RiverPrompts: What Happens When AI Becomes Your Analyst?

Article written by Rohitha Bollipalli , Marketing Manager at Riverflex.

Introduction 

Last month at our internal QBR, we ran an experiment called RiverPrompts, a Hackathon built for the age of next-gen consulting. We asked a simple but high-stakes question: what happens when sharp human minds collaborate with an AI analyst to crack a real business problem? 

Consultants from Riverflex (R1 consultants) were given a fictional brief from “Transvora Logistics,” a global freight management company decommissioning 3,000 m² of warehouse space in key regions. The challenge? Repurpose the space to drive a 35% uplift in client engagement and service uptake within a year across four diverse markets: Mexico, Indonesia, Poland, and the UAE. 

Armed with nothing but their brains, some seriously smart prompting, and access to GPT-powered deep research, R1s had 90 minutes to respond. The results were bold, practical and visionary. 

We’re now opening the doors to share the winning submissions, the strategy behind each one, and crucially the prompts used to generate them. We hope that others can learn from how we work with AI, not against it. 

Problem Statement 

Transvora Logistics is decommissioning 3,000 m² in key regional hubs. The space must be repurposed to generate measurable business value. The core objective: Increase client engagement and service usage by 35% in the next 12 months in Mexico, Indonesia, Poland, and UAE. 

The space must: 

• Reflect regional needs and challenges 

• Be aligned with the brand promise: Efficient. Transparent. Connected. 

• Be cost-effective, scalable, and impactful 

Results 

Alex Astley : Regional Repurposing with Strategic Depth 

Alex tackled the challenge like a true strategist. He built a 5-year roadmap for each region, balancing near-term wins with long-term vision. His insights: 

• Mexico: Convert the space into a Cross-Border Consolidation Hub—with on-site customs, secure transloading, and staging for nearshored factories. He tied this to the U.S.-Mexico trade surge and driver shortages, predicting a 25%+ increase in throughput with lower client risk. 

• Indonesia: Position the space as a Bonded Logistics Centre + 4PL Control Tower—reducing client duty costs while offering centralized distribution coordination. He forecast improved cash flow for importers and higher SME engagement through shared transport planning. 

• Poland: Develop a Value-Added Assembly Centre for nearshored manufacturers—offering kitting, quality checks, and region-specific packaging. This would shorten time-to-market for clients shipping across Europe. 

• UAE: Build an Innovation & Client Experience Lab, showcasing automation tech and offering immersive co-design sessions. He projected that this would not only attract new clients but deepen relationships with existing ones by 30–40%. 

Prompting Style: Alex used a “consultant-as-researcher” approach—treating GPT not like an idea machine, but like a junior analyst. His base prompt was precise and multidimensional: 

“As a strategic logistics consultant for Transvora Logistics, develop a region-specific 5-year plan to repurpose 3,000 m² of warehouse space in Mexico, Indonesia, Poland, and UAE. Your goal is to increase client engagement and service utilization by 35% within 12 months, while ensuring long-term commercial value and alignment with Transvora’s brand promise.” 

Then, he chained prompts to layer depth and context. For every answer he got, he followed up with: 

• “Now list the dependencies for each strategy by region.” 

• “What are the risks, and how would you mitigate them?” 

• “Suggest real-world examples or case studies that support the feasibility of each regional plan.” 

This approach unlocked more than just surface-level answers—it created a decision-ready strategy. By treating prompts like interview questions in a discovery session, Alex pushed the AI to think structurally, not just creatively. 

💡 Takeaways: 

• Don’t stop at the first answer—use follow-ups to layer reasoning. 

• Ask for risks, dependencies, and real-world examples to ground abstract ideas. 

• Think like a consultant in discovery mode—your prompts should sound like strategy questions. 

• The magic isn’t in the prompt—it’s in the sequence. 

Jessica Straeuli (Binns) : Smart Systems, Smart Relationships 

Jessica’s lens was razor-focused: How can technology improve trust and visibility across the supply chain? Her proposals brought the “Transparent” and “Connected” pillars of Transvora’s brand to life: 

• Mexico: Deploy an IoT-enabled Customs Hub—with smart tracking of cross-border shipments, sensor-triggered compliance alerts, and real-time ETA dashboards for clients. It would give clients confidence in notoriously high-risk trade corridors. 

• Indonesia: Set up a Modular E-comm Fulfilment Centre—specifically designed for fast-growing D2C brands. Using automation-light tools and plug-and-play integrations with local marketplaces, Jessica aimed to reduce order-to-ship time by 40%. 

• Poland: Launch a Smart Warehouse Pod—a micro-distribution node connected to a real-time WMS dashboard. She tied this to seasonal surge needs, predicting reduced SLA breaches and 20% faster last-mile delivery. 

