General Ledger, Accounts Receivables & Payable Solutions

"We have started in the last 20 years to deal with computers in higher and higher levels of abstraction but ultimately these levels of abstraction get translated down into these stupid instructions that run really fast” Steve Jobs, 1983

During the last 40 years, since the launch of the Macintosh in 1984, software design has been stuck in a point-and-click framework that heavily relies on a combination of mouse and keyboard operations. We believe that the time and technology is now finally here to break free of the paradigm to allow a more fluent user interface that is not constrained by the boxes and schematics of the graphical user interface.

"We have started in the last 20 years to deal with computers in higher and higher levels of abstraction but ultimately these levels of abstraction get translated down into these stupid instructions that run really fast” Steve Jobs, 1983

During the last 40 years, since the launch of the Macintosh in 1984, software design has been stuck in a point-and-click framework that heavily relies on a combination of mouse and keyboard operations. We believe that the time and technology is now finally here to break free of the paradigm to allow a more fluent user interface that is not constrained by the boxes and schematics of the graphical user interface.

"We have started in the last 20 years to deal with computers in higher and higher levels of abstraction but ultimately these levels of abstraction get translated down into these stupid instructions that run really fast” Steve Jobs, 1983

During the last 40 years, since the launch of the Macintosh in 1984, software design has been stuck in a point-and-click framework that heavily relies on a combination of mouse and keyboard operations. We believe that the time and technology is now finally here to break free of the paradigm to allow a more fluent user interface that is not constrained by the boxes and schematics of the graphical user interface.

MS-DOS and the Terminals

Fig 1. The quick and dirty MS-DOS interface

MS-DOS and the Terminals

Fig 1. The quick and dirty MS-DOS interface

In the early days of personal computing, the command-line interface (CLI) reigned supreme, with MS-DOS dominating the PC market throughout the 1980s and early 1990s. Users interacted with computers by typing specific commands, following a workflow that involved entering a command, pressing Enter, reading the text output, interpreting the results, and deciding on the next command based on the output.

This interface required users to memorize commands and their syntax, with efficiency heavily dependent on the user's knowledge and typing speed. Error messages were often cryptic, leading to troubleshooting through trial and error. Advanced users could create batch files to automate sequences of commands.

While this interaction model demanded a significant learning curve, it offered precise control for those who mastered it.

In the early days of personal computing, the command-line interface (CLI) reigned supreme, with MS-DOS dominating the PC market throughout the 1980s and early 1990s. Users interacted with computers by typing specific commands, following a workflow that involved entering a command, pressing Enter, reading the text output, interpreting the results, and deciding on the next command based on the output.

This interface required users to memorize commands and their syntax, with efficiency heavily dependent on the user's knowledge and typing speed. Error messages were often cryptic, leading to troubleshooting through trial and error. Advanced users could create batch files to automate sequences of commands.

While this interaction model demanded a significant learning curve, it offered precise control for those who mastered it.

In the early days of personal computing, the command-line interface (CLI) reigned supreme, with MS-DOS dominating the PC market throughout the 1980s and early 1990s. Users interacted with computers by typing specific commands, following a workflow that involved entering a command, pressing Enter, reading the text output, interpreting the results, and deciding on the next command based on the output.

This interface required users to memorize commands and their syntax, with efficiency heavily dependent on the user's knowledge and typing speed. Error messages were often cryptic, leading to troubleshooting through trial and error. Advanced users could create batch files to automate sequences of commands.

While this interaction model demanded a significant learning curve, it offered precise control for those who mastered it.

Fig 2. Illustrative schematic of the human-machine interface of MS-DOS. The user had direct access to operate the underlying database by using the terminal, which could potentially lead to crashing the disk.

Fig 3. The early Mackintosh interface with folders and icons, making things more intuitive.

The introduction of graphical user interfaces (GUIs) marked a revolutionary change in human-computer interaction. The mouse became a crucial input device, introducing the "point and click" paradigm. Users now interacted with visual elements on the screen, moving the mouse to navigate and clicking on icons or menu items to perform actions.

