Contact center from zero — without per-seat fees
A fintech client launched outbound sales with no dialer. We designed, deployed, and ran a full VICIdial stack — scaling active agents from 7 to 14 (2×) while talk time grew 65×, lifting SQL conversion 4.6×, with $0 per-seat licensing (open-source VICIdial seating; infrastructure and support separate).
Executive Summary
The client — a fintech company serving individuals with consumer debt obligations — had zero outbound calling infrastructure when planning began in mid-2025. Dialer go-live and the first production talk-time month are July 2025; MQL contact-rate tracking in the charts includes the late-June ramp. They were handling leads manually, with no dialer, no call tracking, and no visibility into agent activity. The goal was to build a scalable outbound operation from the ground up, fast, and without the per-seat costs of commercial platforms.
Over the engagement window (Jun 2025 – Feb 2026), we deployed and managed a full VICIdial environment integrated with GHL (GoHighLevel) as the CRM. The system was configured for predictive outbound dialing against MQL (Marketing Qualified Lead) lists, with dispositions feeding directly back into the sales pipeline. The results speak through the data: contact rates climbed from ~5% to a peak of 57%, sales qualified lead conversion grew from 4.6% to 21%, and total agent talk time scaled 65× (3.74h → 241.88h/month) while active headcount only doubled (7 → 14) — all with no per-seat licensing fees as the team grew.
The Numbers, Month by Month
All data below is sourced directly from VICIdial logs and the GHL (GoHighLevel) CRM dashboard. No figures are projected or estimated — these are the actual metrics recorded during the engagement from June 2025 through February 2026.
The Situation
When this project began in mid-2025, the company had no automated dialing. Sales reps worked a basic CRM list one-by-one — no visibility into who had been called, no recordings, and no performance measurement. A small team carried the whole outbound effort until the dialer went live.
The debt restructuring market is competitive and time-sensitive — leads go cold fast, and the first company to make live contact typically wins the client. With a manual process, the team was reaching less than 5% of their MQL list in any given week. They needed automation, speed, and structure.
- No dialer — agents dialling numbers manually
- Zero quality control
- Difficult to monitor agent activity or talk time
- No lead recycling — attempted leads were often abandoned
- MQL contact rate below 5% per week
- 7 active agents on the first production talk-time month
- No defined MQL → SQL pipeline tracking
- Scaling the team meant manually adding to a spreadsheet process
- Predictive dialer reaching up to 57% of the MQL list per week
- 100% of calls recorded and accessible via agent/supervisor login
- Real-time dashboards showing live agent status and talk time
- Automated lead recycling — unreached leads re-queued by priority
- SQL conversion rate at 21% — up from 4.6%
- 14 active agents on the same infrastructure at zero per-seat license cost
- Full GHL CRM integration — dispositions synced automatically
- New agents onboarded within hours of account creation
How It Unfolded
The deployment wasn't a single big-bang launch — it was a managed build-up. Here's what actually happened, month by month.
Server Setup & First Calls
VICIdial installed and configured on a dedicated server. SIP trunk connected and tested. First campaign created with the initial MQL list imported from GHL. Early agents onboarded onto the stack. Talk time for the month: 3.74 hours across 7 active agents — low, but the infrastructure was live and working.
Workflow Refinement & GHL Sync
Built the GHL integration — leads now flowing into VICIdial automatically as they hit MQL status. Disposition codes configured. Lead recycling rules set up for first time. Talk time reached 64 hours/month by September. Contact rates still in the 4–6% weekly range — system was working but list quality needed attention.
Dialing Optimisation & Team Growth
Adjusted the dialing ratio and answer machine detection settings after analysing call logs. Contact rate broke through 10% for the first time in October, then climbed to 23% by late November. Monthly active agents moved into the high single digits; the pre-sales dialer cohort sat around 2–5 agents in this window. SQL conversion tracking introduced — initial rate: 4.6%.
Scale-Up — The Inflection Point
The biggest single month of growth. Monthly active agent headcount expanded to 13. Contact rates crossed 33% and peaked at 44% in a single week. Monthly talk time jumped from 67 hours to 154 hours. SQL rate hit 15%. This was the month the system proved it could scale.
Mature Operation — Consistent 50%+ Contact Rates
System reached steady-state performance. 14 active agents on the monthly talk-time series (pre-sales cohort ~7–8). Weekly MQL contact rates consistently between 50–57% with a February peak of 57%. SQL conversion rate at 21% by February. Total talk time: 241.88 hours in January alone. Ongoing focus shifted from setup to coaching, reporting, and campaign-level optimisation.
Results
Eight months of sustained improvement across every metric that matters for an outbound sales operation. No projections — all figures are drawn directly from VICIdial logs and the GHL dashboard.
By the Numbers
Jun 2025 → Feb 2026 — Source: VICIdial Log + GHL Dashboard
"Before this system, our agents were spending most of their day waiting on calls that never connected. Now they're in live conversations almost immediately. The difference in what the team can do in a day is hard to overstate."
