The call in one breath

I removed my team and tried to replace them with agents before the agents could finish the work. So everything fell back onto me.

What this is // the frame

The pattern has a name: premature delegation removal + founder-is-the-bottleneck. You did not remove the work, you removed the layer between the work and yourself. That is squarely Mads's wheelhouse (delegation, hiring, "do 50% less"), so this call spends its 45 minutes on the human and management layer, not on the tech.

Apex: fix the operating system - so you can run the $100k machine. Two levers get the airtime in the call: (A) team to agents and (B) time, working ON not IN. The destination is the 6-month climb to $100k/mo by Dec via NEWCO (the $100k by Dec tab); the engine that powers it is on ONE.ie + Tony. The OS fix is what lets you lead that climb instead of being its bottleneck.

Use ← → to flick between tabs.

Format: hybrid

The 45 minutes

Tight scripted opener, hand the floor to Mads on your top lever, force the decisions, capture the homework. Do not let it become a nice chat that changes nothing.

Pick Lever A or Lever B as the primary deep-dive. Park the other on the Decisions tab so you do not feel you are dropping it.

Lever A

Team to Agents: did I fire a delegate, or fire the delegation?

You do not delegate by removing the person. You delegate by moving an outcome to a delegate, then verifying it holds without you. You removed the delegate before the replacement could prove it closes the loop.

Question bank // 15 in 5 groups

Bridge vs push vs hybrid // decision tree

A · Bridge with a human now

Pick when a client is at churn-risk today and no agent closes its loop without you. Mads framing: "A dropped client is permanent. A bridge hire is reversible. De-risk the irreversible one first." It is ONE trusted operator who turns the agents' plans into deliverables - not the old team. The trap: re-hiring the old shape and re-creating the cost/trust problem.

B · Push the agents to autonomy

Pick when nothing is slipping, you have runway, and the gap is spec-quality not capability. It is picking ONE end-to-end loop and making it close with zero you-in-the-middle, hard deadline + kill criteria. The trap: "push" with no deadline becomes infinite rescuing - not closing by the deadline IS your capability answer.

C · Hybrid (the most likely answer)

Pick when live delivery is fragile AND you have no runway to gamble. It is a human bridge on live retainers, agents on net-new/internal work, migrating loops one at a time only after each proves it closes. The trap: letting "hybrid" mean "everything still routes through me."

Agent autonomy // the climb

What Mads will push back on // be ready

• "You fired the team before the replacement existed." Even with the PBN issue, the role should have been bridged.
• "You are solving a management problem with software." The bottleneck is org design - spans of control, owned outcomes, accountability.
• "You will not let go of QA." A delegation-maturity issue, not an AI-capability issue.
• "You skipped the cheap experiment." One loop, one bridge person, two weeks - what tells you which way to go?
• "Where is the data?" How many loops close without you today, and what does each dropped loop cost?
• "This is the same lesson as last time." Nov 2025: set up the funnel yourself, then delegate. You tried to delegate to agents before proving the process closes.

Lever B

Time: not bad management, single-point-of-execution.

Every deliverable, cutover, prospect and R&D thread routes through one person who is also doing the work. With no proven delegate, your only lever for "more output" is "more Donal" - so you clone your attention across 20 chats and they clobber each other.

Do 50% less // owner / delegate / automate

The Delegate column IS your integrator's first job spec. Almost everything cuttable is delivery + ops; almost everything to do MORE of is sales + building a delegate.

Question bank // 12 in 4 groups

"Ideal week" // react to this

Hard cap: 3 concurrent active lanes, not 20. Mornings = ON-business. Afternoons = IN-business, batched. Two protected deep blocks a week. Sales goes first.

Rules: no new chat opens until an old one closes clean; Friday close-out mandatory; R&D + partnership get exactly one block/week until the core is selling; sales is non-negotiable and goes FIRST.

Working ON vs IN // self-diagnostic

Fill before the call: hours/week [YOU FILL] · % delivery+ops [YOU FILL] · % growth+selling [YOU FILL] · concurrent chats/day [YOU FILL] · hours/week on R&D [YOU FILL].

Last 10 minutes

Force a written answer on each before you hang up.

A 45-min call that produces ONE committed decision beats one that produces ten ideas. Pick the top fork as the forcing function; park the rest.

Bring data

The numbers.

The receipts - why $100k is not a hope

A year of building, not a pitch.

The $100k plan rides a machine that already exists. In 2026 I built an operator OS that runs a 9-client agency at near-solo headcount, and Tony brings a production agent substrate plus McKinsey-grade strategy pedigree. The brag here is implicit - it is just the receipts.

