AI-Based Business Process Workflow Automation
Workflow automation is the act of identifying and automating repetetive tasks required to perform as a part of some larger job. It can be as simple as creating a new email address for newly hired employees, or it can be to send en email to the CEO at the end of the month with all KPIs. The point is that it varies from company to company, and no two use cases are the same.
Previously we used to do this with "workflow designers", allowing you to drag and drop "composables" around to illustrate some decision making process, resulting in automated workflows somehow solving the task at hand. Examples of such designers are Make, Zapier, and n8n. Today with generative AI, this approach seems hopelessly outdated.
The above would probably require several hours of dragging and dropping composables around in your "designer". In addition, the learning curve becomes much steeper with composables. With Hyperlambda it's literally as simple as telling Magic the following;
"Create an executable Hyperlambda function that selects some 5 KPIs from chinook database, generates KPIs, and sends these on email to thomas@ainiro.io"
5 seconds later I've got my workflow!
Hyperlambda
Hyperlambda is at the core of Magic. The point about Hyperlambda is that it's "dead simple". And when I say simple, it's literally 5 to 10 times less complex than Python. Something you can verify by measuring token consumption for some Hyperlambda snippet, and compare it towards its Python equivalent. In addition, it doesn't require any "dragging and dropping" of composables. Below is an HTTP CRUD read endpoint to give you an idea of the savings here.
This makes it an ideal "language bridge", moving from technical use cases to working code utilising code generators. Generating business process workflows using Hyperlambda is simply 100 times easier than to create the equivalent code in C# or Python or something. Basically, copy and paste your use cases into Magic's dashboard, and see the system simply solve your tasks - Whatever they are.
CRUD Generator
In addition to our natural language based AI code generators, we've also got our CRUD generator. This component allows you to literally create thousands of lines of code per second. It is not based upon AI, but rather meta programming, so it's consumes 0.01% of the electricity and resources in comparison.
With the CRUD generator you can for instance wrap your existing database into CRUD HTTP endpoints, export the OpenAPI specification, and paste it into other systems to for instance create a CRUD-based AI agent. The CRUD generator also ties into the Hyperlambda generator, allowing you to use the CRUD generator first, for then to add business logic to your CRUD endpoints afterwards.
Then later you can create natural language based AI workflows that are sequentially executing multiple such CRUD functions, and/or other functions generated using our code generators - For then to orchestrate your business workflows together using natural language. An example can be for instance.
- Create a record in x database
- Push contact to HubSpot
- Send confirmation email
- CC sales team
Where the point about the above, is that each individual item is a separate reusable Hyperlambda file, generated using our code generators, allowing you to "chain" such "actions" together into complex and rich business workflows using natural language.
Wrapping up
By combining our CRUD generator with our AI-based code generators, designing workflows visually is simply obsolete. Using natural language to "describe" your workflows is simply easier. Today it's much easier to simply ask the AI to generate code solving whatever problem at hand, and watch it "generate" your workflows instead of having to drag and drop composables around. And if you're afraid of lockins, the whole platform is open source.
If you're interested in hearing more about how we can help you automate your business workflows, you can contact me below.
