As an early-stage asset finance SaaS business on a rapid growth trajectory, it’s crucial that we, as Zeti’s Product team, prioritise the features that we add to ZetiOS and avoid sinking development resources into unnecessary increments. This can be challenging when multiple opportunities pull us in different directions, particularly when our cloud-based asset finance platform already serves distinct user groups such as lenders and asset operators.
We must not only prioritise the most valuable work, but also ensure our justification and processes are robust enough for commercial and operational stakeholders to understand and respect, especially when saying ‘no’ to a feature that would help progress a prospect or a customer in the short-term.
Product discovery and prioritisation aren’t new concepts in enterprise software development. Evaluating ideas by value and complexity has been around for decades within various methodologies. What has dramatically evolved, however, is the toolkit to conduct that work.
Generative AI in product management and modern prototyping tools, some of which were featured in our CTO’s outlook on AI in the last edition of Zeti Pulse, have radically reduced the time that it takes to turn a hypothesis into something customers can interact with. Through that interaction, we can better predict customer outcomes, both positive and just as importantly, negative - and make clearer decisions about whether a feature should be developed at all.
AI-powered product discovery: From notes to knowledge
One practical change has been how we capture and consolidate discovery activity. AI meeting notetakers now ingest discovery call transcripts both prior and during our Product team’s involvement, storing them in a coherent, organised fashion. Using tools like Claude Cowork, these transcripts are combined with email threads and commercial documents stored on SharePoint or Google Drive.
The output is not just a record of conversations, but an evolving synthesis of pain points, decision criteria, edge cases and stakeholder perspectives. This consolidated view removes the need to rely on fragmented memories and isolated conversations. Instead, it allows prioritisation from evidence that is structured and auditable.
For enterprise software and fintech product teams, this kind of AI-powered product discovery creates a far more structured foundation for prioritisation.
Vibecoding:AI- prototyping supercharged
Another key change relates to the increased fidelity and speed of AI-assisted software prototyping enabled by generative AI. Collins Dictionary named ‘vibecoding’ as Word of the Year for 2025, signalling the significance of this AI-assisted design and coding breakthrough.
Prototyping itself is not new. During the 2010s, designers commonly used tools such as Figma, Sketch or InVision to bring concepts to life, relying on years of training and experience to translate ideas into usable design.
Vibecoding has now, to an extent, democratised this activity. At Zeti, our Product team is now routinely and rapidly building interactive, production-like prototypes for lenders and asset operators using AI-enabled tools like Lovable and Claude Code. Crucially, even without a designer’s direct involvement, the detail and quality that can be achieved is so high that prospects and customers - junior or C-suite - can seriously assess what feels like a fully-fledged solution. They can identify what delivers value, deprioritise what does not and in some cases, demonstrate a level of commercial commitment that alone can qualify a prioritisation decision.
In light of this feedback, prototypes can be quickly iterated and finalised, or discarded altogether. This enables development resources to focus on building the right features and, just as importantly, to avoid spending time on the wrong ones.
One recent example involved our work on a digital lending early settlement journey with a lender customer. The existing process for partial or full settlement of hire purchase or finance lease agreements was handled manually across multiple teams - was rife with inefficiency and business rules. Discovery insights, captured consistently from early commercial discussions through to detailed operational calls, allowed us to vibecode both lender- and borrower-facing journeys. By presenting realistic, interactive flows, the client was able to see that the hire purchase journey delivered higher value with lower complexity, and chose to prioritise that path while deprioritising the finance lease journey initially.
Particularly in early-stage SaaS and fintech software development, the challenge is rarely a lack of opportunities, but the discipline to decide which ones deserve attention. While the underlying theory remains familiar, the rapid rise of AI-enabled tools has fundamentally transformed how confidently these decisions can be made. When discovery, prototyping and feedback are this tightly connected, saying no becomes less about constraint and more about focus. And in a fast-moving business, that focus is what allows the right ideas to move forward.
FAQs
1. What is product discovery in SaaS software development?
Product discovery is the process of understanding customer needs, validating opportunities, and assessing value before investing development resources. It helps teams prioritise the features most likely to deliver business and customer outcomes.
2. How is AI changing product management?
AI helps product teams consolidate customer feedback, analyse discovery conversations, generate insights, and rapidly prototype solutions. This allows teams to validate ideas faster and make more informed prioritisation decisions.
3. What is vibecoding?
Vibecoding refers to using AI-powered tools to rapidly create software prototypes and working applications from prompts and requirements. It enables product teams to test ideas quickly before committing development resources.