🤖 run an AI-first sales team

SalesDaily for Leaders
November 15, 2026

This week: AI in your sales playbook, your hiring tests and your call coaching.

In this issue:

• Garrett Marker: Build AI around what your best sellers know

• Ron Gabrisko: Have AE candidates build a demo in the interview

• Kevin Dorsey: Define what good looks like before AI scores a call

• Recommended reading: Four posts from sales leaders

Check your 2027 plan against B2B's top CROs

On December 8 to 10 at 9 AM PT (6 PM CET), HSE is running its free, virtual Agentic Strategy Summit. CROs from ZoomInfo, ElevenLabs, Carta and Vercel are among 100+ speakers taking on questions like:

• How do you run a team of sellers and AI agents?
• Who's accountable when an agent goes rogue?
• What deserves a place in your 2027 budget?

STRATEGY

Build AI around what your best sellers know

Garrett Marker built an internal AI system on the judgment of his top sellers and put it inside the tools his team already used. You get how it works, what he builds versus buys, and where the extra productivity goes.

Start with what top sellers understand

On any sales dashboard, a few sellers bring in 10 times the revenue of the average one. Garrett has spent years asking how to turn average sellers into that group.

Data and tools weren't his blocker. The CROs he talks to don't see their sellers getting much more productive from new go-to-market tools.

In his view, top sellers understand the buyer's journey, and playbooks try to translate that understanding for everyone else. Generic scripts lose it on the way.

↳ The result is bad discovery calls, weak demos and closed-lost deals.

Keep one narrative per account

His team built what it calls the narrative platform. Each account has one narrative, updated from every customer signal and open to every go-to-market function.

→ Facts with sources: AI turns scattered account data into facts sellers can look up. Each one carries a citation, so they can check where it came from.

→ Top sellers in the loop: they train the system and confirm it reads the buyer the way they do.

→ Patterns plus reasons: the system looks across past sales cycles for what worked, and the sellers who won explain why.

↳ Before a demo, it can point to what is likely to land with this buyer, based on their pain points and similar deals.

Put it where sellers already work

The narrative shows up in the chat and AI tools the team already uses, so adopting a new tool is never the hurdle.

Every app syncs both ways. Each interaction writes back to the central model, and sellers enter nothing twice.

Alerts fire on context:

⇢ The CFO you're about to call just opened a support ticket, so today is a bad day for an upsell.

⇢ A feature a buyer asked for on someone else's call shipped yesterday, so bring it up on your next one.

What his company reports so far

→ Sellers who use the platform are 5x more productive.
→ Cost to serve across the customer base is down more than 20%.
→ A new seller sourced a $15 million opportunity while still ramping.

↳ These are his company's own numbers. He's in his third fiscal year there, and he says revenue growth accelerated each year while acquisition cost and cost to serve fell.

Onboard with practice nobody has to wait for

In the past 12 months, his company hired and onboarded more than 250 sellers across 10 teams. The enablement team did it without growing.

Sellers learn by doing, he says, so onboarding is built out from role-play practice. The practice tools were trained on the same data as the narrative platform.

New hires no longer wait for:

→ Product marketing to write a battle card
→ Marketing to build a deck
→ A manager to free up for a role play or to score a call

↳ By his company's numbers, last year's hires were twice as successful as earlier ones.

Build what's specific to you and buy the rest

Buy when the job is generic. His example is call recording.

Build where the way you show up in a customer conversation is specific to your company. At every company he has worked for, he can name two or three sales plays that changed results. Once the team ran them, win rates moved by 20%, 40% or 60%.

Failed AI projects had the wrong people in the room. Nobody could connect product and go-to-market, so the tools automated the wrong work.

↳ His team now ships fewer AI tools, he's closer to each one, and the results are better.

✘ Building whatever the closed-lost data asks for. In his experience, that data is largely sellers explaining away poor discovery or a bad demo.

Use the gains for promotions before quotas

Once quotas reached the point where the unit economics worked, Garrett largely stopped raising them. He wants sellers to use the extra capacity to prove they can do a bigger job.

