“No-code” is more than just hype. It’s a revolution. Prior to no-code, if you wanted to make a website, you’d need a technical web developer. Now, you can use Bubble, Webflow, Carrd, or a myriad of other visual tools. The same has happened in AI. Prior to no-code, you’d need data scientists, data engineers, and machine learning engineers to build and deploy AI models. Now, you can use no-code AI.
While no-code web development has been around for decades (think blogger.com of the late 90s), no-code AI is a lot more recent. As the number of low-to-no code AI tools grows, it's worth understanding the landscape as not all solutions are made equal, and it can still be very difficult for non-technical folk to build and deploy AI.
We’ve explored dozens of no-code AI options. We found that many no-code tools come with a sacrifice. If they delivered a lot of valuable features, they often sacrificed ease-of-use, while if they delivered usability, they sacrificed performance and a real value proposition.
We wanted something that both delivered on performance while still being as broadly usable as Excel. It would need to be end-to-end, with fast, free training to get results quickly, and a deep and effortless integration ecosystem. That product did not exist - so we made it ourselves. To keep ourselves honest to that vision we have made Akkio open for anyone to try without cost.
This article is just our view of the no-code AI market, having experimented with dozens of tools and finally deciding to build our own. We dive deep into a wide range of no-code AI products, giving you our thoughts as a no-code AI company. We’d love to hear your thoughts and feedback as well!
Steve Jobs once famously said that “the line of code that’s the fastest to write, that never breaks, that doesn’t need maintenance, is the line you never had to write.”
Writing code is simply the means to an end: Code is for creating applications. No-code platforms just get you to working applications faster. Even software developers go by the principle of “D.R.Y.,” or “Don’t Repeat Yourself.” No-code takes this to the extreme and uses visual interfaces on-top of codebases, so you reuse elements that have been made before and combine them in novel ways.
It’s now possible to build entire businesses with no-code platforms. Here are some examples of some popular no-code tools:
There’s no doubt that no-code is the future. As little as one-quarter of one percent of the world knows how to code. Anyone can be taught to quickly and easily use no-code.
We’ve all heard about AI, whether in the news, science fiction movies, or your friend raving about their self-driving car.
In reality, AI is just a statistical technique used to answer complex questions about data. Simple questions like these can be answered with simple statistics:
However, once we’re dealing with complex, non-linear interactions among variables, we need to use machine learning. For example, say you operated a retail store and wanted to predict sales. Sales could be impacted by variables like the day of the week, store hours, the level of advertising, and COVID-19 restrictions.
Further, these variables won’t have a linear relationship. A holiday typically boosts sales, especially on a Friday, while perhaps having no effect on a Sunday. Machine learning discovers these complex relationships automatically.
Predicting sales is just one simple example. You could use machine learning to predict customer churn, or employee attrition, or lead conversion, or any other KPI in your data. If you have historical tabular data (such as a CSV with your records or even just a Salesforce account), you can make predictions.
Given the power of AI to optimize any organizational KPI, companies have been implementing it for decades. Since the early 90s, for instance, Sprint has been using AI for “real-time telephone fraud detection.”
Today, virtually everything you do is impacted by AI. When you walk into a Walmart, you’ll see an AI-optimized number of associates to meet demand. When you enter a McDonald’s, you’ll be one small data point in their enormous customer demand models.
When you browse Netflix, Amazon, Tinder, Spotify, Medium, or even Google Search or YouTube, you’re given AI-powered recommendations. Incredibly, Amazon’s recommendation features are responsible for up to 35% of their revenue.
These multi-billion-dollar companies can afford massive teams of data scientists, data engineers, AI engineers, and the like, who spend years on complex machine learning projects.
Those efforts just aren’t feasible for SMEs—especially considering that data scientists command six-figure salaries in the US. For a company that just wants to boost sales, convert more leads, reduce churn, or optimize any other KPI, it shouldn’t take months and cost hundreds of thousands of dollars.
It should be effortless and affordable. That’s where no-code AI comes in.
There are so many analytics and AI tools out there, there’s no way we could cover them all.
Instead, we focused on a number of market segments, including status quo players (FAANG-size companies), data science tools (which sometimes have some no-code functionality), simpler no-code AutoML (which usually focus on one part of the end-to-end process), verticalized offerings, and end-to-end no-code AI.
These groupings allowed us to explore companies in every field of AI, from sentiment analysis to computer vision to Robotics Process Automation.
Akkio is an end-to-end no-code AI platform. This means you can build, deploy, and integrate AI models, all in one place, without any technical expertise. Often, AutoML tools and even no-code AI tools require software engineers and other technical professionals to integrate the models that are created. With Akkio, it’s easy to integrate AI into any workflow by building an “AI flow,” powered by a fully visual interface.
Prevision is a no-code AI solution that aims to “increase the productivity of your data science projects.” In other words, you’re expected to already be on the AI journey, and have some technical capability. Prevision is also focused on AI modeling, and not the end-to-end process that would include AI integration into your business workflow.
