Buzzbloom Media Cuts Missed Enquiries to Zero with SmartFlow AI

The Challenge

Buzzbloom Media is a fast-growing marketing and creative agency running campaigns, content production and paid media for a broad and expanding client roster. As the agency scaled, its front door, the first call or enquiry from a prospective client, quietly became its biggest operational bottleneck.

Account managers were fielding new business calls in the gaps between client work, which meant enquiries were routinely missed, answered late, or picked up by whoever happened to be free rather than someone equipped to qualify the lead properly. Every enquiry then had to be manually logged into the CRM after the fact, often hours later, sometimes not at all, leaving pipeline data incomplete, inconsistent, and impossible to report on with any confidence.

By the time Buzzbloom approached SmartFlow, the agency estimated it was losing a meaningful share of inbound leads before a human ever had a proper conversation with them, not because the work wasn't good enough to win the business, but because the business never got the chance to answer the phone properly.

Key Pain Points

• Missed New Business Calls: enquiries arriving outside office hours, during client meetings, or at peak periods went straight to voicemail with no consistent, timely follow-up.
• Inconsistent Lead Qualification: whoever answered asked different questions in a different order, meaning leads reached account managers with wildly inconsistent information, some over-qualified and over-served, others barely screened before a meeting was booked.
• Manual, Delayed CRM Entry: enquiry details were logged by hand, well after the actual call, creating gaps, duplicate records and reporting errors that made pipeline forecasting unreliable.
• Zero Visibility on Call Outcomes: there was no consistent way to see which enquiries converted, which stalled, or why. Decisions about new business process were being made on instinct rather than data.
• Admin Eating Into Billable Time: senior account staff were losing hours every week to intake admin instead of client strategy, creative direction and delivery, the work that actually drives agency revenue and retention.

“

We were a creative agency accidentally running our new business pipeline like an understaffed call centre. Every missed call was a potential client we'd never get back, and we had no real way of knowing how many we were losing. We needed something that could pick up every single time, ask the right questions properly, and just get it into the system correctly, without anyone having to think about it.

The Solution

SmartFlow AI implemented intelligent CRM automation alongside a tailored AI voice agent to streamline customer interactions, automate manual processes, and create a single connected workflow that lets Buzzbloom's team focus on higher-value creative and strategic work, rather than bolting on two disconnected tools and hoping they'd talk to each other.

Mapping Buzzbloom's Real New Business Process

Before building anything, SmartFlow's team spent time with Buzzbloom's account managers to understand how new business conversations actually happen, the qualifying questions that matter, the language the agency uses with prospective clients, the red flags that signal a poor-fit enquiry, and the specific criteria that separate a lead worth a same-day callback from one that can wait.

This wasn't a generic discovery call; it was a full audit of the agency's existing intake process, CRM structure and pipeline stages, so the finished system would reflect how Buzzbloom actually sells, not a template.

AI Voice Agent, Built Around Buzzbloom's Own Playbook

Rather than deploying an off-the-shelf voice bot with a rigid script, SmartFlow designed a voice agent trained specifically on Buzzbloom's own qualifying framework and tone of voice, built on Retell for call orchestration and ElevenLabs for natural, human-like voice generation, with GPT-4 powering the conversational logic underneath. The agent handles:

• Instant First Response: every inbound call is answered immediately, 24 hours a day, seven days a week, removing the risk of a missed enquiry regardless of time, day, or how busy the team is.
• Consistent, Structured Qualification: a natural but structured conversation captures budget range, campaign type, project timeline and decision-maker status on every single call, regardless of who's on shift.
• Genuine Conversational Handling: the agent understands follow-up questions, handles interruptions and can answer basic questions about Buzzbloom's services in real time, rather than forcing callers through a rigid menu of pre-set options.
• Smart Handoff Logic: when a call signals a high-value or nuanced opportunity, the agent recognises this and books the caller directly into the relevant account manager's calendar, live, on the call, rather than making every enquiry wait for a callback.
• Natural Escalation & Fallback: calls that fall outside the agent's scope, existing client account issues, sensitive matters, or simply a caller who wants to speak to a person, are identified immediately and routed straight to a human, with full context passed along so nobody has to repeat themselves.

