming increasingly important with the promoted data preparation capabilities.
How disappointing data preparation experiences inspired two friends to bring revolution in the industry
For Gandhi, address the problems that he faced for the first time in the previous roles.
Gandhi said, “The big thing about the founding story of the structure is that it is both a personal and professional.” “I (Alex) were telling about the time when I was working as a data analyst and opies and consulting, these really preparing the niche, BSpoke data set for customers – the list of all fitness affected and their following metrics, the list of companies and the jobs that they are posting, they were completing the museums on the East Coast.”
The inability to quickly recur quickly from the idea to the dataset was particularly disappointing. Gandhi said, “What I got was that you could not recur and go to the idea to set data in a quick fashion,” Gandhi said.
His co-founder, Alex Richenback faced similar challenges while working at an investment bank, where data quality issues obstruct the attempts to make models at the top of the structured dataset.
How to use your $ 4.1 million seed funding to change enterprise data preparation
With new funding, plan to develop your technical team and install yourself as “Go-to data tools in industries”. The company currently offers both free and paid levels, which require enterprises options for those who require advanced facilities such as on-radius deployment or highly specific data extraction.
As more companies invest in AI initiative, the importance of high quality, structured data will only increase. recently MIT technology review insight report It was found that four out of five businesses are not ready to capitalize on generic AI due to poor data foundations.
For Gandhi and Structured Team, solving this fundamental challenge can unlock significant value in industries.
Gandhi said, “The fact is that you can also imagine a world that is making data sets, it is a lot for our users.” “At the end of the day, the pitch is about being capable of this control and adaptability.”
