Requirements engineering best practices: From business needs to functional specifications

By Razzaq Onotu

Requirements engineering acts as an unsung hero in producing successful products through its work which transforms stakeholder aims into structured blueprints while resolving complexity. Through my experience heading QuickBucks at Access Bank alongside the Sendbox logistics platforms I discovered business needs transform into functional specifications through methods beyond technical terminology. The path to success requires stakeholders to collaborate through empathetic communication for achieving unambiguous solutions.

The beginning of every product emerges from one primary problem. Fast digital approval processes for SMEs represented the central challenge at Access Bank while legacy approval processes took days. Stakeholders agreed on the goal, accelerate lending, but diverged on execution. The risk team focused on preventing fraud while the growth team demanded speed in all their operations. Bridging this gap required workshops where cross-functional teams mapped user journeys, uncovering pain points like manual document uploads and opaque eligibility criteria. The meetings converted general targets into practical specifications. The requirement to speed up approval processes evolved into the functional requirement which enabled business registration applicants to complete the digital submission process through automated government database connectivity resulting in manual entry reduction. Non-functional requirements appeared alongside the primary system requirements when the developer stated that processing should handle 500 simultaneous applications within a 2-second latency. The detailed approach helped teams avoid unnecessary expansion of work while ensuring all team members are aligned on specific achievement targets.

Requirements engineering functions through the process of asking the right questions. The stakeholders requested real-time transaction tracking when Sendbox integrated with its blockchain remittance partner at first. Probing deeper revealed their true need: merchants required transparency to resolve disputes, not just speed. We moved our project toward designing a permanent audit log system which delivered SMS notifications to users with limited bandwidth capabilities leading to better dispute resolution times. The integration tools of user story mapping together with MoSCoW (Must-have, Should-have, Could-have, Won’t-have) prioritization proved to be required essentials. By categorizing requirements based on business impact, we avoided over-engineering. This approach fostered alignment, as marketers and engineers literally saw the same vision.

The fundamental value of Product Requirement Documents (PRDs) depends on how easily stakeholders can access them during development. At Access Bank, early PRDs for QuickBucks were dense technical tomes, alienating non-technical stakeholders. We transformed them into visual flow that merged flowcharts and mockups while adding acceptance criteria. A visual Figma prototype of an SMS OTP integration system accompanied the PRD acceptance criteria which stated “Users need to complete authentication within 45 seconds.” This approach fostered alignment, as marketers and engineers literally saw the same vision. We established usability metrics as performance indicators to measure user success rates where “90% of users finished loan applications without creating support tickets” and “interface ratings from users exceeded a minimum of 4/5 for ease of use achieved 85% agreement.” Reliable KPIs directed the design process toward making decisions such as switching from dropdown menus to predictive text to reduce form abandonment.

Non-functional requirements can either make a product succeed in the marketplace or let it fail. The requirement at Sendbox for “scalability” needed more specific definition. We measured and established the requirements when testing platform load under peak holiday conditions to support 10,000 concurrent users with 3-second maximum latency. Our implementation of cloud-native architecture enabled seamless scaling thanks to which Black Friday traffic increased by three times without any disruption.

Change will always occur but uncontrollable chaos does not need to exist. The Agile principles taught teams to accept adjusted project requirements while preserving control of development processes. For QuickBucks, regulatory shifts mid-project necessitated adding facial recognition for KYC. Our team conducted impact analysis to understand the trade-offs before diminishing the rewards dashboard schedule by two sprints for compliance reasons. Open sharing of information between stakeholders transformed probable conflicts into productive working relations.

Artificial intelligence technology continues to transform requirements engineering practices. The combination of stakeholder interview analysis powered by NLP technology and generative AI solutions accelerates how feedback processes work. Technology lacks the capacity to adequately substitute human interaction because humans provide essential attributes such as active listening and realistic assumption evaluation together with the skill to blend dreams with achievable solutions.

In the end, requirements engineering isn’t about documentation. It’s about crafting a shared language that turns diverse perspectives into coherent action. Whether building fintech platforms or logistics ecosystems, success hinges on transforming “I need” into “We will,” one precise requirement at a time.

READ MORE FROM: NIGERIAN TRIBUNE