Persistent’s SDLC persona-focus reveals an affordable path to GenAI adoption across the enterprise

Persistent Systems is placing the software development lifecycle (SDLC) at the forefront of generative AI (GenAI) adoption, bridging enterprise challenges through a persona-driven strategy. By targeting key decision-makers such as CTOs, CIOs, and chief product officers (CPO), Persistent aligns its offerings with the tangible pain points of these personas. The result? Tailored GenAI solutions that cut through the noise of big tech’s generic capabilities to deliver actionable, cost-effective outcomes.

Persistent’s SASVA platform—a Service as Software Virtual Agent—epitomizes this approach. Designed to tackle the complexities of SDLC, it delivers value through bespoke applications, addressing both technical debt and cost-efficiency while delivering context-rich insights for decision-making. It is cutting delivery times by up to 40 percent.

While humans remain in the loop, the SASVA platform is evidence of the shift toward service-as-software that HFS predicts for all service providers
(see Exhibit 1).

Exhibit 1: Service-as-software becomes increasingly inevitable as firms seek to deliver bigger and better outcomes without adding to headcount

Source: HFS Research, 2024

SDLC is just the bridgehead to expand GenAI into a broader set of functions across the enterprise

Persistent has been targeting product-based enterprises, crafting solutions that resonated with product managers and CPOs. The company has built a strong foundation of use cases, starting with SDLC, which spans assessment, backlog grooming, roadmap generation, release planning, development, quality assurance, deployment, support, and professional services.

Support across this phase-specific approach enables enterprises to accelerate time-to-market while managing legacy constraints. Persistent has deployed SASVA in more than 30 production use cases already.

Clients often approach Persistent with legacy codebases spanning decades. SASVA uses GenAI to parse repositories with thousands of interconnected files, creating actionable roadmaps for modernization. This includes applying a proprietary deterministic engine with tailored small language models to assess and reverse engineer the repositories with thousands of interconnected files, libraries, and dependencies, creating actionable roadmaps for themes such as modernization aligned with customer requirements.

SASVA cracks a multimillion-dollar backlog, saving time and reducing reliance on experts; CMOs and sales leaders are the next targets

One customer engagement involved solving a multimillion-dollar backlog by deploying SASVA. The outcome? Significant cost savings and reduced dependency on large and expensive teams of experts, eliminating common barriers to entry for many projects.

Persistent has now set its sights on adjacent personas, such as chief marketing officers (CMO) and sales leaders. By extending SASVA’s capabilities to enable sales-to-product alignment or optimize go-to-market strategies, Persistent can drive enterprise-wide transformation. Persistent is focused on enterprise issues—cutting the cost of entry and doing away with the need for token-hungry huge context windows.

Focus on affordability delivers effective small models at small(er) cost

While hyperscalers offer compute-heavy solutions, Persistent has built SASVA to operate efficiently on-premises, leveraging smaller models when most appropriate. This focus on affordability has opened the door for enterprises to experiment with GenAI on smaller projects—a point of difference versus the hefty commitments typically required by larger tech providers.

Persistent cut the need for large context windows in GenAI (and therefore reduced the consumption of tokens) with a hybrid approach that combines iterative learning with hierarchical data structuring. This strategy breaks down complex repositories into manageable pieces, enabling targeted interactions between smaller model components and data.

Firms can deliver GenAI solutions on standard CPUs rather than more costly and energy-hungry GPUs

By avoiding reliance on extensive context windows and computationally heavy large language models, enterprises can operate SASVA on standard CPUs rather than expensive parallel processing and energy-hungry GPUs. This significantly reduces cost and compute requirements, making SASVA more accessible and scalable for enterprises handling even intricate and large-scale datasets.

For a leading enterprise grappling with high compute costs, Persistent deployed a tailored solution that reduced processing times while enabling insights from decades of historical data running on a standard CPU infrastructure.

For another global firm managing repositories with tens of thousands of interrelated files, Persistent deployed SASVA’s iterative learning model. The platform dynamically analyzed hierarchical relationships within the data, enabling the customer to modernize their stack while preserving institutional knowledge in the contextual information it captured.

Low-token approach is practical in targeted use cases—but may not meet all requirements of firms with broader, more complex AI needs

Persistent’s low-token approach can make great sense for enterprises prioritizing cost-efficiency and compliance in targeted use cases such as SDLC.

And, when facing broader or more complex AI needs—such as those requiring holistic, multimodal processing or future-proof adaptability—Persistent can deliver by extending context length.

Balancing cost against functionality and scalability will be the key decision point for enterprises evaluating which route to take.

Persona-based strategy aligns with the careful approach of many enterprise leaders in tackling GenAI

Persistent’s persona-based strategy aligns well with the cautious optimism of enterprises exploring GenAI. Persistent addresses the practical barriers of adoption head-on by focusing on affordability, context precision, security, and replicable success models. Its solutions speak directly to concerns such as: How can I modernize without blowing the budget? How do I align my team with actionable insights rather than theoretical models?

By focusing first on SDLC personas and expanding outward, Persistent is grounding its GenAI journey in practical use cases. Its ability to combine technical innovation with business-specific relevance sets it apart amid a market clamor dominated by the hyperscalers’ catch-all approaches.

The Bottom Line: Focus on real-world business problems to navigate the complexity of GenAI adoption.

GenAI can solve real business problems when it’s rooted in enterprise-specific needs and delivered with a persona-driven approach. For enterprise leaders, Persistent’s journey offers critical lessons in deploying GenAI within SDLC frameworks and beyond.

Focus on outcomes, not just technology. A persona-led, cost-conscious strategy offers a compelling roadmap for enterprises navigating the complexities of GenAI adoption. By solving real-world challenges today, enterprise leaders can lay the groundwork for the Generative Enterprise of tomorrow.