AI • IOT • NETWORK EDGE

Whether the goal is to speed office tasks or impress customers with chatbots, today’s businesses are increasingly eager to deploy AI applications.

Once launched, AI applications can be a boon to productivity. But creating them can be a time sink, especially for generative AI solutions, which are powered by large language models and image recognition systems that require extensive fine-tuning and testing.

Now there’s a better way to bring AI solutions to fruition. Using a low-code platform, businesses can develop custom AI applications much faster. Low-code applications are more straightforward to maintain and customize to accommodate future use cases.

Simplifying AI Solutions Development

In the competitive world of AI applications, timing is a critical factor, says Brian Sathianathan, Co-Founder, Chief Digital Officer, and Chief Technology Officer at low-code AI platform developer Iterate.ai. “A lot of companies want to be the first to market with innovative solutions. But it’s hard to do because their IT and technology teams already have their hands full,” he says.

Sathianathan and his colleagues founded Iterate to simplify the AI application-building process, shortening development time from months to weeks. “On average, it’s eight or nine times faster to take an AI idea from concept to reality,” Sathianathan says. “Less complex AI solutions can be created up to 17 times faster.”

Iterate simplifies the #AI application-building process, shortening #development time from months to weeks. @IterateAI via @insightdottech

Iterate saves time by creating pre-written blocks of code for various AI capabilities, such as chatbots, payment systems, or image recognition. Using the company’s Interplay platform, developers can drag and drop the code blocks into their solutions.

“It’s like building a luxury home from parts delivered on a truck,” Anton says. “We send you entire kitchens, bedrooms, and bathrooms, and you can put them together very quickly.” The code blocks are grouped into customized solutions for industries such as finance and insurance, retail, and automotive.

Saving Time with a Low-Code Platform

Interplay’s enterprise office solution, GenPilot, allows organizations to build their own generative AI large language models (LLMs) from internal data and documents. Many LLMs specialize in tasks, such as financial planning or logistics management, and GenPilot allows companies to select the models they prefer. Though public LLM solutions such as Chat GPT and Microsoft Copilot can also be used for generative AI solutions, some companies hesitate to upload their information to them.

“Public models are shared in a multi-tenant cloud environment. We provide a secure private environment, and companies can run their models on-premises,” Sathianathan says. Banks, insurance companies, and other organizations can also build in compliance rules governing data in various regions.

For employees, GenPilot saves hours of time by gathering and interpreting documents across databases. For example, if an insurance customer emails a company representative with a question, but neglects to supply their policy number, GenPilot will not only find it but determine how the policy applies to the question, how much the customer pays for services, and whether a change would alter the fees. It then composes a reply to the customer’s email.

“It responds intelligently in plain English,” Sathianathan says. Companies can set rules governing tone of voice and level of technicality.

For unstructured documents, such as PDFs, employees can use a different solution, the Interplay OCR Reader. This application translates images into machine-readable text and initiates workflows. For example, when bank employees upload customers’ scanned documents to the OCR Reader, it extracts relevant information and enters it onto a loan application form.

Streamlining Retail AI Management

One of Iterate’s latest solutions is Interplay Drive-Thru, which builds voice-enabled chatbots to take customer orders and make upselling recommendations at busy quick-serve restaurants (QSRs).

Chronic labor shortages often require QSR workers to perform multiple duties, packaging food, taking payments, and serving in-store customers as well as those at the drive-thru. “Chatbots give them a little more breathing room,” Sathianathan says. Orders are processed faster, shortening lines for customers and increasing throughput for restaurants.

Drive-thrus and other retailers can speed payments with Interplay’s LPR (license plate recognition) solution. Customers who opt in supply a photo of their license plate and credit card, and are recognized by computer vision cameras as soon as they arrive at a participating business. Interplay LPR, which complies with GDPR and other privacy regulations, is currently deployed at more than 1,000 gas stations and convenience stores in Europe.

“It will automatically open the pump for customers and charge them for gas. These actions happen within 30 milliseconds,” Sathianathan says.

Interplay’s LLM solutions are deployed on Intel^® processors. Applications that run on high-performance CPUs are more cost-effective for businesses than those that also require GPUs, as many LLM solutions do.

“A system using only CPUs cost $2,500 to $4,000 per machine. An equivalent GPU/CPU combination would be $8,000 to $12,000,” Sathianathan says. Retail IT teams are also more familiar with standard operating systems, reducing training time.

Once a low-code solution is deployed, developers can easily move the same Interplay code blocks into new solutions, instead of having to sort through millions of lines of code to make changes. In addition, Interplay’s code blocks use the Intel^® OpenVINO^™ toolkit, enabling developers to optimize their AI applications more efficiently. “You can use up to 350% less compute power with OpenVINO. That’s a huge benefit,” Sathianathan says.

Bright Future for Low-Code AI Solutions

Today’s AI applications enable companies to automate processes in ways that would have been unthinkable just a few years ago, Sathianathan says. “AI solutions can do sales calls. They can generate legal documents, which are traditionally expensive to produce.”

Using low-code building blocks, small businesses as well as large enterprises can develop solutions like these quickly and affordably. That will help to expand the reach of AI applications and level the playing field, Sathianathan says: “Very soon you will see many new automation capabilities being developed. Startups will be able to punch above their weight, and costs will continue to come down for everyone.”

This article was edited by Georganne Benesch, Editorial Director for insight.tech.