After signing up for a free trial, you can see that there are four steps: Uploading data, training a model, analyzing performance, and creating predictions.
We uploaded the Telco customer churn dataset from Kaggle. The 1 megabyte dataset uploaded immediately, but we had to wait several minutes for it to process in the background. When that was done, it was easy enough to build a churn model and make predictions, but there was no easy export or integration functionality.
That said, it’s clearly a powerful tool for highly-technical data scientists, speeding up the process of building a variety of machine learning models.
Gyana is similar to Prevision, in that there’s a straightforward, visual process to analyze data, but there’s no end-to-end no-code system to integrate those models into your workflow. Gyana is a good fit for basic modeling needs.
Gyana’s process looks like the following:
We uploaded the same Telco customer churn dataset to find insights. Gyana’s current AI functionality is limited to linear regression, so we couldn’t do the binomial classification required to predict churn.
However, if your needs are straightforward, Gyana is a solid no-code tool to create linear regression models from your data.
Peltarion is a highly complex, technical tool, even though code isn’t needed. That may sound counterintuitive, but here’s what we mean.
When you go to build an AI model in Peltarion, there are a huge number of metrics and settings to tweak. Things like “B2 rate” and “batch size” and “convolutions.” It feels like a thinly-veiled wrapper around code, so while it’s technically a no-code solution, it’s not necessarily much easier.
As with Prevision, however, the presence of these complex features could make it a powerful workhorse for more technically inclined users.
Levity provides a no-code AI solution for documents, images, and text, with use-cases like content moderation, insurance claims processing, and analyzing text messages.
Levity is still in a private beta, so we didn’t get a chance to test it out.
Apteo is a vertical focused no-code AI solution, helping “ecommerce companies segment their customer base and predict buyer behavior to increase customer lifetime value and improve retention.”
It integrates with ecommerce tools like Shopify, Stripe, and Square. In short, if you have specific ecommerce needs, Apteo is worth a look.
Syte.ai is another niche no-code AI tool for ecommerce, which uses visual AI, NLP, and hyper-personalization to drive better search and discovery.
If you’re looking to build a smart commerce recommender system, be sure to check out Syte.ai.
Most machine learning platforms work by finding statistical patterns in historical data, but not by finding cause-and-effect relationships. causaLens is taking an innovative approach to AutoML with its causal AI platform. It remains to be seen if causal AI is the way to go, as it seems to be more of a research interest at this stage.
Causaly is another player in the causal AI field, but is more verticalized than causaLens, and is focused on finding causal relationships in biomedical science research. Causaly uses machine learning to analyze over 30 million papers, clinical trials, and side effect databases.
This is the most hyper-specific AI platform we’ve looked at, but if you’re in the biomedical sciences, then it’s definitely worth a look.
PredictNow.ai is a vertical specific no-code AI solution, focused on financial machine learning, allowing you to compute the probability of profit for your next investment.
Don’t expect to get rich with financial machine learning, but if you’re already a technically-inclined investor, PredictNow.ai can help you speed up your game.
Accern is another no-code AI for finance solution, and it’s a much bigger player. This means they’re also a lot more expensive, and their Premium package on AWS will set you back $60,000 a year.
That said, Accern is a lot more feature-rich than PredictNow.ai, with functions for credit risk, ESG investing, quantitative research, financial crimes analysis, and more. If you have heavy-duty finance needs, Accern is worth trying out.
RunwayML is a no-code AI tool specifically made for creators and creatives, with features for image, video, and text data. For example, you can create synthetic images and videos, cut objects out of videos, and more.
Lobe is a Microsoft AI product that lets you make image classification models in clicks. It’s Microsoft’s answer to Google’s Teachable Machine, and is a great way to get started with image classification, but isn’t very feature-rich.
Google’s AI platform includes AutoML Vision, tabular AutoML, and other AI solutions. Since you don’t need to actually dig into Python to build models in the Google AI Platform, we’d say it counts as a no-code solution, but it’s far from easy-to-use.
Google AutoML is a hassle to set up and maintain, which is why it’s such a lucrative skill to have. Employers will pay significant figures just to be able to figure out the product.
Google’s own quickstart guide explains that you need to create a data bucket, enable Cloud AutoML and Storage APIs, set a training budget, and manipulate a number of other settings just to get started. After building a model, you also need to carefully undeploy your model, and delete both the model and dataset, or you might accidentally incur unnecessary Cloud AutoML charges.
In short, the Google AI Platform is about as complex as you can get while still being “no-code.” We also found Google to be on the expensive side for model training. For a deep dive on performance check out our benchmark of Google AutoML vs Microsoft Azure and Amazon Sagemaker.
A little known fact is that IBM Watson Studio Desktop allows you to build machine learning models without code, as Watson is more famous for its question-answering functionality.
That said, Watson’s no-code AI features clearly aren’t a focal point for IBM, as the functionality is limited to model creation, training, and deployment, without any serious integration features.
Amazon’s SageMaker Autopilot is similar to Google’s AI Platform, in that it’s similarly complex for a “no-code” AI tool.
