The Dawn of a Thousand Palantir's
Agentic AI, Anthropic Computer Use, OaaS (Outcome as a Service)
There are whispers in the tech world of a new business model reshaping service-based industries.
Welcome to the era of Outcome as a Service (OaaS), a paradigm that's not just disrupting traditional Software as a Service (SaaS) but reimagining the entire concept of business services.
It’s not fully clear where the term originates from, although Foundamental, a venture capital firm specialising in architecture, engineering, construction, and supply chain (AECS) sectors, claims to have invented the term "Outcome as a Service."
What started as a niche concept is rapidly evolving into a innovation vector, catching the attention of major players in the VC world and beyond. VC firms like Sequoia Capital, Better Capital, are all hailing a major shift towards OaaS models.
OaaS has also been called ‘Service as a Software’ turning SaaS on its head.
Although. I prefer to use ‘Outcome as a Service’ or OaaS as it doesn’t mess with your head and says what its on the tin.
Despite the terminology mixup, the moniker recognizes a crucial distinction - while traditional SaaS typically targets just 1-3% of a company's P&L, OaaS has the potential to address as much as 50% by tackling core business processes such as labor, engineering, logistics and more.
“The cloud transition was software-as-a-service. Software companies became cloud service providers. This was a $350B opportunity. Thanks to agentic reasoning, the AI transition is service-as-a-software. Software companies turn labor into software. That means the addressable market is not the software market, but the services market measured in the trillions of dollars.”
Sequoia Capital
This shift is being driven by several key factors:
1. The rapid advancement of AI/ML and now LLM and now Agentic AI and reasoning technologies.
2. Increasing demand for tangible, measurable business outcomes.
3. The need for solutions that align directly with existing business processes and purchasing patterns
4. The growing complexity of business challenges that require more than just software tools to solve.
OaaS goes beyond providing tools or platforms that require user intervention and management.
Instead, it delivers the end result that users desire, often automated by AI.
This approach addresses the core need for businesses to ’get work done’ by focusing on desired ‘outcomes’ rather than the ‘tools’ used to achieve them.
Traditionally, businesses purchased software tools and hoped for the best.
They paid for features, not results.
What if instead of buying software, you could buy guaranteed business outcomes?
This is the premise of OaaS.
Instead of selling tools, OaaS providers commit to delivering specific, measurable results.
Whether it's increasing revenue by 25%, reducing operational costs by 30%, or improving customer retention by 40%, the focus is on tangible business impact.
"AI-powered OaaS solutions perform tasks autonomously, delivering results directly to users without requiring them to manage or interact with software, thus streamlining the process and saving time and effort."
Better Capital
OaaS Roots in Traditional Business
The concept of outcome-based services has its roots in older business models.
One notable example is the "power by the hour" model introduced by Rolls-Royce in the 1960s for aircraft engines. This model, which guaranteed engine availability for a fixed fee per flight hour, laid the groundwork for outcome-based thinking in industrial services.
Companies like Honeywell launched an outcome-based service model in 2017 in the connected building domain. Their success in providing facilities management services based on connected technology and condition-based monitoring has demonstrated the viability of the OaaS model in industrial settings.
The software industry's shift from on-premise solutions to Software as a Service (SaaS) in the early 2000s paved the way for more service-oriented business models. This transition inspired other industries to rethink their approach to product and service delivery.
While it's challenging to attribute the coining of "Outcome as a Service" to a single entity, it's clear that the concept has emerged from a long history of service-oriented business models and has been shaped by technological advancements and industry-specific needs.
As the model continues to prove its value, it's likely to become an increasingly important paradigm in business strategy across multiple sectors.
The Catalyst: Generative AI
While the concept of outcome-based models isn't entirely new, the rise of Gen AI is supercharging this trend. In reality it has been driven by the convergence of several trends - AI, Cloud, Big Data, IoT to name a few. These technologies have enabled companies to measure, predict, and guarantee specific outcomes with high accuracy.
As we saw in my last post, we are speeding towards a 2e30 FLOPS future which is becoming increasingly feasible. This monumental leap in computational power, approximately 50,000 times that of LLAMA 3.1 405B model, will enable the processing of vast datasets and the execution of complex algorithms at speeds previously unattainable.
The evolution of reasoning techniques in LLMs, has seen significant advancements through various approaches like Chain of Thought (CoT), Tree of Thoughts (ToT), and reinforcement learning (RL) based methods.
These developments collectively enhance AI's problem-solving capabilities, with each technique offering unique strengths. CoT improves transparency and multi-step reasoning, while ToT enables more structured exploration of multiple reasoning paths, incorporating search algorithms like breadth-first and depth-first search.
RL-based approaches, as demonstrated by OpenAI's o1 model, allow for adaptive learning and continuous improvement of reasoning strategies. The integration of these techniques results in AI systems capable of tackling complex, multi-faceted problems with greater efficiency and accuracy.
