How Top Agencies Are Using AI to 10x Output

Executive Summary

1. Time wins deals. Repeatable tasks move from hours to minutes with lightweight AI workflows and a few hard rules on verification.
2. Research fuels pipeline. Deep research stacks surface high fit accounts and talking points that raise reply rates.
3. Trust beats hype. Clear disclosure, show your work, and align AI usage to outcomes clients actually care about.

The 15 Minute Workflow: Compress the Repeats

Claim in the episode: Work that used to take one to two hours now takes 10 to 15 minutes with an AI assisted flow. Here’s how to replicate:

Template the brief:

• A one page prompt brief per task type: inputs, brand voice, must include facts, must avoid claims, and the acceptance checklist.
• Include an example that is good and an example that is unacceptable.
  

Generate, then verify:

• First pass in ChatGPT or equivalent.
• Second pass is a verification checklist that a human must complete. No publish until every box is green.
  

Proof with a tracer:

• Maintain a running log: prompt version, model version, links to sources, who approved. This is your audit trail when clients ask how you used AI.

  
  

Ship a narrow win:

• Move one recurring asset through the pipe end to end: title tags, outreach briefs, weekly summaries, meeting notes. Set a 15 minute SLA.

  
  

Guardrails that keep you safe:

• Every output must be checkable in under five minutes.

• No generated numbers without a source.

• Anything legal, medical, or compliance heavy is out of scope.

Lead Sourcing That Finds “Real” Prospects

Jason’s team uses deep research to decide who to contact and why, before any inbox outreach begins. Key mention of Gemini Deep Research and companion LLMs.

The stack and flow:

• Signals to track: hiring velocity, tech stack changes, content gaps, leadership moves, new funding, shifting location strategy.

• Research workflow:

  • Define ICP and exclusion list.
  • Run deep research for 50 accounts to surface 3 to 5 specific triggers per account.
  • Human verifies the top 10 accounts and trims noisy signals.
  • Draft first outreach lines that cite a real trigger, not a guess.

Outcome: Fewer emails, better replies, shorter time to first qualified call.

Content That Actually Helps Sales

The SEO playbook stays steady, but distribution and measurement shift. Jason suggests moving content closer to revenue events and measuring usefulness by how fast it gets used by sales or clients, not only by traffic.

Use content like a tool, not a magazine:

1. Bottom funnel formats: battlecards, teardown memos, implementation checklists, pricing explainers, objection handlers.
2. Operational rule: if a piece of content is not used in a live deal this week, it is not bottom funnel.
3. Traceability: add a “where used” field to every asset and log the deals it supported.

Navigating Client AI Fears

Client anxiety about AI feels like past waves of “SEO is dead.” The answer is not secrecy, it is education and posture.

A simple client framework:

1. Disclose with outcomes: “We use AI to speed research and drafting. Humans verify facts and final decisions. You get faster delivery and lower cost per experiment.”
2. Show your work: share your acceptance checklist and the audit log for sensitive deliverables.
3. Co create risk rules: define clear red lines with the client on where AI is allowed, where it is not, and who signs off.

Why this works: transparency builds trust, and trust buys you room to move faster.

Coaching Your Team to Adopt AI

Adoption rises when people see what is in it for them, not when you force tools. Early in the episode Jason describes the importance of starting with the win for the individual and making it visible.

Playbook for managers:

1. Start with one job to be done: “Reduce weekly reporting time from 2 hours to 15 minutes.”
2. Pair build: one operator and one AI fluent partner co write the brief and checklist.
3. Share the before and after: time saved, errors prevented, dollars influenced.
4. Reward reuse: celebrate the teammate whose checklist gets adopted across pods.

Reality Check: Your Market Still Buys Like Humans

Jason notes a useful grounding point. Most customers sit outside our tech bubble, which means adoption curves are messy and uneven (Portugal anecdote at 00:06:37 to 00:06:47). Design for real world friction, not only for headlines.

Guideline: build workflows that work even when inputs are imperfect, data is partial, and human review is mandatory.

The Practical Stack

1. LLMs: ChatGPT, Gemini, or your model of choice for generation and critique.
2. Deep research: a research agent that can summarize sources and cite them for human verification.
3. Knowledge base: process docs and checklists in a shared space.
4. Issue tracker: every AI assisted task maps to a ticket with acceptance criteria and the audit log.

Metrics That Matter

1. Time to deliver a repeatable task target 15 minutes for the first shippable draft.
2. Verified error rate decline week over week.
3. Pipeline lift from research more replies per 100 emails, more meetings per 100 replies.
4. Content usage rate assets referenced in live deals per week.
5. Client confidence measured via quarterly survey or renewal notes.

Swipe File: Prompts and Checklists You Can Steal

Research to outreach line

1. Inputs: company, role, trigger signals, public sources
2. Ask: “List 3 concrete reasons this account is high fit. Cite sources. Draft 2 opening lines tied to those reasons. Keep to 40 words.”

Weekly report compression

1. Inputs: analytics links, top 3 goals, change log
2. Ask: “Summarize movement against goals, attribute changes to logged actions, flag anomalies that need human review, and propose 3 experiments. Keep to one page.”

Acceptance checklist for any AI output

1. Factual claims have sources.
2. Brand voice matches example.
3. No unverified numbers.
4. Compliance terms approved.
5. Reviewer initials and date.

Final Word

Winning with AI is not about flashy demos. It is systematic time compression, research that targets the right people, and honest communication that earns trust. Start with one workflow, one checklist, and one measurable win. Then scale it across the team.

Watch the episode and skip to these moments:

1. Hours to minutes claim 00:00:33 to 00:00:36
2. Deep research for lead sourcing 00:03:59 and 00:04:13
3. Move content down funnel 00:05:06 to 00:05:15
4. Ranking on LLMs and AI overviews 00:07:10 to 00:07:14
5. Client AI fears and education 00:09:26 to 00:10:31