The "desktop" metaphor provided a familiar context for organizing files and applications. Users could open multiple windows, switch between tasks visually, use menus to explore available functions, and interact with graphical controls like buttons, checkboxes, and sliders. The learning curve shifted from memorizing commands to understanding visual cues and metaphors, making computers more approachable for novices. The onboarding process became much easier, leveraging well-known design mnemonics such as folders, desktops, and documents.

The introduction of graphical user interfaces (GUIs) marked a revolutionary change in human-computer interaction. The mouse became a crucial input device, introducing the "point and click" paradigm. Users now interacted with visual elements on the screen, moving the mouse to navigate and clicking on icons or menu items to perform actions.

The "desktop" metaphor provided a familiar context for organizing files and applications. Users could open multiple windows, switch between tasks visually, use menus to explore available functions, and interact with graphical controls like buttons, checkboxes, and sliders. The learning curve shifted from memorizing commands to understanding visual cues and metaphors, making computers more approachable for novices. The onboarding process became much easier, leveraging well-known design mnemonics such as folders, desktops, and documents.

Fig 4. With the Graphical User Interface, interfacing with the database is now done via two layers of abstraction, first an API layer that translate codified inputs into backend operations on the database, and on top of the API layer several UI layers that allows the user to create API calls by clicking specific buttons.

Trapped in the designer's maze

The core drawback of the point-and-click paradigm has always been that the workflows or action available to the user needs to fit into the mental model the software designer has envisioned. For any workflow or action the user wants to achieve that does not fit the mental model of the designer, completing the actions will involve a long series of clicks and entering of the same information to achieve the output.

Take Google Workspace as an example. It’s essentially email and calendar that is browser based. In Google Workspace the mental model is 1 tab for email and 1 tab for calendar. You use the email tab for sending and receiving emails, and you use the calendar to manage appointments/meetings.

Say Bob wants to schedule a meeting with Jane. Bob emails Jane asking her for her availability, and gets a reply that Jane can do next week Thursday at 09:00. Bob reads the email and needs to tab over to his calendar to see if he is available at that time before clicking back to email, reply to Jane, and click back to the calendar to create the meeting invitation.

This phenomenon occurs across all work software, where you need to open multiple tabs, click back and forth to copy and paste data between flows and has always been a fundamentally broken flow.

Natural Language Interface

With the advent of large-language models and AI, we have a phenomenal new technology that enables the translation of natural language input into API calls.

This allows us to create a completely new product experience where users can express themselves in natural language and manipulate multiple objects across the interface.

This combines the expressive capabilities of terminals with the ease of use of point-and-click products for a much better product experience. Instead of Bob clicking back and forth between tabs, he simply writes “if I am free on Thursday send a reply to Jane that that works, and create a meeting invite for 30 minutes” and the LLM can create two API calls to both the email (reply to Jane) and the calendar apps (create meeting).

In more complex business products like ERPs, CRMs and HR software suites, the power of the natural language interface grows exponentially. “Reallocate the 100,000 USD invoice from ServiceNow from our US to our Swedish entity” can create all the accounting operations to move a cost between two group entities, ensuring that all of the dates, amounts, cost centers, taxes etc are entered consistently so the transactions reconcile.

Trapped in the designer's maze

The core drawback of the point-and-click paradigm has always been that the workflows or action available to the user needs to fit into the mental model the software designer has envisioned. For any workflow or action the user wants to achieve that does not fit the mental model of the designer, completing the actions will involve a long series of clicks and entering of the same information to achieve the output.

Take Google Workspace as an example. It’s essentially email and calendar that is browser based. In Google Workspace the mental model is 1 tab for email and 1 tab for calendar. You use the email tab for sending and receiving emails, and you use the calendar to manage appointments/meetings.

Say Bob wants to schedule a meeting with Jane. Bob emails Jane asking her for her availability, and gets a reply that Jane can do next week Thursday at 09:00. Bob reads the email and needs to tab over to his calendar to see if he is available at that time before clicking back to email, reply to Jane, and click back to the calendar to create the meeting invitation.

This phenomenon occurs across all work software, where you need to open multiple tabs, click back and forth to copy and paste data between flows and has always been a fundamentally broken flow.

Natural Language Interface

With the advent of large-language models and AI, we have a phenomenal new technology that enables the translation of natural language input into API calls.

