£365k of annual uplift projected from AI buyer engagement and 24/7 sales response for a 30-person global wholesale marketplace

The situation

The brief

"We're losing buyers to the time-zone gap. Build us an AI engagement layer that responds instantly, recommends the right products, and converts, without re-platforming our website or hiring more sales staff."

What we're building

• Mapped the full buyer path: registration, category selection, outreach, product discovery, conversation, order, and identified the four points where buyers drop or sales reps lose hours.
• Designed a buyer-product matching engine that sits behind the existing site, taking each buyer's declared categories and behaviour and surfacing the right SKUs from a multi-thousand-line catalogue.
• Specified an automated WhatsApp + email outreach engine, deep-linked product offers sent to all 340 registered buyers on a behaviour-driven cadence, no front-end changes required.
• Architected a 24/7 conversational AI sales agent that handles first-response, qualifies intent, captures custom-brand order parameters (MOQs, colours, specs) and hands over to a human only when needed.
• Scoped the integration plan with the in-house dev team, we deliver the AI logic and APIs, they handle the front-end and order workflow. Clean ownership lines, no platform rebuild.

Engagement length: Two phases, ~120 days total · Deliverable: Live recommender engine, automated outreach pipeline, 24/7 conversational sales agent, integrated with existing platform via API.

What we found

• Median first-response time of 12 hours on inbound enquiries, effectively unanswered for the entire buyer's working day. On a global marketplace where competitors reply in minutes, this was the single largest source of lost revenue.
• 340 registered buyers, no behavioural outreach. The platform was sitting on declared category preferences for every buyer and using none of it. Outreach was either ad-hoc or skipped entirely once a buyer had registered.
• 12 sales reps doing manual triage. First-response work, "what do you sell", "do you ship to my country", "what's your MOQ", was eating an estimated 40% of sales-team capacity. That's headcount paid to answer questions an AI can answer in seconds.
• In-house chatbot built but not closing the loop. The team had built their own assistant and integrated it with WhatsApp, a real engineering achievement, but it routed everything to a human and inherited the same 12-hour latency. The infrastructure was there; the intelligence layer was missing.
• No discovery layer on the site itself. Buyers landing on the platform had to navigate the full catalogue, cookware, furniture, apparel, footwear, to find the categories they'd already told the platform they buy. The result: drop-off before the basket.

What we recommended

The projected outcome

• £260k/yr in recovered revenue from a step-change in first-response time, 24/7 AI coverage replacing a 12-hour gap, against a buyer base spread across every working day on the planet.
• £105k/yr of sales-team capacity released, ~5 FTE-equivalent of manual first-response work automated, redeployed onto high-value accounts and outbound.
• 40% projected lift in enquiry-to-quote conversion from instant response and behaviour-driven product matching, paired with deep-linked outreach across 340 registered buyers.
• Phase 1 payback inside seven weeks, build cost recovered before the platform subscription has billed twice.
• Zero front-end rebuild. AI logic delivered as APIs the in-house dev team integrates against. The buyer-facing site stays exactly as it is; the intelligence sits behind it.

Status: Phase 1 in build · Phase 2 scoped and contracted · First buyer cohort live in week 6.