For businesses, this translates to more sophisticated decision-making tools that can handle complex scenarios in fields such as strategic planning, financial modeling, operational efficiency etc.
AI can process vast amounts of data, consider multiple outcomes, and provide nuanced insights, potentially speeding up how companies approach problem-solving, risk assessment, and long-term planning.
Anthropic’s Computer Use Breakthrough!
Anthropic's latest announcement of ‘Computer Use’ capability represents a significant advancement in the field of agentic AI and, moving from algorithms through to an AI actually interacting with a computer screen.
This new feature allows AI models to directly interact with and control computer systems. It completely changes how businesses and individuals interact with technology from hereon.
“Computer Use is the first instance of AI truly becoming embodied with access to a cursor, keyboard, computer and the web.”
Here is why Computer Use is important for OaaS:
Direct interaction: The AI can now perform tasks on a computer, such as building websites or editing spreadsheets, based on natural language instructions.
Expanded capabilities: This feature enables AI to handle more complex, multi-step tasks that previously required human intervention.
Increased automation: Businesses can potentially automate a wider range of processes, from data analysis to content creation.
Enhanced problem-solving: The AI can now access and utilize various software tools to solve problems more effectively.
Improved efficiency: By directly manipulating computer systems, the AI can potentially complete tasks faster and with fewer errors than human operators.
This technological leap will allow OaaS providers to offer more ambitious guarantees and tackle a wider range of business challenges.
Palantir: The OaaS Pioneer
Long before the current AI hype cycle, Palantir Technologies pioneered an outcome-focused approach that would later be recognized as Outcome as a Service.
Founded in 2003, Palantir differentiated itself through several key strategies:
Forward Deployed Engineers (FDEs): Palantir's unique approach involved embedding engineers directly within client organizations. This model allowed for deep integration and firsthand understanding of client challenges and industry-specific nuances.
Data Integration Expertise: At its core, Palantir's strength lay in data integration. The company developed tools to access, clean, and transform enterprise data, making it usable and accessible across organizations.
Customized Solutions: Rather than offering one-size-fits-all software, Palantir tailored its solutions to address specific client needs, ensuring relevance and maximizing impact.
Actionable Intelligence: Palantir's solutions were geared towards providing insights that drive tangible business improvements, not just data visualizations.
Continuous Adaptation: The company focused on building enduring relationships with clients, continuously refining and adapting solutions as client needs evolved.
Palantir achieved this without reliance on LLMs or Gen AI.
Their approach focused on
Operational AI focused on enhancing decision-making in complex environments
Emphasizing solid data foundations before applying AI
Human-AI collaboration, leveraging the strengths of both
Key aspects of Palantir's pioneering OaaS model included:
Two-Pronged Engineering Model: Palantir divided its engineering into two types: FDEs who worked directly with customers, and PDEs (product development engineers) who worked on the core product. This model allowed for rapid learning and product development based on real-world needs.
Overcoming Data Integration Challenges: Palantir recognized early on that data integration was not just a technical challenge but often an organizational and political one. They developed strategies to navigate these complexities.
Security-First Approach: Palantir built robust security controls into its data integration layer, addressing genuine data security concerns and often making company data more secure.
Unique Culture: And the most critical of all, Palantir fostered a culture that combined intellectual rigor with intense competitiveness. This attracted talent and created a highly generative environment for innovation. This is what I refer to as ‘cracked employees‘ in my previous 2e30 FLOPS post.
For a deep dive into Palantir philosophy, culture and approach to business read “Reflections on Palantir“ by Nabeel Qureshi (ex Palantir).
Case Study: Optimizing Supply Chain Management
To understand the practicalities of this shift lets consider a manufacturing company struggling with supply chain inefficiencies.
Here's how different approaches address this challenge:
Traditional SaaS Approach
When purchasing supply chain management software from the likes of ‘SAP‘ or similar one goes through the following:
Pay hefty upfront and ongoing subscription costs.
Spend months on implementation.
Train employees on the new system.
Gain better visibility, but the burden of improvement remains on you.
Palantir's Pre-Gen AI OaaS Approach
Palantir offers a tailored solution:
Embeds FDEs to deeply understand your specific challenges and industry nuances.
Implements Foundry, integrating all existing systems and addressing data security concerns.
Works alongside your team to identify bottlenecks and navigate organizational complexities.
Develops custom applications on Foundry, continuously refining based on real-world feedback.
Guarantees a 15% reduction in supply chain costs within a year, with ongoing collaboration for continuous improvement.
The New OaaS Landscape with Gen AI
A startup called "SupplyAI" leverages advanced AI:
Uses LLMs to quickly ingest and understand supply chain data and documentation.
Deploys AI agents and tools similar to Anthropic’s Computer Use to continuously monitor operations and integrate data from disparate sources.
