From 8 to 14 clients without a hire — a case study
An anonymized walk-through of an East Coast MSP that ran our automation playbook for 90 days. Numbers, hours saved, dollars retained.
This is a story about one specific MSP. We've anonymized the name and a few details, but the numbers are real. They went from 8 active clients to 14 in 90 days, on the same headcount, while improving CSAT. Here's what they did, in order.
The starting state
An East Coast MSP, 6 technicians, 1 lead dispatcher, 1 owner. Active clients: 8 SMB accounts in financial services and professional services. Annual contract value range: $40K–$120K. Service desk running on ConnectWise PSA, RMM via Datto, voice via a third-party answering service plus the on-call tech's cell phone after hours.
Their core problem: they had a strong pipeline of new clients, but adding a 9th client meant hiring a 7th technician. The next hire would cost $85K loaded plus 9 months of ramp. They couldn't justify it on the new contract economics.
Day 0 — Where the hours were going
We measured for two weeks before deploying anything. The breakdown:
- 14 hours/week per tech on tickets that turned out to be password resets, MFA challenges, or known-issue lookups.
- 22 hours/week of dispatcher time on triage classification.
- 9 hours/week per tech on average dealing with after-hours escalations from the answering service.
- $1,400/month answering service contract.
- ~6 tickets per week falling through cracks (clients texting techs' personal cells, no ConnectWise record).
Week 1 — AI Triage on inbound
We turned on AI Triage as the only change. Dispatcher reviewed the AI's classification on every ticket. By end of week, accuracy was 96%. Dispatcher's daily triage time dropped from 4.5 hours to under 30 minutes.
Recovered capacity on day 7: ~22 hours/week of dispatcher time.
Week 2 — Multi-channel intake
Turned on SMS and Microsoft Teams as verified intake channels. Within 4 days, the "shadow tickets" (clients texting techs' personal cells) collapsed. Tickets per week jumped 18% — not because new tickets appeared, but because previously invisible tickets started getting logged.
Side effect: client CSAT rose. Clients who had been getting irregular service via personal text now had a verified support channel with SLA timers.
"Our ticket count went up 18% in week 2, but the techs felt less overloaded. Turns out the work was already happening — it just wasn't visible. Now it was visible and managed."
Week 3 — Identity verification on privileged actions
Turned on the in-call verification gate for password resets, MFA resets, account unlocks. The dispatcher's first reaction: "this is going to slow us down." The reality: a 4-second Authenticator push, then the action runs.
End of week 3: zero failed verifications from legitimate users. One flagged attempt from an unknown caller pretending to be a controller at a client — denied automatically, flagged for review, traced to a known phishing campaign.
Week 4 — AI Voice on after-hours
Replaced the answering service with AI Voice. After-hours calls now answered in under 30 seconds, verified, triaged, and either resolved with known-issue responses or escalated to on-call with full context.
Hard numbers from week 4 alone:
- 14 after-hours calls answered.
- 11 resolved without paging a human (printer issues, password resets via verified self-service flow, known-issue responses).
- 3 escalated to on-call. All three were genuine P1s.
- Answering service contract canceled. Saving: $1,400/month.
Days 30-60 — Sales acceleration
The recovered capacity opened up the sales conversation. The owner stopped saying "we'd need to hire" on every new prospect. Pipeline closed faster. Two new clients signed in days 30-45, two more in days 45-60.
By day 60: 12 active clients, same 6 techs, same 1 dispatcher. Margin per client ticking up because the same fixed-cost team was supporting more recurring revenue.
Day 90 — The numbers
- Clients: 8 → 14 (+75%)
- Headcount: 7 → 7 (no hire)
- Recurring revenue: +62%
- Dispatcher triage time: 22 hrs/wk → 1 hr/wk
- Tier-1 ticket volume per tech: -41% (auto-resolved or AI-triaged)
- After-hours pages to humans: -82%
- Answering service: Canceled ($16,800/year saved)
- Tickets falling through cracks: 6/week → 0
- Average CSAT: 4.4 → 4.7
- MSP Process platform cost: ~$28,000/year
- Avoided 7th-tech hire: ~$85,000/year fully loaded
The honest framing
The story isn't that AI replaced humans. The story is that AI replaced the next hire. The 6 existing techs and 1 dispatcher are still there, still doing the work. They just have the capacity to support 75% more revenue without working more hours.
The owner's actual quote: "I would have spent $85K on a hire, ramped them for 9 months, and gotten the same capacity. Instead I spent a fraction of that on the platform, got the capacity in 30 days, and now my team is happier." Same business outcome. Different unit economics.
Ship a verified service desk in 30 days.
Book a 30-minute call with a solutions engineer who came out of an MSP service desk. Bring your stack. We'll model the impact with your numbers.