Intelligent CRM Automation

Every conversation the voice agent handles is automatically logged and structured directly into Go High Level in real time, no manual entry, no delay, no lost detail. SmartFlow built a custom n8n automation layer connecting Retell's call data to Go High Level's pipeline, which:

• Creates or updates the contact record the instant a call ends, fully populated with every qualification detail captured during the conversation.
• Automatically tags and scores each lead based on qualification responses, so account managers can prioritise their follow-up list at a glance instead of re-reading call notes.
• Moves each contact into the correct pipeline stage automatically, replacing the manual drag-and-drop process the team relied on before.
• Triggers automated follow-up sequences for leads that aren't ready to book immediately, so no enquiry goes cold simply because nobody remembered to chase it.
• Flags duplicate or returning contacts automatically, preventing the double-entry and messy records that come with manual logging.

One Connected Workflow, Not Two Bolted-Together Tools

The core design principle behind the entire engagement was connection, not addition. The voice agent and CRM automation were engineered as a single continuous workflow, orchestrated end-to-end through n8n: a call comes in, gets answered and qualified consistently, and the outcome is already sitting correctly in the Go High Level pipeline, scored, tagged and staged, before the account manager has even seen it. There is no admin step in between, because the system was built so that none is needed.

Performance Visibility, Built In From Day One

Because Buzzbloom previously had no consistent way to see what was happening to its own leads, SmartFlow built a reporting layer alongside the automation itself, giving the team real-time visibility into:

• Call volume and answer rate, including out-of-hours performance.
• Lead qualification outcomes and scoring distribution.
• Conversion from enquiry to booked meeting, and meeting to won client.
• Time-to-first-response and time-to-booking, tracked automatically rather than estimated.

For the first time, Buzzbloom's leadership could see exactly what was happening at the top of their funnel, not a rough estimate, but real numbers, updated automatically.

What Made This Different

Buzzbloom had looked at generic voice AI tools before and walked away unconvinced. They were scripted in a way that didn't reflect how the agency actually spoke to prospective clients, and, critically, they weren't connected to the CRM at all. Bolting a voice bot onto the front of the same manual admin process would only have moved the bottleneck, not removed it.

SmartFlow treated the voice agent and the CRM automation as one system from day one, built around Buzzbloom's actual qualifying criteria, tone of voice and pipeline structure rather than a generic script sitting on top of unconnected software. The result wasn't just faster call answering, it was the first time Buzzbloom had ever had consistent, structured, real-time pipeline data, without adding a single extra admin task to get it.

Lessons & Challenges

Building a voice agent that sounded genuinely consistent with Buzzbloom's brand, rather than like a generic call bot, took more iteration than expected. Early versions of the conversation flow captured the right information but didn't sound like Buzzbloom; getting ElevenLabs' voice output and GPT-4's conversational tone to match how the agency actually talks to prospective clients required several rounds of scripting and voice tuning before it felt right on a live call.

Qualification logic also needed real-world calibration. The initial scoring rules, built from account managers' stated criteria, didn't always match how they actually prioritised leads in practice, there's often a gap between what a team says it looks for and what it really responds to. SmartFlow refined the scoring model against real call outcomes in the first few weeks of live use, rather than relying solely on the upfront discovery sessions.

Connecting Retell's call data cleanly into Go High Level via n8n also surfaced edge cases that don't show up in testing, calls that dropped mid-qualification, returning callers with existing CRM records, and enquiries that didn't fit neatly into a single pipeline stage. Each of these needed its own fallback logic rather than a one-size-fits-all rule, built and refined through the first month of live calls rather than assumed upfront.

Finally, not every caller wants to talk to an AI agent, however well it performs. Fallback paths to a human were built in from the start, but real usage data shaped exactly when and how that handoff should trigger, ensuring the agent handled the calls it was built for, and stepped aside cleanly for the ones it wasn't.

Business Outcome

Within the first full month of deployment, Buzzbloom went from missing a meaningful share of inbound enquiries to answering 100% of new business calls, every one qualified consistently, day or night, regardless of who was in the building. Lead qualification accuracy hit 92%, giving account managers confidence that a lead landing on their desk was genuinely worth their time.

The agency now saves an estimated 35+ hours a month in admin time that previously went into manual call handling and CRM entry, translating to over £45,000 a year in reclaimed staff capacity, time now spent on client strategy, creative direction and delivery instead of intake admin.

More than the hours saved, Buzzbloom's leadership now has something they never had before: a real, measurable view of their own new business pipeline, and the confidence that no enquiry, whenever it comes in, will ever go unanswered again.

“

It's not just that we stopped missing calls. We finally know what's actually happening in our pipeline, in real time, without anyone having to chase it. That's been the bigger shift for us.

Implementation Timeline