Outbound, Reimagined: Building Pipeline with Fewer Heads and Better Margins

Why Outbound Got Hard

Outbound hasn’t failed because sellers got lazy. It’s struggling because buyers are saturated. There are more vendors chasing the same decision-makers across email, LinkedIn and phone, and generative AI has multiplied output faster than quality. Decision-makers now receive hundreds of touches a day and increasingly lean on AI filters to triage messages. We’re edging toward an agent-to-agent world, where a seller’s AI engages a buyer’s AI long before a human reads anything.

“It’s not that buyers massively increased. It’s that sellers did.” — Frank Sondors

The implication is simple: orchestrate, don’t blast. Precision, timing and channel mix now matter more than volume.

From Growth-at-All-Costs to Doing More with Less

Frank’s perspective was forged leading a 50-person sales team through the “add-heads-to-hit-targets” era. Most tools were priced per seat and designed for large teams, while feature requests that could lift conversion per rep rarely made the roadmap. SalesForge was built in response: software that raises output per person and cuts the deadwood—the sizeable segment of a team that neither loves nor excels at the work.

The goal is leaner teams, higher leverage per seller, and a structurally lower CAC.

“For every ten salespeople you hire, one is a natural, two or three can be coached, and the rest become attrition. Software should narrow that gap.”

The 2025 Outbound Reality

Saturation is real. On averages, cold email and cold call results have softened.
Winners still win. Teams that adapt quickly, stack channels, and test relentlessly are growing.
Treat outbound like paid media. Budget, experiments, instrumentation, iteration.

“Ask any sales leader how many tests they ran last month. Most say zero.” — Frank

A Practical Playbook for Modern Outbound

1) Warm the Market, Then Reach Out

Response rates climb when prospects encounter your brand before the first touch.

Daily POV posts from founders/leaders on LinkedIn
Presence where your buyers actually hang out (YouTube, Reddit, communities)
Targeted paid/retargeting to your ICP
Offline tactics that cut through (events, curated direct mail)

When you do reach out, it’s no longer truly “cold”.

2) Build the Backbone Automation

Remove repetitive clicks so humans can sell.

Stand up an automation layer (e.g., N8n, Make.com)
Ship 5–10 flows per week to handle enrichment, routing, alerts, nudges, data hygiene
Trigger follow-ups and handoffs off events, not memory

“We joke daily: should we automate this? The answer is usually yes.”

3) Run a Test & Learn Cadence

Operate like a performance marketer.

A/B subject lines, angles, offers, CTAs, send windows, sequences
Instrument opens, replies, meetings booked, down-funnel quality, and cost per meeting
Kill what doesn’t move the needle; double down on what does

Should You Use AI SDRs? A Brutally Honest Fit Check

Great fit when:

Large reachable market (typically 5,000+ accounts)
Mid-market ACV (around £/$50k or below)
Short sales cycles (2–6 weeks) from first touch to next step
Differentiated offer (not a commodity in a red-ocean category)
You’ve already proved channel fit (cold email/LI has booked meetings before)

Likely a miss when:

You have ≤3,000 accounts (enterprise coverage can be done manually)
ACV ≥ £/$50k, multi-threaded deals, long cycles, heavy procurement
No PMF or channel proof (agents won’t fix a weak offer or fuzzy ICP)
You’re competing in an undifferentiated, crowded category

“AI SDRs are incredible once you’ve proved the channel. They’re not a magic wand.”

Vendor Reality: Why You Must Test More Than One

There is no “Google of AI agents” yet. Vendors differ in:

Data sources and enrichment logic
Email infrastructure and send orchestration
Timing, throttles, warm-up patterns
Prompting strategies, refusal handling and safety rails

Run 3-month pilots and, where possible, parallel tests. Compare meetings/booked per £/$ and down-funnel quality, not just reply rates. Move on quickly if a vendor doesn’t pay back.

Human + AI > Human or AI

The winning pattern is augmentation:

Humans for nuance, negotiation, advanced qualification and multi-threading
Agents for first-touch, research, enrichment, scaffolding and timely follow-ups
Both operating in one environment so signals compound and costs fall

“The smartest teams will run humans and agents together to maximise output and minimise unit cost.”

Team Design for the Next 12–24 Months

Expect leaner teams with more technical firepower.

Must-have competencies

RevOps / Automation Lead
Builds the backbone automation, instrumentation and alerting so sellers aren’t clicking for a living.
GTM Engineer (Sales Development Engineer)
API-literate builder who stitches systems, ships flows, and removes manual work from the funnel.

Sales expectations

Fewer reps, higher output per rep
Clear rules of engagement for agent vs human touches
Weekly automation releases, with alerts for both positive and negative signals

“If no one on your team can script or work with APIs, you’ll over-hire to compensate.”

Cost Reality: Why This Can’t Wait

Acquiring net-new revenue is getting more expensive; price rises are harder in crowded markets. If you can’t lift price, you must reduce unit cost—which means less manual labour per meeting, per opportunity, per pound of pipeline. Automation and AI make that possible now, not someday.