The operator OS // 1 operator, 8-person output

What I built this year // systems, not heroics

Tony's side // the platform + the pedigree

The destination - what the OS fix is FOR

$14k to $100k by December.

The NEWCO machine: a 33/33/33 company (Donal CEO, Tony CTO, Tomas CRO; John funded toward equity) built on Tony's ONE platform, sequenced ACTIVATE → MONETIZE → SCALE. This is why the operating-system fix comes first: you cannot run a $100k machine as the single bottleneck. Levers A and B are how you free yourself to be its CEO.

The climb // 3 quarters to $100k

How $100k is built // 5 stacked levers, not one

The 4 founders // one north star each (manage the team)

The channels // big networks, Tomas, ads, custom links

Questions for Mads // set up + manage the climb

This tab is the strategic overlay; the Decisions tab is still where you force the 2-3 calls. The point of showing Mads the climb is to get the setup + management right and find the gaps, not to add more to your plate.

The growth engine (not the build trap)

Scaling ONE.ie with Tony.

The both-halves play: OO sells the services (the $6 of a budget), ONE.ie is the software (the $1) - together the whole $7. A new 33/33/33 company (Donal / Tony / Tomas), launching ~July 2026, with Tony keeping the ONE.ie tech. This tab is the SELL side - scaling distribution is the opposite of hiding in R&D (see the Upside contrarian flag).

Site stats // ONE.ie platform, as of 2026-06-08

Pricing model: credits (not flat $/mo) - a $500 pilot = 5M credits, ~$0.0001/credit + Tony's 10% platform margin. Honest read: ONE.ie is a pre-test tester platform; the live site shows no public usage metrics yet. Lead with the loop logic + the org chart, not customer counts.

The managed-growth loop // the machine

Distribution // 2 affiliate channels + white-label

Lead-list inventory // what is already in the pipe

Questions for Mads // gap-find, then let him lean in

If he offers to introduce or promote it, that is the channel door opening on his terms. Note it, thank him, follow up after - do not close on the call.

Speak his language

Mads's playbook.

Map your mess onto his models so the conversation runs in his frameworks, not abstract complaints. "Do 50% less, achieve 150% more."

DISC // the 4 styles

Effective Delegation // the real bottleneck

• Delegate the outcome, not the task. "I want you to own this, and if you think it can be done better, improve it."
• Why founders fail: "the spider" - a hand on every pie. The block is psychological: fear they get it wrong, so you will not let go.
• Delegate-first priorities: tasks you hate; tasks eating 50%+ of your hours; low-dollar admin. That is your agent-allocation list.

Hiring // not "the best candidate"

Attitude over skill: "I will take 80% less skill for 10% more attitude." He would rather promote from within than import an outside expert who imposes a foreign system. Define before you hire: the role's outcome (not title), the DISC style it needs, the attitude floor, whether you can grow into it.

Cadence // 1:1s + feedback + meetings

• 1:1s - weekly, a 10/10/10 shape (their stuff / your stuff / development). Encouragement, not surveillance.
• Feedback - a Ternary format; a 10-50 second thing, frequent and low-drama. For agents: a short weekly review of outputs + tight corrective feedback into the spec.
• Team meetings - align the unit (the "phalanx"), not status-report.

His exact 10/10/10 + Ternary internals live behind his paid course; do not over-claim the detail.

Do 50% less // the mechanism is leverage

Your problem → his model

/10x divergent menu

Upside - the plays.

Several plays are already folded into the tabs (cut-list to Lever B, one-decision to Agenda, accountability loop to Capture). The rest are a menu to action after the call.

Top of the stack: #1 the cut-list is the input every other play depends on; #4 keeps the call from being a nice chat with no change; #3 names the root cause. #12 is the system-building-instead-of-selling trap - park it until 90 days of proof. [COMPLIANCE-CHECK Rule 12: #7's cron stays internal pipeline-prep, never auto-drafts a client email.]

During + after the call

Capture.

A coach's value is in the homework actually getting done. Type into this live; read the commitment back to Mads before you hang up, with a date.

Mads's key advice

My homework / commitment

Commitment:
Due date:

Decisions made

• Stop doing:
• Bridge with a human (Y/N + who):
• Integrator (identify / hire / who):
• Reprice this week:
• Re-open intake (now / hold until):

The defining fork (pick one)

Agent-led now / bridge first / hybrid:

Follow-up booked?

Y / N · date: