Your pipeline has changed shape in eighteen months. The enquiries are now for GPU capacity and higher rack densities, for interconnection into AI workloads from customers who never asked about interconnection before, and from enterprises asking their managed service provider who will run their AI estate. The sales, provisioning and support processes handling them were built before AI was the reason anyone called you.
AI has changed what your customers buy. Whether it has changed how you sell, provision and support it is the question that decides who earns from the demand.
You are selling AI-era capacity through pre-AI systems.
Most of the commercial side has not caught up with the demand side. A sales team quoting AI-driven capacity is often still working from a spreadsheet and a phone call to an engineer, rather than live data on power, density and availability. Support fields the same technical questions about rack density and interconnection week after week, questions a properly built system could answer correctly in seconds. Provisioning built for predictable, standard orders starts to strain the moment volume and urgency both increase at once, and with AI-driven demand, they usually do.
The cost of getting this wrong is rising with the cost of capacity. Data centre construction has climbed from $7.7 million to $10.7 million per megawatt between 2020 and 2025, and JLL forecasts a further 6% rise to $11.3 million in 2026. Quoting that capacity from a spreadsheet and a phone call, rather than from live numbers, is where the margin leaks.
The same gap is a cost problem as much as a revenue problem. Every hour a salesperson or engineer spends on work a system could do reliably is an hour not spent on the next customer. Fixing it earns new revenue in some places and lowers the cost to serve in others, and most operators will find both apply. Name them separately rather than folding both into one innovation line on a board slide.
AlixPartners surveyed more than 400 data centre executives and found a split forming: operators that get strategy, integration and execution right pull ahead, and the rest fall behind however much they have already spent on racks and power.
The AI capex debate stops at the supply side.
This year’s trade press tells one story on repeat: power density, GPU estate, grid connection delays, construction cost inflation. It is the capacity story, and the operator’s job in it is to build and wait for the capacity to sell. Whether your own sales, provisioning and support systems use AI barely gets a mention.
Put AI to work where demand becomes revenue.
The shift is already measurable inside telecoms. STL Partners’ adoption tracker recorded a tenfold rise in telco generative AI projects in the year to May 2025, with half of them creating new products rather than cutting costs. Telcos, data centre operators and MSPs are the same buyer here, so the lesson carries across all of them.
Document generation turns a capacity enquiry into a quote without someone drafting it line by line. Support answers a question about rack density, power availability or turnaround time correctly at 2am, with no ticket waiting in a queue until morning. Provisioning keeps up with a surge in AI-driven orders because the system understands the order, instead of routing it to a person to work out what it means. None of it needs new GPU estate or a sovereign hosting decision, only AI put to work inside the systems you already run, on the workloads you already understand.
Take a routine example. A customer emails asking whether a site can support a 40kW rack within six weeks. Today, that question probably reaches a salesperson, who checks with an engineer, who checks a spreadsheet, and replies two days later. None of those steps needed a person. A system that knows the site’s power headroom and lead times could answer immediately and correctly, freeing the salesperson and the engineer for the next customer instead. Swap the rack for a 10G wavelength or a firewall change on a managed estate, and the story is identical.
It has to see real capacity, real orders and real customers.
None of it holds up as a chatbot bolted onto an unchanged back end. The data model has to hold consistently across sales, provisioning and support, so an AI-handled answer matches what the network and billing systems know. The AI layer has to sit inside the OSS/BSS you already run rather than beside it, because a system that cannot see real capacity, real orders and real customer history cannot make a good decision on any of them. Escalation matters as much: when the AI cannot resolve something with confidence, the workflow hands it to a person cleanly, so a customer never meets the system’s limit at the worst possible moment.
The roadmap question
The capacity is being built and the demand is still growing. Whether the systems that sell, provision and support that capacity have kept pace with the systems that generate it is the open question, and for most operators the honest answer today is no.
Closing that gap is product work, not a rip-and-replace: define what your commercial engine needs to do, build it into the systems you already run, and keep it working once it is live. That is what an outsourced product-development team for digital infrastructure does, from idea to revenue, and it is the work PTX takes on.
