🤖 How top leaders use AI SDRs

In partnership with

Where sales teams win next year

2026 will reward teams that understand agentic systems.

This event shows what that looks like in practice.

→ AI handling sourcing and follow-up
→ Systems driving execution
→ Leaders rebuilding orgs around coordination

Quick, tactical, and free:

⇢ Reserve your spot

What six months of AI SDRs actually looks like

Jason Lemkin breaks down what happens when you deploy AI SDRs at scale, based on six months of real data and over $1M in closed revenue.

AI SDRs amplify what already works

They do not fix broken processes.

If your messaging fails with humans, AI will scale that failure.

Before deploying AI, you need:

• Proven messaging that already converts

• Clear ICP definition

• Clean RevOps processes

• Training based on real conversations that worked

↳ The AI can only repeat your best practices infinitely. You have to give it best practices first.

Human oversight is not optional

Expect 15-20 hours weekly managing these agents.

Performance rises and falls with human attention. Weeks with less oversight produce noticeably weaker results.

✔ Daily monitoring of outputs
✔ Weekly training updates
✔ Constant refinement of messaging
✔ Regular review of conversations and responses

↳ This is coaching five SDRs who never sleep. Not automation you can forget about.

Specialization beats all-in-one

Running five different AI agents trained for specific use cases outperforms one general tool.

Each agent gets completely different training:

• Cold outbound to new prospects

• Lapsed customer re-engagement

• Active nurture for email openers who haven't converted

• Inbound qualification on website

• Ghosted lead recovery

↳ Specialized tools go deeper. Three A+ tools beat one B+ tool.

The outbound numbers

6.7% overall response rate. Industry average sits around 3%.

The AI sends more emails in one month than a human SDR sends in 40 months. With better response rates.

But results required massive human input. Sub-agents within outbound each had different training:

⇢ Lapsed sponsors: Reference past relationship
⇢ Current sponsors: Cross-sell to additional events
⇢ Previous attendees: Offer early access
⇢ Engaged non-converters: Follow up on email opens
⇢ Pure cold: Personalize based on company signals

Inbound AI changes meeting prep

Before each meeting, the AI provides complete context:

→ Full conversation history
→ Every page the prospect visited
→ How many times they engaged
→ Other people from their company who visited
→ Specific content they consumed

No more discovery calls. The AI already did discovery. Jump straight to solution discussions.

↳ In one month, 70% of closed revenue came through the AI SDR.

Direct selling on lower-ticket items

For products under $1,000, the AI can close deals without human involvement.

For higher ASP deals, it qualifies and books meetings. Then hands to humans.

The workflow that emerged:

1. Prospect asks for discount

2. AI gives discount code

3. Prospect leaves without buying

4. Day 3: AI follows up about unused code

5. Prospect converts

↳ 20% of ticket revenue now comes from AI agents selling directly.

Deliverability matters more than message

A 2-3 week warm-up period before sending at volume keeps emails hitting primary inbox.

Skip this step and your best message lands in spam.

Feed the AI fresh contacts

Upload new contact lists twice per week when possible.

About 90% of contacts should come from your own database. Not scraped from intent data providers.

↳ The AI performs better with fresh, trusted data.

The deployment playbook

Start with one agent and one use case. Get it working before adding a second.

Limit vendor selection to two maximum. Training ten vendors simultaneously means none get trained properly.

Begin in draft mode:

• AI suggests messages

• You approve and send

• You correct and train constantly

After 30-60 days, start empowering for low-stakes interactions.

The common mistakes

✘ Expecting magic without putting in the work

✘ Deploying AI to fix what humans couldn't fix

✘ Generic training instead of specific proof points

✘ Set and forget instead of constant iteration

✘ Ignoring vendor expertise on platform best practices

👉 Power up your sales team

If you lead a sales team, run training sessions, or coach reps - this is for you.

With SalesDaily Elite, you get full white-label rights to my complete library of sales cheat sheets, guides, and templates.

✔️ Editable Canva files you can brand as your own

✔️ Ready-to-use for team onboarding, and workshops

✔️ Built for high-impact sales team enablement

Stop wasting time building resources from scratch.

➔ Upgrade to SalesDaily Elite

Why copying Predictable Revenue failed

Aaron Ross, author of Predictable Revenue, reveals why most companies failed when copying his framework and what the future of B2B sales actually requires.

The framework got misunderstood

Companies copied tactics. SDR teams. Email cadences. Meeting quotas.

They bolted these onto commercial systems that were never designed to support them.

Marketing doesn't know what sales is learning. Customer success has no idea what was promised in the deal. Information dies in functional handoffs.

↳ Even Aaron admits the market got it wrong. The book wasn't really about outbound prospecting.

Tactics without architecture creates activity, not outcomes

You can nail the sales execution and still plateau.

The commercial system underneath is the problem. When SDRs learn what messaging works, does that intelligence reach marketing? Or does it depend on someone remembering to mention it in a meeting?

