As artificial intelligence continues to reshape the software landscape, businesses are grappling with an emerging challenge: every action involving AI—from generating prompts and images to executing automated workflows—incurs a cost.
In contrast to conventional software-as-a-service (SaaS) models, where income typically correlated with user counts or subscription tiers, AI-driven solutions necessitate real-time monitoring of computational expenses linked to customer utilization. This paradigm shift compels software developers to reevaluate their strategies for managing access, usage, and pricing.
Addressing this emerging issue is Stigg, an Israeli startup launched in Tel Aviv in 2021 by former New Relic colleagues Dor Sasson and Anton Zagrebelny. The firm has created a software system that enables vendors to oversee, in real time, which customers, teams, or AI agents have access to certain features and the extent of their service consumption.
The idea for Stigg stemmed from Sasson's tenure leading AI product development at New Relic. While the company successfully rolled out AI-driven log analysis tools, it faced challenges in determining user access, establishing pricing structures, and controlling usage.
Sasson recounted that existing entitlement management had resorted to spreadsheets, revealing a significant issue within the software sector: despite advancements in product development, regulatory systems surrounding pricing, feature access, and customer entitlements remained disorganized across billing platforms, application code, and manual processes.
The swift rise of AI technology has amplified these difficulties. Traditional SaaS companies operated under a framework of charging customers based on subscription tiers or user counts. Conversely, AI solutions often require tracking multiple factors simultaneously, such as prompts, image generation, automated workflows, actions from AI agents, user credits, and spending limits.
This added complexity is driving enterprise clients to demand more control over their AI expenditures. Organizations increasingly seek to allocate AI budgets among departments, restrict individual employee or AI agent consumption, and eliminate unanticipated computing costs.
Stigg reports situations where a single user can deplete an organization's AI budget in just one day, catching administrators off guard. Earlier this year, OpenAI shared insights into how major AI developers are tackling this challenge, showcasing a system that verifies user limits, credits, and entitlements prior to processing requests. Stigg's founders noted that their platform's framework aligns closely with this model.
Recently, during the AI World Fair conference, Stigg unveiled its "Usage Runtime" platform, aimed at managing everything from tracking AI usage to customer billing. The system boasts features such as real-time usage tracking, spending controls, high-volume metering, and the flexibility for enterprise clients to implement the platform within their own cloud environments.
One notable success story comes from collaboration software firm Miro, which utilized Stigg's platform to implement an AI credit model in under six weeks—a project the company estimated would have taken around 5,000 engineering hours to develop in-house.
Stigg has also launched a complimentary version for early-stage companies, while continuing to cater to larger enterprises whose AI infrastructure needs become increasingly intricate as they scale.
Industry experts are increasingly recognizing usage management as essential infrastructure for AI businesses rather than merely a facet of billing systems. As demand for AI services continues to surge, organizations are exploring methods to prevent unexpected operational costs from escalating alongside.