This allows us to create a completely new product experience where users can express themselves in natural language and manipulate multiple objects across the interface.

This combines the expressive capabilities of terminals with the ease of use of point-and-click products for a much better product experience. Instead of Bob clicking back and forth between tabs, he simply writes “if I am free on Thursday send a reply to Jane that that works, and create a meeting invite for 30 minutes” and the LLM can create two API calls to both the email (reply to Jane) and the calendar apps (create meeting).

In more complex business products like ERPs, CRMs and HR software suites, the power of the natural language interface grows exponentially. “Reallocate the 100,000 USD invoice from ServiceNow from our US to our Swedish entity” can create all the accounting operations to move a cost between two group entities, ensuring that all of the dates, amounts, cost centers, taxes etc are entered consistently so the transactions reconcile.

Trapped in the designer's maze

The core drawback of the point-and-click paradigm has always been that the workflows or action available to the user needs to fit into the mental model the software designer has envisioned. For any workflow or action the user wants to achieve that does not fit the mental model of the designer, completing the actions will involve a long series of clicks and entering of the same information to achieve the output.

Take Google Workspace as an example. It’s essentially email and calendar that is browser based. In Google Workspace the mental model is 1 tab for email and 1 tab for calendar. You use the email tab for sending and receiving emails, and you use the calendar to manage appointments/meetings.

Say Bob wants to schedule a meeting with Jane. Bob emails Jane asking her for her availability, and gets a reply that Jane can do next week Thursday at 09:00. Bob reads the email and needs to tab over to his calendar to see if he is available at that time before clicking back to email, reply to Jane, and click back to the calendar to create the meeting invitation.

This phenomenon occurs across all work software, where you need to open multiple tabs, click back and forth to copy and paste data between flows and has always been a fundamentally broken flow.

Natural Language Interface

With the advent of large-language models and AI, we have a phenomenal new technology that enables the translation of natural language input into API calls.

This allows us to create a completely new product experience where users can express themselves in natural language and manipulate multiple objects across the interface.

This combines the expressive capabilities of terminals with the ease of use of point-and-click products for a much better product experience. Instead of Bob clicking back and forth between tabs, he simply writes “if I am free on Thursday send a reply to Jane that that works, and create a meeting invite for 30 minutes” and the LLM can create two API calls to both the email (reply to Jane) and the calendar apps (create meeting).

In more complex business products like ERPs, CRMs and HR software suites, the power of the natural language interface grows exponentially. “Reallocate the 100,000 USD invoice from ServiceNow from our US to our Swedish entity” can create all the accounting operations to move a cost between two group entities, ensuring that all of the dates, amounts, cost centers, taxes etc are entered consistently so the transactions reconcile.

Fig 5. Illustrative learning curve and output with terminal products (red), point-and-click (blue) and natural language products (green)

Leveraging natural language enables a fast onboarding to the full capabilities of command language. As the AI is trained on the entire API space of the underlying application, can do retrievals from the database before executing a command etc, it enables a new level of productivity than what was previously feasible for ordinary users to achieve.

Fig 6. Natural Language Interface allows the user to interact with all the available APIs of the product anywhere, all at once.

Light
Command Interface

Light Command
Interface

Light
Command Interface

Light
Command Interface

The death of the point-and-click interface

Looking for something else

Accounts Payable

Expenses

Cards

Procurement

Accounts Receivable

Vendor Management

Carbon Footprint

Accounts Payable

Expenses

Cards

Procurement

Accounts Receivable

Vendor Management

Carbon Footprint

Accounts Payable

Expenses

Cards

Procurement

Accounts Receivable

Vendor Management

Carbon Footprint

LightCommand Interface

Light CommandInterface

LightCommand Interface

LightCommand Interface

The death of the point-and-click interface

Looking for something else

Accounts Payable

Expenses

Cards

Procurement

Accounts Receivable

Vendor Management

Carbon Footprint

Accounts Payable

Expenses

Cards

Procurement

Accounts Receivable

Vendor Management

Carbon Footprint

Accounts Payable

Expenses

Cards

Procurement

Accounts Receivable

Vendor Management

Carbon Footprint

Light
Command Interface

Light Command
Interface

Light
Command Interface

Light
Command Interface