Cognizant’s first-quarter results suggest that Ravi Kumar S has moved the company beyond the defensive phase of its turnaround. Revenue rose 5.8 per cent year on year to $5.413 billion, or 3.9 per cent in constant currency. Quarterly bookings increased 21 per cent, trailing twelve-month bookings reached $29.6 billion and seven large deals were signed. Management retained constant-currency revenue-growth guidance of 4 to 6.5 per cent for 2026 while raising the adjusted operating-margin range to 16 to 16.2 per cent.
The numbers show renewed commercial momentum, particularly in financial services. They also raise the standard for Cognizant’s AI builder strategy. The company argues that enterprises face an AI velocity gap between the money they invest and the business results they realise. Cognizant wants to bridge that gap through engineering, industry expertise and partnerships. The proposition is credible because most large companies do not lack access to models; they lack modern data, redesigned processes, governance and the ability to operate systems safely at scale.
For Kumar, the challenge is to make that role economically distinctive. Every global services competitor claims to help clients move from pilots to production. Cloud and software providers are building more implementation capability, while clients are hiring their own AI teams. Cognizant needs to prove that its combination of technology operations, engineering and sector knowledge accelerates outcomes enough to support pricing, expansion and margin.
Bookings are a test of delivery, not a victory over it
The 21 per cent quarterly increase in bookings and more than 70 per cent growth in large-deal contract value indicate stronger sales execution. Large programmes can reshape Cognizant’s position inside clients, particularly when they combine modernisation, cost reduction and AI-enabled operations. But bookings create future obligations before they create profit. Transition cost, productivity commitments and client dependencies determine value.
Kumar should focus on the conversion and quality of the seven large deals. Each needs a clear baseline, staged milestones and an accountable executive across sales and delivery. If the contract promises automation savings, Cognizant must know which tools are production-ready and which still require investment. Benefits shared with a client should be measured after security, quality and change-management costs, not against a theoretical reduction in coding time.
Trailing bookings of $29.6 billion, up 11 per cent, provide useful coverage but can conceal concentration or long durations. Management should track the proportion that is net new, the time to revenue and margin progression by cohort. It should also preserve a pipeline of smaller AI and engineering projects that can expand quickly. Depending entirely on large outsourcing deals would expose Cognizant to slower ramp-ups and price competition.
The financial-services segment again led performance, benefiting from the company’s established expertise. Banks and insurers have large modernisation needs and strong incentives to use AI in service, fraud, underwriting and operations. They also impose demanding regulatory and resilience standards. Cognizant can turn governance into premium work if it offers repeatable controls rather than recreating compliance for every client.
AI building begins below the model
The most valuable enterprise work is often unglamorous. Data must be cleaned, applications connected, permissions mapped and processes redesigned before an agent can act reliably. Kumar’s builder framing is strongest when it emphasises these foundations. Cognizant already manages critical systems and understands the organisational constraints that a model provider may not see.
The company should productise that knowledge. Common tools for data readiness, model evaluation, agent monitoring, security and process discovery can reduce delivery time. Industry-specific components for claims, payments, healthcare administration or manufacturing engineering can deepen differentiation. The goal is not to sell a single proprietary model. It is to create a controlled system in which clients can change models without rebuilding the entire operating environment.
Partnership breadth helps maintain choice, but it can create inconsistent architecture. Cognizant works with major cloud, software and platform groups. Sales incentives from those partners should not determine technical selection. The company needs transparent criteria covering accuracy, cost, latency, data rules and exit options. Its value depends on being the client’s accountable architect rather than an extension of a vendor’s channel.
Agentic systems also change liability. An assistant that drafts information is different from an agent that moves money, alters a production schedule or communicates with a patient. Cognizant must define approval thresholds, maintain audit trails and design recovery. It should price the cost of assurance and resist autonomy where controls are immature. Responsible restraint can support trust and longer relationships even if it slows initial deployment.
Engineering breadth must create cross-selling
Cognizant’s acquisition of Belcan expanded aerospace, defence and industrial engineering capability. That gives the group access to physical products and operational technology, areas where AI interacts with safety, hardware and complex regulation. The strategic value depends on integration. Belcan’s specialists should gain access to Cognizant’s global delivery, data and software capabilities, while Cognizant’s clients gain deeper engineering services.
