Commentary: Artificial intelligence is hard. Low-code options like Akkio aim to make AI much easier.
Every company may want to put artificial intelligence to work, but most companies aren’t blessed with the ability to hire battalions of data scientists–nor is that necessarily the right approach. As Gartner analyst Svetlana Sicular once argued, often the best possible data scientist is the person you already employ who knows your data and simply needs help figuring out how to unlock it. For many business line owners, it’s this kind of approach that may make the most sense, as they seek to be smarter with the data they already have.
One company working to enable this vision is Cambridge, Massachusetts-based machine learning startup Akkio, which pairs AI with low code in an attempt to democratize AI. I caught up with company co-founder and COO Jon Reilly to learn more.
SEE: Research: Increased use of low-code/no-code platforms poses no threat to developers (TechRepublic Premium)
“There’s a revolution underway in how basic software gets built,” Reilly said. “Just as the clickable icons of Windows and then macOS replaced typing obscure DOS commands into a black computer-terminal screen, new ‘no-code’ platforms replace programming languages with simple drag and drop features.”
Though there are definite successes where AI has dramatically improved how technology gets built, in general the platforms are not yet powerful enough to create complicated software (Copilot can help on suggestions during coding, while Diffblue can automate Java unit test writing), but for simple needs like managing schedules or tracking customers they may be enough. The inevitable abstraction from code to graphical interfaces opens the door for any tech-minded person to build a software solution for basic tasks, rather than entering a ticket in the IT department’s crowded queue.
Even software developers are using these tools to solve simple problems fast.
“Every company can see productivity gains as they digitize their business,” Reilly said. “The problem is that not every company has a tech team capable of building software. Off-the-shelf products are rarely a perfect fit.”
Google, Microsoft and Amazon have all introduced products that allow users to create applications without writing a line of computer code. They target people who know the business far better than the IT department, yet lack deep technical expertise to write software. By dragging and dropping templates and combining them into logic trees, users can build software to do things like process invoices or present live data in a user-friendly format.
SEE: Business leader as developer: The rise of no-code and low-code software (ZDNet/TechRepublic special feature) | Download the free PDF version (TechRepublic)
The no-code movement is even putting the power of AI within the reach of marketing managers, sales staff and financial analysts. Feed these no-code platforms data, and they will spit out predictions in seconds, giving anyone the power normally reserved for companies or research institutions manned by data scientists.
This is particularly important now given the shortage of data scientists and software developers. While demand for developers is surging, schools have fallen behind. According to Computer Science Education Week, computer science does not even count toward high school graduation in 35 out of 50 states. Management consulting firm Korn Ferry predicts a global shortage of 4.3 million tech, media and telecom workers by 2030.
“But even better than hiring scarce and expensive developers, no-code platforms are incredibly cheap for what they do,” Reilly said. “They are volume plays, betting that once businesses understand how easy they are to use, everyone will use them to address the dozens of processes or decisions for which software can help. In time, no-code platforms will be as ubiquitous as word-processing or spreadsheet software is today.”
Removing friction from adoption has the potential to unleash the power of AI across all new industries and allow non-specialists to produce work much faster with higher precision while literally predicting the future. How so? These AI platforms promise to take much of the guesswork out of forecasting and increase the clarity of the near term future for many businesses. They typically work on any kind of tabular data, predicting everything from winners in a horse race to how much steel a factory is going to need next year.
Such AI platforms are already allowing sales teams to prioritize leads. Instead of relying on intuition, a machine-learning algorithm can predict which leads are worth spending time chasing and which can wait. It’s a matter of increasing business velocity–something that all organizations need.
Reilly shared an example about a manufacturer. One of its vendors spent months with application developer teams and AI specialists to build clever software to sort through the manufacturer’s online advertising and decide which targets were worth spending more money on. That took months and never worked very well. Then the vendor tried a no-code platform and came up with a solution within hours. The AI platform lets them simply upload a spreadsheet of advertising targets and other metrics and, presto, the targets worth reaching rise to the top.
“Under the hood what is happening is called predictive analytics using state-of the-art machine-learning algorithms,” Reilly said. “But here’s the good news: You don’t have to care or even know that. And you don’t need to go to your IT department begging for hardware, software or specialists. You just do it yourself, no computer coding required.”
Disclosure: I work for AWS, but the views expressed herein are mine.
Learn the latest news and best practices about data science, big data analytics, and artificial intelligence. Delivered Mondays
Codecademy partners with Black and Brilliant for 10-week AI course (TechRepublic)
5 programming languages application solutions developers should learn (free PDF) (TechRepublic)
A guide to The Open Source Index and GitHub projects checklist (TechRepublic Premium)
Programming languages: Developers reveal most loved, most loathed, what pays best (ZDNet)
Programming languages and developer career resources (TechRepublic on Flipboard)
Matt Asay is a veteran technology columnist who has written for CNET, ReadWrite, and other tech media. Asay has also held a variety of executive roles with leading mobile and big data software companies.
8 must-have tools for developers on Linux
Hiring Kit: IT finance manager/budget director
Feature comparison: Time tracking software and systems
7 steps: Learn how to migrate to Linux desktop