Artificial Intelligence (AI) Models recognize patterns based on previous data, which in turn can be used to make predictions, decisions, or interpretations. Although AI models are usually built based on various suitable machine learning algorithms, Zoho Creator has simplified the process by combining these algorithms into an intuitive interface that completely hides their complexity. This helps low-code developers utilize machine learning without having to worry about the underlying mechanism.

When applied on Zoho Creator, AI models can optimize business processes by forecasting future outcomes or automating operations that normally need human intervention. For example, a CRM application built on Zoho Creator can use AI to predict if a prospective client can be converted based on present and past data. This insight can be used by the sales team to devise ways to approach the client. Alternatively, a Survey application built on Zoho Creator can use AI to automatically analyze the sentiment expressed in the feedbacks.

Model Types

Zoho Creator provides support for  five common machine-learning models that can work efficiently for business applications. Some of these models are prebuilt, which are readily available to be deployed in applications, whereas others are customizable according to your requirement. Neither require any prior machine learning or coding skills to utilize them. Based on this build-type, the models are classified into:

• Ready-to-use AI models
• Custom AI models
  
• Prediction model
  
• OCR model
  
• Object detection model

The following are the model types supported by Zoho Creator:

Generally, in order to deploy an AI model, it needs to be built based on suitable ML algorithms, fed with a dataset of appropriate size and quality, trained so it recognizes patterns, and tested to ensure accurate results. For ready-to-use AI models, all these steps are properly implemented and made available in Zoho Creator applications as fields in forms. 

However, in some scenarios, AI models need to be tailor-made for a specific requirement. To satisfy this purpose, Zoho Creator's AI Modeler  can be used to build custom AI models to get desirable results. This can be viewed as a builder that lets you easily create AI models without directly working with any deep learning concepts.

For example, let's assume when a picture of a printed receipt is uploaded, all of its content needs to be fetched and be inserted in an email template. For this scenario, the ready-to-use OCR model can be used because there is no custom need and it simply extracts all available texts from the picture. Now, consider receipts of the same template will be uploaded every time, and only the IDs of the receipts need to be extracted and populated in the Receipt ID field of the same form. Here, the custom OCR model can be used to serve this purpose as the model needs to be trained for one particular receipt template. 

Business Scenarios

Let's consider a few business situations in which the AI modeler would come in handy.

1. Keyword Extraction: Auto-categorize support tickets

A delivery partner company called Zylker has built a ticketing application using Zoho Creator. They use this application to track incoming tickets and provide immaculate  customer service. Every time they receive a ticket, they'll categorize them and add tags based on the type of query. This makes it easier for the experts in the area to look into and address it. 

With Creator's keyword extraction AI model, this process can be automated. For example, if the ticket contains words like 'faulty', 'spoiled', 'damaged', 'defective', and 'refund', the ticket can be auto-tagged as 'Return & Refund'. Similarly, if 'delayed', 'missing', 'rude', and so on appears, the ticket can be tagged as 'Delivery Agent'. This helps assigning the right customer support agent to tickets and drastically reduces the manual labor of categorizing tickets. 

2. Sentiment Analysis: Analyze  customers' feedback after a service launch

Zylker's delivery company recently launched a 10 minute grocery delivery service in a couple of locations. To learn about the reach of this service and to know how customers feel about it, they circulated a survey form built using Zoho Creator. 

When there are too many survey results, the sentiment analysis AI model can be used to quickly get an overall insight into the customers' opinion about the new service. For example:

• Average percentage of customers who gave positive response can be identified.
• If most of the responses are positive, negative responses can be separated so it is easy for the team to spot the areas of improvement.

3. Optical character recognition (OCR): Digitization of business cards

4. Object detection: 

5. Prediction: Forecast selling price of a new product

A processed food manufacturing company called Zylker uses a Zoho Creator application to manage their organization process. Their research teams continuously work on new recipes and updates a form with the specifications of the recipes that are approved to be launched as a product. The marketing team has to take it from here and decide upon selling prices based on market research. In this scenario, a custom AI prediction model can make their job easier by suggesting a tentative price based on various influencing factors like market value of similar competitor products, weight, ingredients, and choice of packaging.

6.  Custom OCR: Extract order ID from sales order

Zylker processed food manufacturing company manages their orders using an order management application that is built using Zoho Creator. They supply their products to many buyers in bulk and send out auto-generated sales orders as acknowledgement. All sales orders are maintained in a form along with other essential information like order ID and buyer name.

Although maintaining records of sales orders is an efficient process, doing it manually is monotonous and inefficient. In this scenario, the custom OCR model can be trained to work exclusively on the sales order layout that Zylker issues. Upon upload of Zylker's sales order, layout-specific information like order ID and buyer name can be automatically extracted and maintained in the same form, saving time and manual labor.

7. Custom object detection: Auto-tag damage level

An automobile resale company, Zylker buys and sells used vehicles online. They embedded a Zoho Creator form in their website for the sellers to provide information about the vehicle that needs to be sold including its image. Before a proper inspection is done, a custom object detection AI model can be used to tag and categorize the proposals based on the damage level of the vehicle. For example, a vehicle with too many dents can be tagged as 'Severe damage', vehicles with a few dents can be tagged as 'Moderate damage', and so on. 

