Artificial Intelligence (AI) is poised to violently disrupt the Information-as-a-Service (IaaS) industry. The announcement this week that Google’s Deepmind project GenCast performed better than the ENS forecasts produced by European Centre for Medium-Range Weather Forecasts (ECMWF) was a shot heard around the world.
Predicting day-to-day weather, paths of hurricanes and cyclones is the most complex Information as a Service (IaaS) business imaginable. It requires state of the art supercomputers, algorithms few humans can comprehend and vast treasure troves of data.
Google’s Gencast performed 20% better than the best weather forecasting system humanity could produce.
AI tools have become increasingly adept at collecting, processing, and generating valuable insights from data. These activities lie at the heart of IaaS business models.
Large language models (LLMs) and agentic AI solutions from major players (like ChatGPT and Google’s Gemini) threaten to commoditize the core offerings of specialized information providers.
Is this the end for IaaS? Not necessarily.
If companies understand the nature of AI’s disruption and adapt with urgency, they can secure success not just survive.
The Existential Threat of Agentic AI
One of the most critical disruptions stems from agentic AI platforms. These tools strive to become a one-stop shop for information by consolidating vast datasets and providing user-friendly interfaces.
As these platforms mature, they will give users near-instant access to insights and forecasts, diminishing the need for specialized IaaS providers who historically offered data and analysis as a paid service.
Agentic AI solutions even allow clients to create their own “AI agents” that replicate tasks once performed by IaaS companies.
For instance, if an AI agent is trained to compile specialized data and generate niche reports, the creator of that agent can distribute the agent widely, making once-costly information available for free.
The barrier to producing professional insights plummets, and with it, traditional revenue models in the IaaS space face a direct threat.
Feeding the Beast – Strip-Mining and Conflicts of Interest
Even if an IaaS provider decides to join forces and integrate AI into its service, the strategy can backfire.
Services like ChatGPT or Google’s Gemini train their models on the queries users submit and any data provided.
In effect, an IaaS provider using ChatGPT or Gemini to enhance its service will systematically expose all of its proprietary data to the AI, feeding the beast that’s poised to destroy them.
Some companies may consider AI solutions like Amazon’s Bedrock, which promise not to use customer data for model training.
But these same tech giants have a long history of replicating successful partner models and rolling out competing offerings—often with far greater scale and resources. Relying on these solutions may spare your data from being strip-mined, but you could still be handing over your business playbook to a future competitor.
Companies providing AI services are in the same phase of development the Internet was in the early 2000’s. These companies are burning through billions of dollars and they are hungry for business models that will pay the bills. Yours could be next.
As Warren Buffet said, “If you’ve been playing poker for half an hour and you still don’t know who the patsy is, you’re the patsy.” If you’re an IaaS business and you’re playing with AI without taking the right precautions, you may be the “patsy”.
Eroding Value Proposition
Traditional differentiators of IaaS companies, such as proprietary content or expert interpretations, are losing their edge. AI models trained on large swaths of domain-specific data can replicate (or nearly replicate) many research and analytical tasks.
AI’s accessibility makes it incredibly easy for customers to question why they should pay a subscription for insights they might get for free—or close to it—by querying a general-purpose AI.
While generic AI tools might not perfectly match the depth and accuracy of bespoke industry reports, the perceived “good enough” factor risks chipping away at the IaaS market.
When customers see AI generating near-instant results that appear on par with curated reports, the pricing advantage of a paid IaaS product becomes increasingly difficult to justify.
AI’s Weaknesses
Although AI poses an existential threat, it’s not invincible. AI has several weaknesses.
For IaaS companies willing to double down on quality and innovation, AI’s inherent weaknesses can be turned into opportunities:
- Query Engineering Complexity: Effective use of AI depends on how well users engineer their queries. Complex business applications often require advanced prompt design and workflow orchestration to yield useful results. This opens the door for specialized service providers to offer AI consulting and integrated workflows that eliminate user guesswork.
- Limited Access to Proprietary Datasets: Public AI models thrive on publicly available data. They don’t have access to specialized, private datasets unless given explicit permission. IaaS providers can leverage proprietary data and domain expertise to deliver validated, high-accuracy outputs that generic AI simply can’t match.
- Validation and Accuracy: AI models can’t always validate their own outputs. They may provide erroneous results because they lack an inherent mechanism to weigh or rank the credibility of the data they process. An IaaS company can incorporate robust quality control steps, using proprietary knowledge or curated workflows to fact-check AI outputs.
Measures for IaaS Providers to Counteract AI Threats
1. Evolve from Data Vending to Results Delivery
IaaS companies must shift their focus from simply selling information to providing tangible results. By understanding and integrating how clients actually use the data, you can package insights in a format that speeds decision-making.
For example, instead of delivering spreadsheets or PDF reports, consider offering automated dashboards or AI-driven analytics workflows that solve specific business problems.
IaaS companies need to look for ways to move deeper into the client’s workflow. By offering higher degrees of customization, the IaaS provider can make it harder for someone querying an AI to duplicate the results.
2. Implement Private AI Environments
To safeguard proprietary data, deploy self-hosted open-source AI models rather than relying on third-party platforms.
Owning your AI environment end-to-end protects against data leaks and ensures you’re not helping train a future competitor. This approach also allows for advanced customization and tight integration with your existing datasets.
3. Build Accurate Validation Workflows
Accuracy becomes a major differentiator in the era of AI commoditization.
Embed rigorous validation steps—powered by your proprietary knowledge base—into the AI pipeline so clients can trust the output. Provide transparent quality checks, or have subject matter experts sign off on critical insights before they reach end users.
4. Enhance Your Data Ecosystem
Supplement your proprietary datasets with curated external sources by implementing secure web-crawling and data ingestion pipelines. Integrate these sources into a private, high-performance repository (such as a vector database). By enriching your data environment, you can generate deeper, more comprehensive insights than generic AI.
5. Offer NLP-Driven Interfaces and Confidentiality
Customers increasingly expect AI-style, natural language-driven experiences.
Failing to offer an AI front end risks seeming antiquated. Deliver user-friendly AI interfaces while ensuring client queries remain private and fully controlled within your own infrastructure. This addresses growing concerns about data confidentiality and prevents “strip-mining” by public platforms.
Conclusion
The IaaS industry stands at a crossroads. Either embrace AI’s disruptive potential through strategic adaptation or risk irrelevance in a rapidly changing market.
About Verlicity AI
At Verlicity we specialize in incorporating real-time data into Self-Hosted AI solutions. Our platform is designed to keep your proprietary information confidential while delivering real-time insights from domain-specific and client data.