To get started with SageMaker, you’ll need to set up an AWS account, set up Amazon SageMaker Studio with an IAM role that has the required permissions, run code to extract data from Amazon S3, create an experiment specifying details like “S3 location of input data” and “S3 location for output data,” and more. Be sure to check out the Sagemaker benchmark.
Azure AI is the last of the “status quo” AutoML platforms we’ll cover. It’s similarly complex, and even signing up requires phone verification, adding a credit card and personal address, and signing a lengthy agreement.
Once signed up, you’ll need to make a machine learning workspace, with details like a workspace name, subscription, and resource group. Then, you need to create a new Automated Machine Learning model, create an experiment, create a new compute (specifying compute name, virtual machine size, minimum and maximum number of nodes), select the compute, upload a dataset, select the prediction task, run the model, and more.
That’s only covering the gist of getting started for a test drive, so using Azure for your AI needs will be a serious endeavor. We found Azure to have similar performance to Google and Amazon Sagemaker in our benchmark test.
H2O.ai is a top contender in the traditional AutoML space, and a leading rival to the likes of Google AutoML and Azure AI. To demo the platform, you can sign up for a “Driverless AI Test Drive,” which gives you 2 hours of access, though the AWS marketplace estimates your infrastructure costs using H2O will be over $500 a month.
C3.AI is another leading AutoML contender, which recently conducted an IPO. They bill themselves as much more than just AutoML, claiming that the offerings of companies like H2O.ai are just a feature within C3.AI. C3.AI doesn’t offer a hands-on trial, so there’s no way to analyze the features first-hand, but C3.AI boasts a number of large enterprise customers.
That said, C3.AI doesn’t exactly claim that it’s effortless, and their own timeline suggests it could take over half a year before you deploy C3.AI models into production.
Splunk is another massive AutoML player, and is publicly traded, with a market cap of almost $30 billion, as of writing. Splunk refers to its AutoML features as the “Machine Learning Toolkit,” or MLTK.
Note that Splunk just makes it onto this list of no-code AI tools, as MLTK involves using SPL (Search Processing Language) commands to build machine learning models.
Coding isn’t strictly needed, but Splunk is still a relatively technical solution.
DotData calls itself the “AutoML 2.0” solution, referencing its “feature engineering automation” as the ‘2.0’ part. That said, most other solutions we’ve looked at offer some degree of feature engineering automation as well.
DotData is designed to “empower your BI & Analytics teams,” so while there are no-code features, it’s among the more technical solutions in this guide.
DataRobot is another popular enterprise AutoML platform, but it doesn’t come cheap. On the AWS marketplace, we can see that the DataRobot Managed Cloud AutoML will set you back almost $100,000 a year.
While DataRobot offers no-code features, it’s intended for a technical audience, including analytics leaders, data scientists, business analysts, software engineers, and IT operations teams. After all, if you’ll be spending six-figures on a single software subscription, you’ll probably want a technical user to make sure you’re taking advantage of it.
Dataiku is a leading data science platform, which comes with visual AutoML functionality as well. That said, Dataiku is more focused on general data science tasks like data integration, building data pipelines, data visualizations, and statistical analysis.
Auger AI is hyper-focused on creating accurate predictive models, and their main selling point is outperforming the accuracy of many other AutoML platforms. However, you won’t find a full set of end-to-end features, and Auger AI leaves something to be desired in terms of integrating AI in your workflow.
BigML, like Auger AI, is laser-focused on AI modeling, but less on integrations. If you’re more technical, and already in a role that involves building AI models, BigML can be a great way to speed up your process.
If you’re new to building, deploying, and especially integrating AI, then BigML may have a steeper learning curve.
MLJar is last, but definitely not least, on our list. Like the previous two tools, MLJar is highly focused on modeling, automating feature engineering, algorithm selection, documentation, and explanations.
Again, however, there’s no end-to-end solution suite to integrate the models you build into your workflow.
Implementing AI used to be a matter of careful, prolonged consideration. After all, AI projects used to cost hundreds of thousands of dollars and took several months, if not years, to get off the ground.
Today, it’s so much simpler. A non-technical employee can build and deploy AI models over their lunch break. It’s no longer make-or-break in deciding what AI tool to use, and instead, execution and creativity are key.
No matter what tool you pick, you want to have an AI use-case that’s meaningful for your organization. With Akkio, for instance, sales teams can score leads or forecast sales, marketing teams can classify customer text or reduce churn, operations teams can reduce employee attrition, and more.
Whatever your goals are, you want to have clarity on your specific use-case before selecting a no-code AI tool.
No-code AI is the most affordable way to implement AI, which has the power to optimize any organizational KPI. It’s now easier and faster than ever to build AI models.
That said, not all no-code AI tools are made the same and the right tool for you depends on your business needs. We’ve explored tools that range from just a few dollars a month, to enterprise platforms that cost six-figures a year. By gaining clarity on why you want to use AI, you can find the right tool for you.
If we’re missing a leading no-code AI tool you’d like to see featured, please let us know!