Engineering work can be more defensible than generic application services because it requires domain knowledge and long product lifecycles. It can also carry lower scalability and project risk. Kumar needs to identify reusable methods in simulation, digital twins, embedded systems and verification. Cross-selling should be measured through shared clients and combined wins, not simply consolidated revenue.
There are cultural differences to manage. Engineering organisations often rely on specialised, locally embedded teams, while IT services have been optimised around global delivery. Forcing all work into the same model could damage quality. Cognizant should combine global tools and back-end capacity with domain leadership close to the client. Integration should remove duplicated functions and expand capability without erasing the operating practices that made the acquired expertise valuable.
Physical AI also offers a route beyond the crowded market for office productivity. Manufacturers, aerospace groups and other industrial clients need intelligence in design, maintenance and operations. Errors have real-world consequences, so verification and cyber-physical security become core. Cognizant can create higher-value services if it connects software, data and engineering assurance under one accountable programme.
Margin improvement must come with reinvestment
First-quarter GAAP and adjusted operating margin were both 15.6 per cent. Adjusted margin increased slightly year on year, and the full-year guidance was raised to 16 to 16.2 per cent, representing expected expansion of 20 to 40 basis points. Adjusted earnings per share rose 13.8 per cent. These figures indicate that growth and discipline are coexisting, but Cognizant still operates below the margin level of some India-based peers.
Kumar should not chase parity mechanically. Cognizant’s geography, acquisition mix and onshore engineering exposure differ. The relevant test is whether margin improves through pricing, utilisation, delivery productivity and higher-value work while the company continues to invest. Cutting sales, training or innovation could meet a near-term target and weaken the recovery.
Approximately 357,600 employees support the operating model. AI can increase revenue per employee, but workforce actions need to preserve domain expertise and the apprenticeship pipeline. Cognizant should redesign teams around human-agent workflows, reward reusable assets and train managers to supervise automated work. General AI course completion will not establish readiness for regulated production systems.
The margin opportunity also lies in reducing organisational friction. Large deals often cross industry, service and geography lines. One account leader should be able to assemble capabilities without transferring cost and responsibility through layers of internal negotiation. Kumar’s builder strategy requires the company to behave as one organisation at the client interface, even while specialist groups retain technical authority.
Growth must become more durable than a rebound
Cognizant’s guidance implies stronger full-year momentum than the company delivered in the first quarter on a constant-currency basis. Bookings support confidence, but macroeconomic uncertainty can delay starts and reduce discretionary work. Kumar needs a balanced portfolio of cost-saving managed services and growth-oriented transformation. Cost programmes are resilient but price-sensitive; innovation work can expand faster but is easier for clients to postpone.
The AI velocity gap is a useful description of client frustration, but it must lead to measurable outcomes. Cognizant should report examples through aggregate operating indicators: time from pilot to production, reuse of common platforms, expansion after deployment and the effect of automation on delivery economics. Labelling a growing share of work as AI-related will become less informative as AI becomes embedded everywhere.
Kumar has rebuilt commercial energy and brought acquisitions, partnerships and an industry focus into a clearer narrative. The first-quarter combination of revenue growth, bookings and higher margin guidance suggests that the turnaround is gaining substance. The danger is that momentum encourages complexity, with too many partner solutions and specialised units competing for the same client budget.
Cognizant can differentiate by accepting responsibility for the difficult middle between a model and a business outcome. That requires modernisation, engineering, governance and long-term operation. It is less glamorous than releasing a foundation model and more valuable to most enterprises. If Kumar can make that work repeatable, the company can grow faster without surrendering margin. If each engagement remains bespoke, the builder strategy will describe effort rather than an economic advantage.
In 2026, Cognizant no longer needs to prove that it can return to growth. It needs to prove that the growth is structurally better: supported by net-new work, reusable capability, strong cash economics and deeper client relevance. The first-quarter results provide evidence, not completion. Kumar’s leadership will be judged by whether bookings convert into a more valuable company rather than merely a busier one.
