As an engineer, I worked with many startups over the years. The main thing I started telling people who reached out to me and asked if my team could help them build their product after learning what they wanted to do was not to hire us. Why? Because making software products is extremely expensive, and for a startup that still needs to validate the idea, it’s not the right thing to do at the early stage. First, it makes sense to validate the concept and only then start building the product with minimal expenses.
Similarly, for companies looking to improve their team’s productivity and reap the benefits of streamlined internal workflows and data collection, building custom internal tools is too much of an investment, at least initially. Hiring a company like ours to build software or an in-house tech team is a serious endeavor, and there are other things to focus on initially. Most of the processes can be at least prototyped and tested using existing off-the-shelf solutions, which will be a much cheaper and faster way to get started. Let’s look at some of them.
Spreadsheet as a Database
First and foremost, you need a databse. For example, Google Sheets is one of the most popular tools for prototyping business software. There’s a saying that every SaaS app is created to replace a process that companies typically have in Excel. Well, Google Sheets is much like Excel but in the cloud. They have most (if not all) formulas Excel has, anyone on your team can access them, plus they have flexible access control if you want fine-grained permissions.
Most processes can be modeled as a spreadsheet. But the best thing is that you can connect any number of Google Sheets, turning them into a visual relational database. You can even query them as a database, write and read from a spreadsheet much like you do with a database, and keep using the UI for manual processes. You can also connect forms to collect the data.
Of course, Google Sheets has its limitations, one of the biggest being the UI. If that’s not enough, you can use Airtable, a spreadsheet on steroids. Airtable gives you more than just a spreadsheet view — you can display your data as a Kanban board, as a timeline, as a calendar view, and add special rules for each view to be helpful for a specific task. Moreover, Airtable recently added a feature to sync their database with a regular PostgreSQL database, making it even more valuable for future integrations with other tools.
Automation Using Zapier and Make
Once you have your database, you can start making things more interesting by automating some tasks your team performs manually. Automation rules range from simple one-step scenarios of creating an entry in Google Sheets or Airtable when a client sends you an email to further process it to complex procedures that include looking for attachments in that email, uploading it to Google Drive, converting to text, generating a summary using an LLM and adding that summary to the spreadsheet along with the outline of the email itself.
Zapier is quite simple to start with and has many valuable features. You can have multi-step scenarios that vary based on the conditions you define (e.g., the email in the example above can be handled differently based on the sender if you have an important client that requires special treatment). You can also run your zaps on a schedule. For example, once a week, you can look at all the client inquiries added to the spreadsheet and generate a report showing the ones that still need to be handled.
Make (formerly Integromat) is a more advanced tool with even more robust features than Zapier but a steeper learning curve. It has a visual automation builder where you can see all the steps in your scenario on a single screen and even see how scenarios are executed live. It makes editing the scenarios and finding problems much more effortless. It also includes some useful features like aggregating and merging data arrays, which solves some of the issues Zapier has.
Notion for Textual Data
Spreadsheet or database-based application design is excellent, but it only works for some use cases. Every team needs a way to work with textual data, be it internal team documentation, meeting notes, or SOPs. This is precisely the problem Notion solves. Everything is a document in Notion, which can be organized however you want. It also includes a powerful feature to refer to documents right in the middle of each other or create news ones on the fly from a block text.
Notion also has a concept of “databases,” where each is a collection of documents of a particular type. You can define all the attributes a document has, and choose between different views to present the documents, much like you do in Airtable. But while Airtable’s database is more of a spreadsheet with different views, Notion’s database is better used to organize the documents.
Let’s further extend the example above. Let’s say we’re building a helpdesk for your clients because off-the-shelf products don’t work well anymore, and you need something custom. A client sends an email to support@example.com, which triggers a Zapier automation. This automation adds a row to an Airtable database, marking this inquiry as new. It also generates a PDF version of the email and saves it to GDrive. Finally, the automation adds a summary of the email and a link to the GDrive folder to the same row in Airtable.
Now, if your support team wants to collaborate on that inquiry, the best they can do is write comments on Airtable, which is quite limiting. Instead, you can use Notion for that scenario and add a document to the “Inquiries” database with the attributes you’d create in Airtable. Then, your support team can add more details to the document as they work on the support request. They can leave comments on individual blocks in the document, refer to other inquiries, and do many more things Notion allows you to do.
I merely scratched the surface here, showing you a few tools requiring zero coding. Even the simple scenario of modeling a support workflow with a Gmail mailbox, a Notion or Airtable database, and automation created in Zapier or Make would take a team of developers a few months to implement.
It’s not because developers are slow or lazy. Mostly, engineers must follow specific processes to deliver reliable and maintainable software to you, the end user. The products I listed above handle the reliability for you, plus they give you a simple UI to create everything you need yourself.
Of course, everything comes with a price. No-code tools have limitations, and some things are impossible, at least yet. Also, using multiple tools at a time and maintaining and editing all of the automation can become cumbersome and sometimes unmanageable.
But if you’re starting with building internal tools, it’s a great way to start, and it will likely be a very long time until you hit any of the limitations.
Originally published on Medium.com
