Case Study — AI Workflow & Product Strategy at Tovuti LMS
An AI-Native Design Practice at Tovuti
Most designers who say they "use AI" mean they've added a tool to their workflow — faster, but the same shape of work. I had a different question: what if AI wasn't a tool in the workflow, but the infrastructure the workflow ran on? Three interlocking builds over one year answered it.
The problem#
Three problems were sitting in front of me simultaneously, and none of them were design problems a new Figma component would solve. Research synthesis took 10–15 hours per feature cycle — pulling from Zendesk, transcripts, KB articles, and meeting notes into a coherent brief. Journey maps were static artifacts disconnected from the evidence that should have built them. And Tovuti's analytics product had no strategic vision — no defined north star for what data should surface, to whom, and why.
The action#
I built three interlocking things over the course of the year, each scoped to eliminate a specific structural bottleneck rather than just speed it up.
JEM (Journey Experience Mapper)
A 0→1 web product that ingests multi-source research and generates living, evidence-backed journey maps through guided AI conversation. Full product ownership: architecture, data model, scoping interface, map canvas, engineering collaboration. Shipped to jem-test.tovuti.ai.
10 custom Claude skills
A library of purpose-built AI skills covering research synthesis, competitive analysis, component specification review, and design handoff — each scoped to a specific slow point in the design cycle and built to eliminate it, not just speed it up.
Analytics vision
A product strategy defining what Tovuti's analytics platform should become: an AI-assisted intelligence layer surfacing friction by feature, persona, and stage — audience-by-surface coverage maps, journey maps for analytics users, and a strategic brief positioning analytics as a product capability, not a reporting tab.
The bar for each skill was: does this eliminate the translation step, or does it just make it faster? General-purpose AI tools were rejected in favor of skills scoped tightly enough to require minimal editing of their output — narrower scope, but output a designer could actually trust.
Results#
10–15 hrs → <2 hrs
research synthesis time per cycle
Across the 10 custom Claude skills
20
friction points surfaced in JEM's first week
10 shipped as quick wins immediately
10
custom Claude skills shipped
In active use as part of the standard design workflow
More than the individual outputs: a research-to-artifact pipeline that didn't exist at the start of the year, journey maps built from evidence instead of memory, and a product strategy for analytics that connected data to decisions rather than adding another report. The demonstration isn't that AI tools were used — it's that AI-native design practice means building the tools you need, not just using the ones that exist.
Lessons#
The leverage point is structural, not tool-specific
Synthesis isn't slow because it's hard — it's slow because information comes in the wrong shape. Building a skill that reshapes it, rather than a faster way to do the reshaping by hand, is where the time actually comes back.
AI infrastructure should eliminate steps, not replace judgment
The skills that stuck were the ones that removed a mechanical translation step and handed a designer a structured draft to validate — not the ones that tried to make the decision for them.
A product strategy is still design work, even with no shipped UI
The analytics vision produced no interface, but defining whose decision every report should serve — audience-first rather than data-first — is the same design thinking that shaped JEM's data model.
FAQ#
Is this the same case study as JEM?
JEM is one of the three builds this piece covers, with its own full case study elsewhere on this site. This piece is the broader story: JEM plus the 10 custom Claude skills plus the analytics vision, told as one narrative about building AI infrastructure rather than adopting AI tools one at a time.
Did the analytics vision ship as a product?
No — it's a strategic brief with a defined north star and an AI-integration roadmap, not a shipped interface. This case study represents the vision and the method, not post-launch metrics, and doesn't claim otherwise.
Want the details?
Happy to walk through any of the three pieces — JEM's architecture, how the Claude skills were scoped, or the analytics strategy brief.