We're a young firm — so instead of parading NDAs we can't share, here's the published evidence behind the AI systems we build. Every metric on this page links to a primary source: a peer-reviewed study, a company press release, or a documented client case study. Click through and verify.
One month after going live globally, Klarna's OpenAI-powered assistant took on two-thirds of customer service chats, matched human CSAT, and cut resolution time from 11 minutes to under 2.
Brynjolfsson, Li & Raymond studied a staggered rollout of a generative-AI assistant to real customer-support agents. Issues resolved per hour rose 15% on average — with the largest gains for less-experienced staff.
A randomized field experiment with 758 BCG consultants — run with researchers from Harvard, Wharton, MIT and Warwick — measured what happens when knowledge workers get access to frontier AI on real tasks.
McKinsey's 2024 review of production gen-AI programs in service operations documents double-digit reductions in average handle time and meaningful lifts in first-contact resolution among disciplined adopters.
A ten-person specialist retailer replaced advice-intensive chat overflow with an AI product consultant. The lift was measured causally in an A/B test averaged over six months, with a 33× ROI.
A Cardiff coatings SME rebuilt its operations around an AI-assisted ERP, resolving 30,000 data-duplication points and returning five hours a week per team member. Documented by smedigital.ai.
Note · The engagements above are published third-party results, not AI Optimal Solution client work. We reference them because they represent the same categories of AI system we design and deploy — conversational agents, support copilots, knowledge-worker assistants, and operational automations. Once our own client work is cleared to share, it will replace these entries.
