Desky started because the founders kept finding standing desks that failed at the component level, usually cheap motors, thin leg columns, weak crossbars, or controllers that died within a year, and they decided to build a company that controlled the supply chain instead of just branding someone else’s product. I joined to help formalize that approach, which means we qualify factories directly, audit motor and frame production, specify our own tolerances, and hold inventory in regional warehouses so we can actually honor warranties and ship replacement parts when something goes wrong.

What makes us genuinely different is that we do not dropship from a mystery factory and hope for the best. We own the sourcing, quality inspection, freight, customs, and local fulfillment process, so we can guarantee stability, lifting performance, and long term reliability because we know exactly which components are in every desk and where they came from. That supply chain control is the real product difference, even if customers mostly notice it as a desk that does not wobble and a warranty that actually gets honored.

Most of our new business today comes from direct online orders that we fulfill from our regional warehouses, driven by search, reviews, repeat customers, and small teams who buy a few desks at a time. That demand is real but it is also lumpy, which makes forecasting difficult because a viral review or a competitor stockout can spike orders overnight and leave us managing motor and desktop inventory week by week.

I wish more of our new business came from enterprise contracts with scheduled deliveries, because predictable volume lets us plan container loads, negotiate better component pricing, and hold the right safety stock instead of reacting to chaos. If facilities managers and procurement teams committed to quarterly rollouts, we could smooth production, reduce expedited freight, and pass those savings into better pricing and faster lead times for everyone.

One thing I have actually handed over to AI in the last year is demand forecasting and reorder point suggestions. The system pulls sales history, seasonality, promotion calendars, supplier lead times, and freight transit data, then recommends when to reorder motors, controllers, desktops, and accessories so we do not run out of core SKUs. That has reduced manual spreadsheet work and helped us catch a few stockout risks earlier than we would have on our own.

One thing I tried on AI and went back to doing the human way is supplier quality issue resolution and factory negotiations. AI can summarize defect rates and draft emails, but it cannot read a factory manager’s hesitation, understand a cultural refusal to admit a production problem, or know when a delay is real versus a negotiating tactic. Those conversations require trust, history, and judgment, so I kept the human relationship in charge and use AI only for preparation and documentation.

Customers are arriving much more informed about supply chain details than they used to be. They now ask about stock availability, lead times, shipping damage rates, replacement part policies, assembly requirements, and whether a desk ships from a local warehouse or from overseas. They have often checked reviews and compared warranties before they contact us, so the conversation is less about basic features and more about whether we can actually deliver the right desk on time and support it for years.

Yes, we are seeing a small but measurable number of people arrive through AI assistants like ChatGPT. They will mention that an AI told them Desky has strong warranties or fast shipping, or they will ask why one desk is backordered when another is in stock. It is not yet a major channel, but it is growing, and it means our inventory data, lead times, and product specs have to be accurate and structured so AI tools do not send customers to us with wrong expectations.

I would tell them to build supply chain resilience as their real competitive advantage. That means dual sourcing critical components like motors and controllers, holding safety stock in regional warehouses, integrating inventory data with sales channels so customers and AI assistants see accurate availability, and auditing factories before a quality crisis hits. The brands that win will be the ones that can actually deliver a stable desk on time, not just talk about ergonomics on a website.

I would tell them to ignore chasing every AI gimmick for supplier relationships or pretending a chatbot can replace human negotiation and quality control. Use AI for forecasting, route optimization, documentation, and inventory alerts, but keep humans in charge of factory audits, defect disputes, and commercial commitments. The supply chain is still a people business, and no algorithm changes that.