How EBB Charge is scaling from 20 to 100 sites without growing the ops team behind them, with AI built into the WhatsApp channel the crews already work in.
sites, same ops team
crews and sales grow, ops doesn’t
crew adoption
nothing new to install, learn, or open
of morning relay, gone
replaced by a summary: done, stuck, silent
from kickoff to crews live
first SOP answers in production
EBB Charge installs EV charging infrastructure across Sweden, twenty sites in 2024, a hundred planned for 2026. Coordination breaks first at that pace: keeping every site visible while the crews multiply.
You could see it at seven every morning. An hour of WhatsApp calls, manager by manager, to assemble the picture of what was happening across the sites, then managers relaying the answers on to their electricians. The procedures lived in systems built for a desk, not for the field. At twenty sites that holds. Toward a hundred it becomes the tax on every new site: more status to gather, more exceptions, more of the same questions.
If you run a growing company with people in the field, you know the cost that creeps in around the headcount: the chasing, the same answers repeated, the work paused while someone digs up the current procedure or checks whether the job changed.
Sasha Djakovic came asking for an AI assistant. The other pitches were what vendors always pitch: a dashboard, a training portal, an SOP database. One more login between the crew and the answer. After the first call we proposed the opposite: no new app at all. The electricians live in WhatsApp, so the AI became a WhatsApp contact, a number the crew messages like a colleague.
The whole day now runs through one WhatsApp number.
Morning check-ins go out on their own, evening completion reports come back, and leadership starts the day with a summary of who responded, who didn’t, and what’s blocked. Nine in ten evening reports arrive before anyone has to chase them. In between, the crew asks it procedure questions and gets answers in Swedish, median 24 seconds, grounded in EBB’s own SOPs and wired into Monday.com so the live job context comes attached. No new app, no login, no training.
Now an electrician on a rooftop just texts. The question goes in like any other message, and the answer comes back in seconds, the right procedure with the current job context attached. You can watch that moment in the video above. To the crew it became the sharpest colleague on the team: one that has read every procedure, never loses the thread of a job, and answers in whatever language you ask.
Crews trust it because it is constrained. It is held to EBB’s own documentation and job context, it is not a general chatbot, and when the answer is not there, it says so. That matters on a live site, where a confident wrong answer costs more than no answer.
The fragmented systems started behaving like one, without a single tool replaced.
“Vendors pitched portals nobody opens. Luxdone built AI inside the channel my team already lives in.”
The easy way to absorb growth is more coordinators, more chasing, more of leadership’s day routing information. EBB chose a different model: the business load grows while the ops layer stays flat, because the system absorbs the repeatable coordination.
EBB still runs in the physical world: crews, electricians, charging sites, parts, grid connections, and deadlines. The AI does not do that work. It removes the coordination drag around the work, so the operation gets easier to see as it gets bigger.
People kept the judgment: site visits, customer relationships, the calls that need someone who has seen the failure before, now with time to do them properly. What moved into the system was the repeatable part: the chasing, the relaying, the same question answered again, the status that used to run through leadership.
Not fewer humans. Fewer human bottlenecks.
The key was building in a way that leads to adoption, and every step of the playbook follows from it.
As EBB scales toward a hundred sites, the same system takes on more of the operation: sales-to-site handoff, live escalation when a job gets stuck, an operational view that replaces the morning call chain. Sasha puts the direction more bluntly: “AI in the channel your team already uses. That’s the new ERP.”
“Luxdone understood how my people actually work. Everyone uses it, every day.”
Varje berättelse på den här sidan började med ett 30-minuterssamtal. Berätta om er situation, så berättar vi hur det första steget skulle se ut.
Boka introsamtal