Designing for Outcomes, Not Outputs: The New Standard in AI Commercial Models

Aug 04, 2025 | min read

Roberta Silveira da Mota

We are now firmly in the era of enterprise AI—not just experimentation, but real value creation. And yet, many organisations remain tied to outdated commercial models. While the technology powering AI has leapt forward, the way we commission, contract, and measure success still often reflects a legacy mindset: one of deliverables, hours and headcount.

At CI&T, we’ve observed a fundamental shift. Business leaders are no longer asking what we can build; they’re asking what we can achieve. “Can you help us reduce fraud by 30%?” “Can we halve our time to market?” “How do we use AI to boost customer retention?” These aren’t requests for platforms or dashboards—they’re outcome-oriented goals. And they demand a new kind of engagement: more agile, more integrated, and commercially aligned to the value created, not the effort expended.

The commercial pivot: From scope to strategy

The traditional fixed-scope, time-and-materials model no longer serves the needs of organisations seeking to embed AI effectively. AI development is inherently iterative—it requires continuous learning, refinement and scaling. It is less a project, more a programme of capability-building. For this reason, models like Team-as-a-Service (TaaS) and AI Pods are increasingly becoming the standard for delivery.

TaaS offers organisations flexible access to cross-functional, high-performing teams tailored to the needs of the moment. These are not external suppliers fulfilling tasks—they are embedded, collaborative partners aligned to the client’s strategic goals. According to both Gartner and McKinsey, TaaS is fast becoming the preferred approach as enterprises grapple with the realities of talent shortages: more than 70% of companies pursuing AI initiatives report major skills gaps. TaaS helps bridge these gaps—not just by supplying talent, but by embedding delivery capability where it matters most.

AI Pods extend this concept even further. They are outcome-focused squads built around a specific business ambition—be it reducing operational costs, increasing conversion, or enhancing user experience. These pods blend engineers, data scientists, product strategists, designers and domain experts into a single, cohesive unit, united by a common goal. What makes them effective is not just their skills, but their structure: they work in agile, iterative cycles, often side-by-side with internal teams, and are measured against business KPIs—not just technical outputs.

Pricing for impact: A necessary evolution

As delivery models evolve, so too must the way we fund and measure them. There is a clear and growing shift away from hourly billing and activity-based contracts. A 2025 Gartner report revealed that over 60% of enterprise clients now prefer outcome-based pricing models for AI services. This often means blending fixed retainers with success-based fees or adopting flexible frameworks that link compensation to performance indicators.

At the same time, usage-based pricing is taking hold, particularly within software and AI-enabled platforms. Nearly 85% of SaaS firms are either implementing or trialling usage-based models, which better reflect the value generated rather than the number of users or seats. And yet, most large organisations are still not set up to commercialise AI effectively. A global survey of CFOs in early 2025 found that while 90% regard AI as business-critical, only 29% have defined monetisation strategies. The blockers are not technological—they are commercial. The solution lies in developing real-time metering, building pricing models around value creation, and ensuring tight alignment between product, delivery and finance teams.

Why this matters now

The AI landscape is moving at breakneck speed. Foundation models are improving weekly, open-source tools are democratising innovation, and regulation is becoming more sophisticated. In this context, long, rigid project timelines are no longer fit for purpose. They risk obsolescence before delivery.

What CIOs and digital leaders need today are delivery models that allow for experimentation without runaway costs; flexible teams that can adapt in real time; and commercial frameworks that reward value creation, not just volume of work. This is particularly pressing in regulated industries like financial services, where trust, explainability and compliance are paramount. The ability to deploy AI with both speed and accountability is becoming a source of strategic advantage.

Case in point: AI in financial services

Few sectors illustrate this shift more clearly than financial services. Here, AI is not a back-office enhancement—it is core to commercial strategy. At JPMorgan Chase, AI tools such as Coach AI and internal large language models (LLMs) have driven a 20% uplift in asset and wealth management sales. Advisers using these tools are projected to support 50% more clients over the next few years. The firm expects to deliver £2.0 billion (US$2.5 billion) in annual AI value by the end of 2025.

In Singapore, DBS Bank now runs more than 1,500 AI models across 370 use cases. Their AI time-to-market has dropped from 15 months to just three, thanks to internal innovations like DBS-GPT, which supports enterprise content workflows. These aren’t isolated pilots—they are embedded, at-scale capabilities delivered through outcome-oriented models.

Similarly, Commonwealth Bank of Australia has modernised its financial crime alert system using cloud-based AI visualisation tools, dramatically improving the speed and accuracy of fraud and sanctions monitoring. Mastercard’s AI infrastructure now secures over 159 billion transactions annually, increasing fraud detection rates by up to 300% and significantly reducing false declines that previously affected nearly one in five attempted transactions.

These achievements are not the result of traditional delivery models. They are the product of agile teams, aligned incentives, and partnerships designed for outcomes—not just outputs.

What it takes to be an outcome partner

The most significant transformation in AI today is not in the technology itself, but in how we work together. To be an effective partner in this space, providers must evolve beyond the role of vendor.

We must align with our clients’ strategic ambitions from the outset; we must be prepared to flex as the brief changes and the data unfolds; and above all, we must share in the risk—and the success—of the journey.

At CI&T, this is how we deliver every AI engagement. Whether through TaaS, AI Pods, value-linked pricing or embedded capability-building, our focus is always the same: sustainable business value, delivered at speed.

Conclusion: Leading the Shift

This is not a trend waiting on the horizon—it’s the reality shaping competitive advantage today. The organisations that succeed in this AI-powered economy will not be those who build the most, but those who deliver the most value. That requires rethinking how we structure teams, how we define success, and how we form partnerships.

The shift from delivering outputs to delivering outcomes is well underway. Let’s embrace it—and lead it—together.