Globant, a large tech consulting firm, has torn traditional billing strategies to embrace AI use cases and projects more fully.
This is the trend I told you about my return to April. The Generation AI boom has encouraged businesses to adopt a variety of pricing models, as the latest models are becoming so expensive.
This has already become entrenched in the SaaS industry. Now it's starting to happen in consulting.
A competing IT services expert with Accenture and other consulting companies, Globant deploys AI Pods, a subscription to AI-powered offerings such as software engineering, product definition, design, and testing.
Clients can subscribe to AI PODs each month, with each subscription containing token-based capacity, similar to how some of the major AI lab structures are used.
Whenever a company uses a major AI model, prompts and other requests fall under the new language of generating AI, “Tokens,” which it needs to be processed. When complex projects are handled this way, the cost of token processing can be added quickly, especially when new “inference” models are used.
Traditional consulting projects often come with a time-collecting billing structure. The work is scoped, employees are assigned and time is tracked to calculate what to bill clients. However, in the world of generation AI where output is driven by token-based usage and cost, it depends on the model used. Older approaches can create inconsistencies between providers and clients.
“This is a departure from a traditional effort-based approach,” explained Martín Migoya, CEO on Earth, describing the move as “a fundamental departure from what others in our industry are offering.”
Tokens and effort
Globant's new AI pod package This new reality is packaged into a managed service model. For monthly subscription fees, clients access a virtual workforce of AI agents overseen by Globalt Human Tech consultants. Each subscription comes with a specific number of tokens to use.
This is just a theoretical example and doesn't represent what Globant actually bills its clients, but we get the point. Clients can pay $20,000 a month and include 100 million tokens in their subscription.
“My business is creating software for you and creating technology for you,” Migoya added in an interview. “We are proposing new ways of engagement that are more connected to throughput, not just the time and effort I'm consuming.”
Migoya said that because of the high token limit, Globant Consultants and its clients are high enough to complete the project. If some customers blow up the token allocation, Globant will resolve the solution, but the goal is to stabilize pricing in this new generative AI era and minimize surprises, he explained.
“Streaming” consulting service
He calls it “streaming engineering,” and likens it to a way of paying a monthly fee to stream Netflix movies and TV shows. Instead of spinning up custom teams for each client's needs, customers now subscribe to digital capacity.
Related Stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
Under the hood, the AI pod is powered by the Global Enterprise AI platform, which utilizes a variety of major AI models.
The company also uses CODA, a proprietary agent framework that ensures high quality enterprise output while automating aspects of code generation, testing and deployment. This approach allows Grobant to reuse existing software components, reducing the need to ping the underlying AI model (reducing the cost of handling these tokens!).
Globant says the new model is already being used by financial, entertainment and energy clients. Early recruits include oil and gas company YPF and holding company JM Family Enterprises.
Although subscription-based and token-based models are known in the SaaS world, Globant's movement is a signal that enterprise consulting may be heading in the same way.
The largest consulting firms include Deloitte, EY, PWC and KPMG. We invest heavily in automation and AI Over the past few years. recently, Deloitte And EY has unveiled a new Agent AI platform built in collaboration with NVIDIA.
By redefineing how AI builds and deploys software, it may feel as outdated as the dial-up internet simply by billing based on human time and effort.