Employs agent-based predictive analytics to anticipate disruptions and suggest optimizations quicker than traditional AI/ML and analytics.
Offers a hybrid model of AI automation and human expertise for complex decision-making.
Guarantees a 20% cost reduction and 30% improvement in on-time deliveries.
Pioneering OaaS Startups
As the OaaS model gains traction, several innovative startups are leading the charge, each with their unique approach to delivering outcomes:
Viridien: Offers OaaS for AI and HPC production, providing guaranteed pricing based on business outcomes. They focus on enhancing productivity and cost efficiency for complex modeling and simulation workloads.
InCommon: Delivers global talent outcomes for mid-market companies, leveraging AI across sourcing, screening, and matching to help companies build teams across countries.
Schnitt: Uses AI to drive business outcomes through rich media content creation and analysis.
SuperAGI: Their Supercoder product delivers end-to-end software engineering outcomes to businesses, leveraging an AI agent framework and large coding models.
RapidClaims: RapidClaims uses AI to automate medical coding and streamline healthcare claim submissions, helping healthcare providers reduce claim denials and get paid faster by insurance companies.
Mezink: Provides guaranteed influencer marketing outcomes to brands, leveraging AI across discovery, evaluation, activation, and reporting.
These companies represent genuine Outcome-as-a-Service models by fundamentally shifting away from providing tools or expertise to delivering guaranteed business results.
Instead of selling software or capabilities, they take full accountability for specific outcomes - whether that's successful insurance claims, completed software development, or achieved marketing results.
The key differentiator is that they've eliminated the traditional software interaction layer and user management burden, focusing entirely on delivering measurable end results while assuming the execution risk.
Their payment models directly reflect this approach - for instance, RapidClaims gets paid per successful claim approval rather than for software usage, InCommon earns fees only upon successful candidate placement and retention, and Viridien charges based on completed computational tasks rather than compute time.
This represents a major shift from the conventional ‘vertical SaaS’ or ‘specialist service provider’ model, as these companies align their entire business model and revenue structure around guaranteed outcomes rather than access to tools or expertise.
Common Principles with Palantir
These startups share several key principles with Palantir's OaaS approach:
Focus on Outcomes: Like Palantir, these startups prioritize delivering specific, measurable results rather than just providing software tools.
Deep Integration: Many of these startups, similar to Palantir's FDEs, work closely with clients to understand their unique challenges and tailor solutions accordingly.
Data-Driven Approach: All these companies leverage data integration and analysis as a core part of their offering, a principle that has been central to Palantir's success.
Continuous Improvement: Like Palantir's model of ongoing collaboration, many of these startups emphasize continuous refinement of their solutions based on real-world feedback.
Key Differences
While these startups share common principles with Palantir, there are some notable differences:
Specialization: Many of these startups focus on specific industries or functions, whereas Palantir has a broader reach across various sectors.
Scale and Maturity: Palantir is a well-established company with a track record in highly sensitive environments. These startups are at earlier stages and may not have the same level of experience or security clearances.
Technology Stack: While Palantir has developed its own proprietary platforms like Foundry, many of these startups are leveraging more recent technologies, particularly in AI/ML, LLMs and Agentic AI. This is more akin to a workflow currently than a stack.
Target Market: Some of these startups are making sophisticated solutions accessible to smaller companies, potentially democratizing capabilities that were once available only to large enterprises.
The Future: A Thousand Palantir’s Blooming
Despite the differences between emerging startups and established players like Palantir, we are entering a new age where a thousand Palantir’s will bloom.
The democratization of AI technologies and the speed of development towards 2e30 FLOPS, coupled with the growing demand for outcome-based solutions, is creating fertile ground for innovative companies to emerge across various sectors.
Each startup, in its own way, is becoming a ‘Palantir’ for its specific niche or industry.
Think verticalized Palantir’s.
These new entrants may not replicate Palantir's exact model or scale, but they embody the same core principle i.e. delivering tangible, measurable outcomes rather than just providing software tools.
This proliferation of OaaS providers is likely to accelerate innovation, drive competition, and ultimately benefit businesses across the spectrum.
As this landscape evolves, we can expect to see not just one Palantir, but hundreds or even thousands of companies adopting and adapting the OaaS model. Each will bring its unique blend of industry expertise, technological innovation, and outcome-focused approach to solve complex business challenges.
In this new era, the success of these "thousand Palantirs" will be measured not by the sophistication of their software, but by the tangible value they deliver to their clients.
Welcome to the age of Outcome as a Service, powered by Gen AI.
Finding this valuable? Pass it along to a colleague who's also figuring out the AI landscape. As Gen AI picks up speed, I'll be sharing insights more frequently to help you stay in the loop. Haven't subscribed yet? Jump in—there's plenty more practical guidance coming on how to tackle AI decisions in your organization.