A UK company needed to generate qualified pipeline through outbound but struggled to reach the right accounts at scale. Generic outreach produced weak response rates and few booked conversations. They needed a highly targeted, enriched outbound and account-based motion. Rafiki built a sector-specific AI rankings layer on OpenAI's API to score and prioritise target accounts. This was combined with Apollo's MCP and the Companies House API to enrich and precisely target prospects. The result was a data-driven outbound and ABM strategy tailored to the client's ideal sectors. The outbound motion achieved a 9% response rate and a 3% call-booking rate. The AI-ranked, enriched targeting drove meaningfully higher engagement than typical cold outbound. The client gained a repeatable, targeted GTM engine combining AI scoring with live company data.

Project Details

Vault22 is one of South Africa's leading budgeting, financial-wellness and investment platforms, now expanding into the UAE. Entering a new market required sharper positioning and a clear go-to-market structure. They needed ongoing GTM strategy and execution support. Rafiki refined Vault22's positioning to fit its growth ambitions and new market. The team built out a GTM architecture and executes against it on an ongoing basis. This provided Vault22 with both strategic direction and hands-on delivery. Vault22 gained refined positioning and a structured GTM architecture supporting its UAE expansion. Rafiki continues to build and execute the go-to-market on an ongoing basis. The partnership gives Vault22 sustained GTM capacity as it scales internationally.

Project Details

Tessa, a building and construction software company, needed its platform stabilised and reliably deployed in the cloud. The software was not yet ready to demonstrate to investors and prospective customers. They needed engineering support to get it demo-ready and to sustain ongoing development. Rafiki came in to fix Tessa's cloud deployment and prepare the software for investor and customer demos. The team successfully deployed the platform and established a continuous delivery workflow. Using best-in-class AI-enabled engineering, they built a foundation for ongoing iteration. With the software successfully deployed, Tessa could confidently run investor and customer demos. Rafiki now augments and iterates on the platform month to month. The engagement gave Tessa reliable, continuously delivered engineering capacity.

Project Details

Safe Space Talent, a UK government-approved training provider specialising in apprenticeships, needed to scale networking and candidate-to-opportunity matching across a growing network. Manual coordination limited how effectively they could connect learners, employers, and opportunities. They wanted intelligent, conversational infrastructure that met their users where they already communicated. Rafiki built a custom MCP server tailored to Safe Space Talent's data and workflows. The team then deployed AI agents directly into WhatsApp, creating an agentic-led, human-in-the-loop networking layer. The system powered intelligent matching across the provider's network, surfacing relevant connections and opportunities automatically. The deployment gave Safe Space Talent a chat-native, agent-driven matching infrastructure operating inside WhatsApp. Networking and matching that previously required manual effort became automated and conversational. The provider gained scalable infrastructure to connect its apprenticeship network through a familiar channel.

Project Details

Sixley aimed to enhance its recruitment processes, specifically improving ATS integrations, data quality, and candidate processing speed. They sought agentic, human-on-the-loop hiring workflows to scale their network-led, human-centric approach. The goal was to outperform traditional job boards and streamline their talent acquisition. Their network-led model, functioning as a headhunter at scale, needed further optimization to maintain its competitive edge. The vendor built out and optimized Sixley's ATS integrations and recruitment automation. This included implementing solutions for better data quality and faster candidate processing. They also developed agentic, human-on-the-loop hiring workflows to support Sixley's unique headhunter-at-scale model. These improvements were designed to seamlessly integrate with their existing systems and enhance operational efficiency. As a result of these improvements, Sixley achieved a significant offer rate of 20-35%, which surpassed the performance of major job boards. The enhanced recruitment processes led to more efficient candidate handling and higher quality hires. Additionally, Sixley was awarded Silver at the In-house Recruitment Live Innovation Showcase in Manchester, acknowledging their innovative approach.

Project Details

Modern work had shifted to project-based and agent-driven models, relying on diverse teams of contractors, specialists, and AI agents. However, existing payment infrastructure was designed for traditional salaried employees or single freelancers, making multi-party invoicing and cross-border settlement slow and manual. This created significant challenges for splitting client payments efficiently and maintaining compliance across distributed project teams. The growing involvement of AI agents in billable work further exacerbated this payment gap. A chat-native project finance platform, Petl Pay, was developed to address these issues. The platform integrated with communication tools like WhatsApp, Slack, Claude, and ChatGPT to log work and automate invoicing. It was designed to consolidate multiple contributor payouts from a single client invoice, ensuring compliant and efficient cross-border settlements. This infrastructure utilized regulated open banking and stablecoin rails, with built-in backend orchestration and a compliance layer for automated, optimized payments. The platform successfully validated the market need by supporting over 55 initial projects through Rafiki Works. During its first three months in private beta, Petl Pay facilitated transactions for over 30 international projects monthly. Users experienced typical settlement times of under 2 minutes for both bank-to-bank and cross-border payments. The system also offered near-instant cross-border payouts in multiple currencies, including USD, EUR, and ZAR, with regulated stablecoin rails providing approximately 60-second settlement times at around 0.5% fees.

Skills

Project Details