North Energy, a residential solar company, kept everything it knew in 73 files on a shared drive. The warranty terms were in there, along with a separate permitting contact for every municipality it builds in, the sales scripts, the finance program details and the training decks. Some of the files were scanned PDFs that no text search could read.
That setup charged the company twice for the same question. In the office, someone who needed a warranty term or a permitting contact either dug through folders or pulled a supervisor off their own work. In the field, a rep at a customer's kitchen table with a question about a finance program's terms had nothing to reach for, because the training binder doesn't come to the appointment.
What got built
North Energy's team worked with Level 99 to build an AI support agent that lives inside the company's internal tools and works from a phone. A staff member types a question in plain language and gets an answer pulled from the company's own documents, with the source files named under every answer.
It took about a week. A pipeline re-syncs from the shared drive every night, picking up new and edited documents and leaving the rest alone. Text recognition ran over the scanned PDFs, so their contents became searchable for the first time. In total, 2,570 searchable passages were indexed from the 73 source documents. The chat widget sits inside internal pages the team already used, so nobody had to install an app or remember another login.
What changed for ops and sales
For ops, the questions that used to interrupt a supervisor now go to the chat box first. The knowledge base keeps itself current: someone edits a document, and the agent knows about the change by the next morning. Nobody uploads anything to a separate system.
For sales, the training library now travels to the appointment. A rep can pull the exact objection-handling language or a finance program's real terms mid-conversation instead of promising to follow up later. That promise is the gap the sixth-call problem describes, and it's where a lot of deals start to die. New reps also learn from the same source material the top closers use, instead of a binder nobody updates.
Every answer cites the file it came from, so a staff member can open the policy and check it. Nobody has to take the agent's word for it.
The math, shown so you can check it
Lookups dropped from 2 to 10 minutes down to under 30 seconds, which works out to as much as 95% faster. Digging through folders was the fast option before. Waiting for a supervisor to free up was the slow one.
That adds up per person. One employee making three lookups a day at roughly five minutes each gets back about 15 minutes a day, or more than an hour a week. Multiply by headcount.
Escalations were the expensive part, because a question that reaches a supervisor uses two people's time. If even half of daily questions used to escalate, cutting that in half gives supervisors hours back for work only they can do. Upkeep is zero, since the knowledge base re-syncs on its own every night.
What North Energy hasn't measured, and what this post won't claim, is close rate. Nobody tracked whether a rep who answers a finance question on the spot closes more deals than one who promises to follow up. What we can say is that every "let me get back to you" that becomes an answer at the table is one fewer follow-up call, and follow-up calls are where deals stall. That gives a reason to expect a lift. We haven't measured one.
Where else this problem shows up
North Energy's documents happened to be warranty terms and permitting contacts. The same problem appears anywhere a company's real answers sit in files nobody can search quickly: service manuals, compliance checklists, pricing sheets, old email threads that never made it into a wiki. A better filing system wouldn't have helped much here. What helped was an agent that reads the files already there and answers in plain language, with the source attached so anyone can check it.
The break-even math on AI versus hiring runs on the same logic. A slow answer costs time, and time has a price whether it's a customer on hold or an employee waiting on a supervisor. An internal support agent is the same tool pointed at a company's own staff instead of its callers. Level 99 builds both kinds, on the same plans starting at LVL 10.
The full case study has the complete build details and the rest of the math.
FAQ
Did lookup time actually drop, or is that an estimate?
It's measured against the company's own prior process: 2 to 10 minutes of folder digging or waiting on a supervisor, compared with under 30 seconds through the agent. The 95% figure is arithmetic on those two numbers.
Does this agent replace the supervisors who used to answer questions?
No. It takes the first-line questions that already have a written answer, so supervisors stop getting interrupted for them. Anything outside the documents still goes to a person.
What happens if a document is wrong or out of date?
The agent only knows what's in the source files, and it names the file for every answer. If a document is wrong, someone edits that document, the same as before. The agent re-syncs nightly, so the correction shows up the next morning without anyone touching the chat system.
Why not just train staff to search the shared drive better?
Part of the problem was searchability. The scanned PDFs couldn't be searched at all until text recognition ran over them. The larger part was access. A rep in a customer's kitchen doesn't have a laptop open to a shared drive, and a supervisor interrupted twenty times a day hasn't been badly trained. The agent puts the answers where people already are.