The Death of RPA, Fundraising Benchmarks, AGI Timelines, and How to Ship
It's easy to want the results. It's harder to pay the costs.
Good morning
Back to the previous schedule and on to some new exciting stuff. As we wrap up 2024, as I do every year, there will be more about current and 2025 trends in coming editions.
In today's edition, among other things:
The Death of RPA
State of European Tech 2024
Startups Fundraising Benchmarks
AGI Timelines
The Role of Humans in an AI-Driven Talent Acquisition Process
How to Ship
The Future of Programming
State of AI Agents
Onwards!
The Death of RPA
From 2010 to 2020+, Robotic Process Automation (RPA) dominated the conversation about process automation at companies. The sales pitch worked: software (non-AI) would handle repetitive computer tasks, letting employees focus on more important work. Companies spent billions. But then AI happened.
RPA is software that replicates human actions on computer systems. It creates programs (often called "bots") that perform the same repetitive tasks that office workers do every day. Here's what RPA bots actually do:
Click buttons and fill forms
Copy and paste data between applications
Read and process structured data from documents
Navigate between different software applications
Follow precise, predefined rules to complete tasks
Work 24/7 without breaks or errors
Think of RPA as a virtual workforce that sits at virtual computers, using the same interfaces that human employees use. For example, an RPA bot might:
Log into an email system
Download attachments from specific senders
Extract data from these attachments
Input this data into an enterprise system
Send confirmation emails once complete
Companies adopted RPA because it automated tasks without needing to change their existing software systems. The most common uses include:
Data entry and validation
Legacy system integration
Report generation and distribution
Invoice processing
Employee onboarding documentation
Claims processing
Customer record updates
Major players like UiPath rode this wave to billion-dollar valuations, promising to deliver a "fully automated enterprise." The technology gained massive adoption because it could provide quick wins without requiring massive IT overhauls.
While RPA proved valuable for many organizations, its fundamental limitations became increasingly apparent: limited scope and capability, high maintenance requirements, implementation challenges, scalability issues etc.
The emergence of LLMs and advanced AI has quickly changed the automation landscape. Unlike traditional RPA, AI-powered automation can:
Understand Context: Process unstructured data and adapt to changing conditions
Make Decisions: Apply judgment to handle edge cases and variations
Learn and Improve: Get better at tasks over time through continuous learning
Handle Complex Tasks: Take on sophisticated processes that require understanding and reasoning
Work More Independently: Operate with less human oversight and intervention
Here’s more from a16z:
Instead of hard-coding each deterministic step in a process, AI agents will instead be prompted with an end goal (e.g., book an appointment for the customer, transfer data from this document into this database), and then be empowered with the right tooling and context to take those actions on behalf of the company. They’ll be adaptable to various data inputs and capable of handling changes in business processes. And because of this flexibility, they will be far easier to implement and maintain than traditional RPA systems.
And the market is already growing:
It’s a shift from a bot-driven process that requires explicit programming of every step to AI automation that can be given high-level goals and figured out the steps.
This doesn't mean the end of automation - the opposite. It means almost everything can and will be automated with a power grab opportunity for companies that will question how they work sooner than others.
State of European Tech 2024
Keep reading with a 7-day free trial
Subscribe to Bartek Pucek to keep reading this post and get 7 days of free access to the full post archives.