Setting up an AI model

No set up is required to deploy ready-to-use models in your application. However, for  custom models , the following procedure needs to be followed before it can be added to your applications as fields in forms.

1. Create

In the  Microservices section, choose the required custom model type to be implemented in your application. A straightforward builder is offered to create the selected AI model with custom intelligence catering to your business needs.

2. Train 

The next step is to upload training data and train the model. Training data is the initial dataset that must be fed to teach the model to identify patterns so it can make predictions or perform a required task. The principle behind a good AI model is collecting sufficient, relevant, and high-quality data to train it. Keep in mind one of the oldest and most used phrases in data science: "garbage in, garbage out". Low quality training data will compromise the accuracy of the model's outcome. Some factors that can ensure the quality of training data are:

• Relevance - Narrow down the dataset to include only values that are required.
• Validation - Make sure that the values you input your dataset are correct. 
• Consistency  - Values in your dataset must consistently follow the same format.
• Comprehensiveness - The dataset must be large enough and have the proper scope and range to encompass all of the model’s desired use cases.
• Data cleaning  - Remove or modify data that is incorrect, incomplete, or duplicated. Ensure that no missing values are present in your dataset. You can substitute the missing values with the most frequent values and in the case of numerical values, mean figures.

Once the dataset is ready, upload it and train the model. Training takes a while based on factors like model size, queued models, etc., however you can navigate away and come back later once the model is trained.

3. Test 

After successful training, the model is practically ready to be deployed. However, it is important to evaluate the model's performance and quality. You can test the model by giving sample inputs; if the test result is not satisfactory, it could be because your model is either underfitting or overfitting.

Your model is  underfitting if it performs poorly on the training data. This is because the model is unable to understand the relationship between the input and output variables. Your model is  overfitting if it performs well on the training data but produces incorrect outputs on the test data. This is because the model has memorized the data it had seen and is unable to generalize to unseen examples. 

In other words, underfitting models learn the variations in the training data rather than the patterns, whereas the overfitting models learn the training data by heart that it wouldn't perform well on any other inputs. 

Both overfitting and underfitting are problems that eventually cause poor results on new data and the key to fixing this is to regularize the training data, i.e., the training data should neither have too much variations nor have values that are too close to the expected outcome.

See how you can improve the accuracy of each custom model type:

• Prediction
• OCR
• Object Detection

4. Publish model

The next and final step to set up an AI model is to publish it. Until this point, the model is accessible only by the admins. If you are satisfied with your model, make it available for all users by publishing it. This means the model is ready to be implemented in your Zoho Creator applications.

Managing an AI model

A custom AI model , once configured, can always be edited, renamed, retrained, or deleted.  A good AI model  must be built, fed with quality data, and deployed. However, it doesn't end there. Retaining the quality of an AI model is a continuous process. Over time, as the business environment and data change, it is inevitable that AI models experience  model drift . In other words, the accuracy of model will begin to degrade. For example, a company's sales before covid and after covid are different. An AI model that is built to forecast sales must be trained considering the covid factor too.

For this reason, it's crucial to test and retrain the AI models with updated new data periodically or whenever an evident change in trend occurs. Options to  edit, rename, retrain , and  delete the model are available in the model details page. This page also shows other details like versions and applications list where the model is deployed.

Using an AI Model

To implement an AI model in your Zoho Creator applications, use one of the following methods:

• Use the required model from the  AI modeler page. You will be redirected to the application builder where the model needs to be implemented. 
• Add the required AI fields to the form. When you pick a custom model, all published models will be listed while adding the AI field.

Note: For the applications that have environments enabled, you can use custom AI model only if the application has at least one version published in the production environment. 

AI Calls

AI calls are calculated for your account based on the API calls made while accessing the AI services in Creator. 

1 AI Call = 1 API call made i.e., every time an AI model is deployed in your Creator application, an AI call is made and the required output is fetched.

Users in Ultimate/Professional plan can use upto 1000 AI calls per month. The remaining no. of requests per account can be monitored anytime from the Billing section. 

Points to note

• As the admin, you can create and use custom AI models, whereas your users can only consume the models you've created.  Any user with permission to edit your applications can indirectly utilize AI models by adding  AI fields to forms.
• As a low-code platform, Creator doesn't require you and your users to have prior coding and machine-learning skills to create, utilize, or access AI models.
• You should know enough about all your business requirements to determine and gather data that'll be used for training your custom AI models.
• It is preferred that you do an exploratory data analysis (EDA) while collecting training data for custom AI models.
• AI Modeler is available for  Professional or  Ultimate plans.  Refer to our pricing page .
• You must be using Zoho Creator 6 (C6) to be able to create custom AI models from the  Microservices section whereas, deploying the AI models as fields are supported in both C6 and C5.
• View guidelines for each model type:

• Prediction
• OCR
• Object Detection

Sample Data

Get started and build your custom AI models using the sample data we've provided in the attachment.

Related Topics

• Ready-to-use model
• Understand prediction model
• Understand OCR model
• Understand object detection model
• Understand AI fields