Transcript
When it comes to the database, we need to consider how we structure the data of our app so that we can build an interface on top. I always recommend starting with a table for users because that is where we will have fields for their email, name, and role.
We will also be able to connect users to other objects of our app. Our app will feature specific objects for contacts, companies, and deals. This is very typical for B2B setups where you want to link multiple contacts, like a head of procurement and a CFO, to the same company.
We want different objects rather than trying to fit emails directly into the company object because it makes handling multiple contacts much easier. For deals, we create a separate object because one company can have multiple commercial opportunities, such as a pilot project followed by a second negotiation.
I have added two more objects: one for tasks and one for interactions, which are related to relationship management. Interactions will store one record per email exchange or meeting, allowing us to see everything that happened with a timestamp.
[.blog-callout]
If you are just getting started, the AI co-builder can automatically generate these database schemas for you. You just prompt the AI with your requirements, and it builds the tables and fields instantly.
[.blog-callout]
By storing these interactions, we can read through them or have AI help us understand past history. We can also bring that data into the deals and consolidate them for AI-powered insights, like summarizing interactions at the deal level.
The tasks object will be for the team to manage their to-dos. For example, a task could be following up with a contact after a demo request or once finance confirms a quote.
These six objects compose our database. The power of a relational database like Softr Databases is that we can easily interconnect them in a scalable way before building the interface.
[.blog-callout]
While Softr integrates with 17+ external sources, using native Softr Databases provides the best performance and closest integration for your application data.
[.blog-callout]
I have prepared a detailed list of all the fields needed for each object, such as the deal amount, close dates, and interaction dates. These columns are essential for every table.
In the Softr Studio, I will start by hitting create an app and choosing Softr Databases as the data source. You could connect other sources or services with a REST API, but Softr Databases offers the best performance.
Softr automatically creates the first blocks and utility pages. I will connect a dynamic block to our new database while giving a clear name to both the app and the database components.
I will add one table per object: users, contacts, companies, deals, interactions, and tasks. Now we have our six tables which match our database design.
Next, we create the fields for each table and the connections between them. The first column is always the primary field, where values should be unique. For users, I use the email address.
I will also add a short text field for the full name. For the user roles, we will use a select field with options for admin and sales reps. We can also add a file field for an avatar profile picture.
We will now create a related record in the relations section to connect to our contacts table. This field, called assigned contacts, allows us to link specific records together seamlessly.
I will add more relations for deals and tasks. I also recommend adding a record ID field. In Softr Databases, every row has a unique identifier that helps identify specific records even if names are duplicated.
We can add dummy data to test the interface. Moving to the contacts table, I will use an email field as the primary field and add a connection called contact owner to define which sales rep owns the contact.
Additional fields for contacts include first name, last name, job title, and phone number. We also want to connect contacts to companies so we know where they work.
In this CRM, we will restrict a contact to only one company. I will also add a related record for interactions and keep the toggle on so a contact can be linked to multiple email exchanges or meetings.
For the companies table, we will have the company name, a URL field for their website, and a select field for the industry. This allows users to filter companies by specific industries in the interface.
[.blog-callout]
To manage these connections and view your relationships visually, check out our AI CRM template which comes pre-configured with these relational structures.
[.blog-callout]
Creating your own CRM is great because you can use your own taxonomy instead of being stuck with default categories. Now we will look at Softr's unique AI agent field.
This field can automatically generate insights based on AI whenever a record is created. For example, we can write an AI prompt to get a concise description of a company's activity using the company name and website.
By allowing web search, the AI agent finds online information to fill the field. We can then decide when the agent should run, such as only when a website URL is available.
If I add a company like IBM with its website, the AI agent automatically runs and generates activity bullet points. We can refine the prompt formatting to get the perfect result.
We can also use AI to find the number of employees. While internet data may not be exact, it gives a good idea of whether the company is an enterprise or a small business.
This helps us multiply the relevance of the database without manual work. After setting up the company links, I will add a related record connection between companies and deals.
For the deals table, we add properties like the ARR amount using a currency field in US dollars. We also need a pipeline using a select field for the deal stage, such as new opportunity or negotiation.
Our CRM will host a demo request form. When a prospect submits this, we will store their email, name, and message. We will later use an automation to check if this contact already exists.
[.blog-callout]
Instead of using third-party tools, you can use native Softr Workflows to handle the logic of checking for existing contacts and creating new records automatically.
[.blog-callout]
I will also add date fields for the demo request and the close date. We will link tasks to the deal level and add a file field for storing quotes or slide decks.
I want an interaction digest field at the deal level. We will build an automation so that a button click finds all connected interactions and creates an AI summary of where the deal stands.
In the interactions table, we will have a select field for the type, such as email or meeting. The description field will hold the content, and we can track if the interaction was inbound or outbound.
I will use a formula to automatically generate the title of the interaction based on the direction, type, and contact. This keeps the database organized and readable.
Finally, we prepare the tasks table with a title, description, and status. Our V1 database structure for the AI CRM is now ready with well-connected objects.
[.blog-callout]
If you need a specialized view that isn't available in the standard library, the Vibe-Coding block lets you prompt for custom components that connect directly to your database.
[.blog-callout]
This database is now AI-ready, allowing us to build workflows that fill fields or send emails automatically. I will add more placeholder data so we can move to the interface building phase.
