Viston helped us deploy personalization, dynamic pricing, and content generation to lift conversions during peak sales while keeping infra costs predictable.

Business challenge

Conversion stuck at 1.6%, AOV at ₹1,980, and rising CAC with high cart abandonment on mobile.

Our approach and solution

Viston built a Recommendations API (session-based + user cold-start), dynamic pricing guardrailed by margin floors, and automated product copy/SEO generation with brand-safe prompts.

A multilingual chat assistant deflected routine returns/sizing questions and routed exceptions to human agents.

AI applications and benefits delivered

Recommenders (sequence models + embeddings), pricing elasticity estimation, LLM content generation with toxicity/brand checks, and conversational support.

Benefits: higher AOV/conversion, lower CAC, faster content ops.

Cost of implementation

USD 40,000 including Vertex AI usage, feature store, and 6 months of observability.

Time to implement

10 weeks from design to full rollout.

Tools and technologies used

Google Cloud Vertex AI, BigQuery + Feature Store, Recommendations AI, Cloud Run, Redis, LangChain, Cloud Armor, Firebase A/B testing, Datadog.

Quantitative outcomes

AOV increased 18% to ₹2,336.

Conversion rate up 42% to 2.3%.

CAC reduced 21%.

Content throughput 5x with editorial approval under 24 hours.

Revenue per session +24%.

Key performance indicators (KPIs) tracked

CTR on recs, revenue per session, NDCG@10, cart abandonment rate, CAC, CSAT, LLM toxicity rate.

Pre- and post-implementation metrics

Conversion: 1.6% → 2.3%.

Abandonment: 71% → 58%.

Time-to-publish PDPs: 3.5 days → 16 hours.

A/B uplift in RPS: +24% (p<0.05).

Stakeholder quotes or testimonials

“The personalization engine felt ‘native’ to our brand. Editorial control never took a back seat.” — Chief Growth Officer

Regulatory or compliance considerations

India DPDP Act 2023, PCI DSS scope isolation, content safety filters, opt-out and data minimization controls.