When sales closes deals with specific buyer profiles, does outbound adjust targeting? Or do they keep prospecting the old ICP because intelligence never flows back?

↳ No amount of better SDR coaching fixes structural dysfunction.

What happens when information gets trapped

SDRs prospect. They learn what messaging lands. Which verticals buy. What objections surface.

That's market validation happening in real time.

In most companies, that intelligence dies in the sales silo.

→ Marketing keeps running the same campaigns → The next SDR hired learns the same lessons from scratch → CS has no idea what was promised during prospecting

The metric that changes everything

Old approach: Measure meetings booked.

New approach: Measure market intelligence captured.

When SDRs get measured on meetings, they push conversations toward artificial urgency. When they get measured on intelligence, they have permission to be curious instead of aggressive.

Joey Gilkey runs SDR teams where the North Star is market intelligence captured, not meetings booked.

His SDRs catalogue after every call:

• What messaging landed

• Where the prospect is in their buying process

• What objections surfaced

• Which competitors they're evaluating

↳ That intelligence flows back to AEs for deal strategy. It flows to marketing for campaign adjustments. The system gets smarter, not just busier.

Sales as a communication channel

Most companies see sales as closers. They miss the bigger opportunity.

Sales teams have live conversations with buyers every day. They're learning what works. What objections matter. What timing looks like. Which competitors are in play.

That intelligence is more valuable than any intent platform or survey.

But it gets trapped in individual reps' heads because nobody measures it.

Aaron's take:

⇢ You can't learn what surveys, polls, and market research can't capture ⇢ Live conversations produce insights automation never will ⇢ Companies do not take advantage of this

Human-led beats automation

After fifteen years watching companies try to automate their way to predictable revenue, Aaron's advice: have more conversations.

Not sales conversations. Validation conversations.

✔ Learning what buyers actually care about ✔ Understanding timing without forcing it ✔ Capturing competitive intel without pitching

The future isn't choosing between human and automation. It's humans doing what only humans can do, supported by systems that distribute what those humans learn.

What's actually durable

Tactics change. Tools change. What survives?

Relationships. Creativity. Intuition.

Aaron referenced Colombo, the TV detective who acted dumb and asked curious questions instead of coming in hard. That approach builds relationships. Disarming curiosity. Operating on the buyer's timeline.

↳ The predictable revenue framework got weaponized into aggressive prospecting. What actually works is the opposite.

The diagnostic questions

Before implementing any GTM framework, assess whether your commercial architecture can support it.

Ask:

• When SDRs learn what messaging works, does that intelligence systematically reach marketing?

• When sales closes deals, does outbound adjust targeting based on what's actually working?

• When CS identifies which customers expand or churn, do SDRs prioritize account types that match expansion patterns?

If the answer is no systematic flow, you have an architecture problem. Not an execution problem.

Partnering with:

• Navan: The leading travel and expense platform

• Smartlead: Automate outreach without losing the human touch

• Close: Manage relationships and scale outreach in one smart CRM

• La Growth Machine: Multichannel prospecting that actually scales

(include affiliate links)

The 8 stories I tell every sales leader

Mark Roberge compiles lessons from scaling HubSpot's sales team and teaching at Harvard Business School.

Walk in the buyer's shoes

HubSpot's sales training had new reps start their own blogs. They wrote 15-20 posts. Set up landing pages. Experienced exactly what they'd ask prospects to do.

When a buyer said "what if my writing is bad?" the rep could say "I just went through that."

The 30-page diagnosis

One rep had a territory of two accounts. His job: convert Target from Oracle to Salesforce.

He didn't call them for two months.

Instead, he shopped at Target every day. Screenshotted every email. Built a 30-page diagnosis of their marketing strategy. Mailed 25 copies to every digital marketing executive.

Within three days, meetings with all of them.

There is no universal amazing salesperson

Roberge hired the #1 rep from an 800-person team. Six months later, middle of the road.

That rep came from a 20-year-old company running Super Bowl ads. Prospects knew what was coming.

HubSpot was year one. "What is inbound marketing?"

↳ There's only optimal for your context.

Comp plans at product-market fit stage

Founders copy enterprise comp structures. Commission based on revenue.

Completely misaligned.

At product-market fit stage, you need 20 very healthy customers. Not maximum revenue.

Commission on revenue incentivizes the opposite behavior.

Expansion without experimentation

Q4. Founder does the math. ICP only gets them to 7M, need 9M.

So they commit to the board: "We'll launch Europe and get 2M from there."

They've never sold a single deal in Europe.

The fix: Run 10% experiments on expansion while scaling 90% into known ICP. Put a rep in Europe a year before you need the revenue.

TO-GO

Matt Green: The 30-day pipeline experiment

Marcus Chan: I used to give feedback all wrong

Mike Groeneveld: Kickoffs don’t build culture

Dan Richards: Run a great sales interview process