What Drove the Results
The Contact Rate Story
- Predictive dialing eliminated dead wait time — agents only connect when someone picks up
- Multi-attempt recycling logic kept unreached leads in the queue instead of abandoning them
- Answering machine detection tuned to minimise wasted connections on voicemail
- Contact rate crossed 30% once dialing ratios were calibrated in December — and held above 50% through January–February
The Conversion Rate Story
- Call recordings gave supervisors the data to coach agents on specific call moments
- Custom disposition codes revealed which lead segments converted at higher rates
- GHL sync meant sales advisors could follow up on warm leads within minutes of a positive disposition
- SQL rate grew every single month — 4.6% → 12% → 15% → 18% → 21%
Buyer FAQ — extractable answers
Standalone answers for procurement and ops leads. Metrics below match the Chart.js series on this page (active agents 7→14, MQL contact ~5%→57% peak, SQL conversion 4.6%→21%). If your existing floor needs better dialing, dispositions, or supervisor visibility, see our VICIdial optimization service. Install path detail: ViciBox 12 on GCE.
VICIdial vs per-seat commercial dialers — what is actually $0?
Answer: Open-source VICIdial seating has $0 per-seat license at 14 active agents on this engagement. Infrastructure, carrier SIP, numbers, and support are separate — do not read “$0” as total cost of ownership.
- Commercial per-seat stacks scale license cost with headcount; OSS seating does not.
- You still pay servers (or GCE), trunks, DIDs, and people who run the floor.
- Security posture here is practice-informed — not a certification claim.
How does the GHL disposition loop work in one chunk?
Answer: Agents disposition in VICIdial; webhooks update GoHighLevel so pipeline stages and advisor notifications move without spreadsheet handoff.
- Example: “Interest Confirmed” / “Callback Scheduled” → GHL contact stage updates.
- Goal: zero manual CRM typing after the call for the core disposition set.
- Lesson learned: build the sync in week one — delayed integration created cleanup debt.
7→14 agents: what actually changed?
Answer: Active headcount doubled (7→14) while talk time grew ~65× (3.74h → 241.88h / month). Growth was not “more seats alone” — dialing ratio, recycling, AMD tuning, and supervisor live view compounded.
- Pre-sales cohort on the dialer: 2 agents (Oct 2025) → 8 (Jan 2026) as a subset of total active.
- Same OSS seating economics at 14 — no per-seat license cliff.
- Charts on this page are the source of truth for weekly series (not marketing round-numbers).
What does MQL contact rate 5%→57% mean?
Answer: Share of Marketing Qualified Lead dials that reached a counted contact outcome in the tracked window — launch-era ~5% rising to a peak of 57% (Jun 2025 ramp through Feb 2026 peak on the chart).
- It is an operational contact metric on MQL lists, not a sales close rate.
- SQL conversion is separate: 4.6% → 21% (about 4.6×) in the Oct 2025–Feb 2026 window shown.
- Inflection included dialing-ratio calibration (e.g. contact-rate step-up in Dec 2025).
When should you not hire us?
Answer: Skip this engagement model if you need a pure license reseller, a black-box SaaS with no ops ownership, or a guarantee of a specific contact rate without list quality, carriers, and supervisors in the room.
- We design and run VICIdial around workflow — not a one-week “install and disappear” unless scoped that way.
- If you only want cloud ISO mythology with no carrier/floor work, read the GCE install log instead of buying a floor program.
- Regulated claims: we do not sell “certified” badges we do not hold.
What We Learned
Honest observations from 8 months of building and running this operation — including what we'd do differently.
What Made the Biggest Difference
- Dialing ratio tuning in December was the inflection point. A single configuration change drove a week-over-week contact rate jump from 23% to 33% — the biggest single-week gain in the entire project.
- Call recordings changed agent behaviour. Once agents knew their calls were being reviewed, script quality improved noticeably within two weeks — without formal retraining.
- Disposition data is the most underused asset in most outbound operations. Mining which dispositions correlated with eventual SQL conversion helped the team stop wasting time on lead segments with low potential.
- Starting lean on the pre-sales cohort (2 agents in October) was the right call. It allowed the workflow, integration, and recycling logic to be stress-tested before scaling the broader active headcount to 14 — avoiding failures that would have been much costlier later.
What We'd Do Differently
- Set up the GHL integration in week one, not week four. The manual period before the sync was built created data inconsistencies that took time to clean up later.
- Introduce AMD tuning earlier. We ran with default settings for the first two months. The contact rate improvements in Q4 suggest we left gains on the table in August and September.
- Build the supervisor dashboard before the first agent logs in. Operating blind for early weeks meant decisions were reactive rather than proactive.
- Document the disposition logic in a shared runbook. As the team grew, agents were using disposition codes inconsistently — a defined guide from the start would have kept the pipeline data cleaner.
Building something similar?
Starting from zero or replacing an existing stack — we design and run VICIdial around your workflow, with real integration and no per-seat fees. For the cloud install path, read ViciBox 12 on Google Compute Engine.