AI may be the “IT” of the moment. But that doesn't mean it's easier to implement.
According to the 2023 S&P Global Survey, approximately half of companies with at least one AI project in operation are still in the pilot or proof-of-concept stage. There are many reasons for slow startups, but commonly cited challenges include data management, security, and computing resources.
Roughly half of companies responding to an S&P poll said they are not ready to implement AI and do not intend to do so within the next five years or more.
Fortunately, there is a growing number of products from startups and big tech vendors alike that aim to address these obstacles to AI adoption. (See Google's ML Hub, Kore.ai, Viso, to name a few.) One new entrant is a platform that allows users to build and deploy models without writing code. This is his Pienso.
Virago Jones and Kartik Dinakar founded Pienso in 2016 based on their research at MIT (where they are alumni). The two met several years ago as graduate students at MIT's Media Lab.
“We collaborated on a class project to build tools to help social media platforms moderate and flag bullying content,” Jones, CEO of Pienso, said in an interview with TechCrunch. . “He only had one problem. The model itself worked as expected, but it wasn't trained on the right data, so it couldn't identify harmful content that used teen slang. ”
Jones and Dinakar eventually realized that the solution was to enlist an expert in the problem, in this case a teenager, to help train the model. They built tools for this purpose, and a few years later, Jones and Dinakar worked together to commercialize these tools.
The result was Pienso. Jones refers to this as “non-technical talent” – researchers, marketers, and customers who have access to large amounts of data for AI training but lack the resources needed to structure and analyze it. Support describes it as an AI suite built for his team. that.
“The majority of AI conversations have been dominated by large language models,” Jones says. “But the reality is that no one model can do it all. To extend the full potential of AI to manage business processes and interact with customers, you need to train and train models. You need to be able to fine-tune it, and Pienso believes that experts in any field should be able to do that, not just AI engineers.”
Pienso guides users through the process of annotating and labeling training data for preconditioned open source or custom AI models. (Depending on the model, AI typically requires a label (such as an image of a bird combined with the label “finch”) to learn how to perform a task.) The platform can be deployed in the cloud or on-premises. can. Integrate with enterprise systems via APIs. However, it can also work without APIs or third-party services to keep your data safe.
British broadcaster Sky is using Pienso to analyze customer service calls, Jones said, while an unnamed US government agency is testing Pienso to monitor tracking illegal weapons. It is said that there is.
“Pienso's flexible no-code interface allows teams to train models directly using their own data,” said Jones. “This reduces privacy concerns when using the … model and also makes it more accurate and able to capture the nuances of individual companies.”
Companies pay Pienso an annual license based on the number of AI models they deploy. As the number of models increases, licensing costs also increase.
“We intentionally designed our pricing to allow customers to test our models upfront and understand how AI can help them without making a large investment upfront,” Jones said. added. “We wanted to give our customers the freedom to experiment building new models before deploying them.”
This appears to be an attractive business model for investors. Pienso recently raised $10 million in a Series A funding round led by Latimer Ventures with participation from Gideon Capital, SRI, Uncork, and Good Growth Capital.
Bringing Pienso's total funding to $17 million, Jones said the new funding will be used to scale Pienso's sales, marketing and customer success teams, hire engineering talent and build new features for the platform. .
Luke Cooper of Latimer Ventures said in a statement: “We hear all the time about the need to democratize AI, but what makes Pienso stand out is how they think about the role of domain experts in this equation. A future where smarter AI models for specific applications are built by the people who know best the problems they're trying to solve. will be encouraged.”