• UAE: Build a Client Learning & Innovation Centre—where logistics teams could test systems before onboarding, reducing friction and increasing retention. She pitched this as a pre-sales and post-sales powerhouse, reinforcing Transvora’s long-term stickiness. 

Prompting Style:  Jessica worked backwards from the brand. Rather than dive straight into ideas, she grounded her approach in Transvora’s promise: Efficient. Transparent. Connected. Her base prompt read more like a creative brief than a command: 

“What AI-powered initiatives could Transvora launch to repurpose 3,000 m² of logistics space across Mexico, Indonesia, Poland, and UAE? Focus on cost-effective, scalable solutions that improve operational visibility, customer experience, and trust, aligned with ‘Efficient. Transparent. Connected.’” 

What made her style shine was her brand-first prompt refinement. She didn’t just iterate—she zoomed in, aligning each follow-up to a brand value: 

• “Give me solutions that emphasize operational transparency in cross-border logistics.” 

• “How can a logistics provider build client loyalty through visibility tools?” 

• “Suggest warehouse automation options that small businesses can afford and adopt.” 

By narrowing the aperture this way, she created output that felt tailored, consistent, and trustworthy—like Transvora itself had written it. 

💡 Takeaways: 

• Start from the brand lens, not just the business ask. 

• Use follow-ups to tune the tone and values, not just the features. 

• Prompt for client experience, not just operational features—especially in B2B. 

• Think of GPT like a brand strategist: what would this company say, build, or do? 

Maite Zanasi : Modular Value & Innovation Ecosystems 

Maite pushed the brief into bold territory—reimagining warehouse space as platforms for partnership, experimentation, and ecosystem growth. Her results felt more like business model innovations than facility redesigns: 

• Mexico: Transform the space into a Client Co-Creation Lab—a hands-on space for clients to prototype fulfilment flows, test packaging, and simulate cross-border delays. She framed it as part logistics, part relationship design. 

• Indonesia: Launch a Gig Fulfilment Zone—an AI-routed micro-hub for crowdsourced drivers and pick-packers. This would allow SMEs and emerging DTC brands to access enterprise-level fulfilment with zero fixed cost, radically levelling the playing field. 

• Poland: Build a Shared Manufacturing & Postponement Centre—supporting light assembly, localization, and compliance tagging for EU-bound goods. Clients could hold generic inventory and customize just-in-time, saving working capital and storage. 

• UAE: Create a Sustainability-First Innovation Hub—solar-powered, LEED-certified, and equipped with pilot zones for testing green logistics (EV fleets, blockchain tracking, drone last-mile). She tied this directly to ESG procurement trends in the GCC. 

Prompting Style:  Maite didn’t just prompt for answers—she prompted for better thinking. Her base prompt was visionary: 

“You are advising a global logistics firm that is freeing up 3,000 m² per site in four diverse markets. Design space-as-a-service models that offer strategic value for clients and position the company as a co-creator in logistics innovation.” 

But she didn’t stop there. She prompted the AI to improve its own process. That’s what made her approach stand out: 

• “What types of space-as-a-service models have been used successfully in logistics or retail?” 

• “Generate five follow-up prompts to help me stress-test my idea for a co-creation zone.” 

• “What assumptions are we making about client behaviour in emerging markets?” 

This wasn’t just recursive—it was reflective. She made GPT challenge its own output and reframe the question. The result? Proposals that were modular, future-focused, and rooted in system design thinking. 

💡 Takeaways: 

• Don’t just prompt for answers—prompt for better questions. 

• Use GPT to stress-test your thinking, not just generate content. 

• Challenge assumptions. Ask, “What are we missing?” 

• Think of AI like a strategist in a whiteboard session—it should help you explore, expand, and reframe. 

Conclusion 

Each winning consultant used AI differently, and that’s what made the results shine: 

• Alex treated GPT like a junior analyst, layering prompts to build depth, structure, and risk-mitigated plans. He didn’t just ask for solutions—he built them, step by step, in conversation with the model. 

• Jessica reverse-engineered from the brand promise. Her prompts weren’t just about logistics—they were about trust, visibility, and client experience. She used AI like a systems thinker would use a whiteboard. 

• Maite used meta-prompts to stretch what the AI could do—asking not just “what’s a good idea?” but “how do we get to a better one?” Her approach was generative, modular, and deeply human in its creativity. 

Across the board, what worked wasn’t a one-size-fits-all formula. It was intentionality. Craft. Curiosity. These weren’t generic prompts—they were precise, shaped by sharp minds who knew what they were after. That’s the lesson: AI is only as good as the questions you ask—and the way you ask them. 

At RiverPrompts, we learned how to ask better questions. And that, more than anything, is the skill of the future.