→ An SDR who beats target by a set percentage three months in a row gets an AE's book of business in month four.

→ If they close what a ramping AE would while still hitting the SDR number, they get the job. Open headcount doesn't matter.

→ The same logic moves AEs from small deals to large ones.

Sellers don't interview for the next job. Overperform in your role, complete a set list of activities, and you move up.

Before you buy another AI tool

✔ List the two or three plays that move your win rate.
✔ Ask your best sellers why those plays work, and write it down.
✔ Put account context in the tools your team already uses.
✔ Buy the generic tools and build only what's specific to you.
✔ Decide where extra capacity goes before you raise a quota.

HIRING

Have AE candidates build a demo in the interview

Ron Gabrisko tests AE candidates on the work they'll do with AI, then certifies them again and again after they join. In this conversation with Sam Blond, he also makes the case for keeping a human SDR team.

Keep the fundamentals and add one requirement

Ron's list of what still matters: preparing well for customers, building relationships, researching accounts, grit, EQ and IQ.

The new requirement is being AI-first. His AEs use the company's own AI product on their phones and build demos with it themselves.

They also have to hold a CEO-level conversation. AI has become a CEO-level initiative, and his buyers now include CFOs and CMOs.

↳ Slides alone no longer carry a meeting. Buyers expect to see what you'll deliver, built on sample data that looks like their industry.

↳ Sam sees a new group at the top of his own leaderboard. These sellers are AI-curious and can walk a buyer through how the product works under the hood.

Run the interview as a live build

→ Pick an account. The candidate chooses a company.

→ Pitch the CEO. They give a CEO-level AI pitch for that company, specific to its industry.

→ Build the demo. They get access to the product, build a demo on the spot and walk the panel through it and the outcomes.

Ron's reason: the interview tests, in real time, the qualities he wants in front of customers.

Back-channel references and hiring from known networks still count. Every candidate also meets the team in person.

Certify again after the hire

New hires start with boot camps on the full product range. Certification follows and continues through the year:

✔ The CEO-level pitch, because he wants storytellers who can paint what's possible
✔ Industry-based demos built in the company's AI product

A former baseball player, Ron expects sellers to practice 10 times more than they play.

↳ Sam's takeaway for his own team: a technical test in onboarding that sellers pass before they run a demo.

Keep the SDR team as your AE academy

Ron still has a large SDR team. Its headcount has stayed roughly flat while he kept adding AEs and solution architects, because output per SDR went up.

Outbound is now automated and personalized down to the person. By his account, response rates rose from 1% or 2% with mass emails to between 6% and 10%.

He won't take the team to zero, though it may get smaller over time. Many of his SDRs graduate into AE roles, and without them nobody teaches new sellers how to sell the product.

↳ Go lean there, and you either hire every AE from outside or find another way to train and mentor the next generation.

↳ Some of his best sellers came up through those ranks. They grew up with AI and write their own apps.

⇢ Sam runs the opposite setup at his two-year-old company: one SDR books about 400 outbound meetings a month through automation.

Plan for 60% of seller time with customers

Ron pushes for at least 60% of seller time in front of customers, in person where possible.

→ Monday and Friday: preparation and admin
→ Tuesday to Thursday: with customers, if not more

Agents now handle account research, presentation prep, executive briefing docs and proposals. The time saved goes back to customers.

↳ Reaching new departments and stakeholders stays the AE's job. He expects each AE to act as the CEO of the account.

Five changes before your next AE hire

✔ Add a live build to the interview: one account, a CEO pitch and a demo.
✔ Meet every finalist in person.
✔ Certify new hires on the pitch and the demo, then repeat during the year.
✔ Set a technical test sellers pass before their first demo.
✔ Name who will train your next AEs before you shrink the SDR team.

Cheat sheet of the week

COACHING

Define what good looks like before AI scores a call

Kevin Dorsey starts every AI project by working out what his best sellers do differently. From there he builds call scorecards, diagnosis checklists and one metric per seller for AI to track.

Tell AI what to do

Kevin sorts leaders into three groups: those who deny AI, those who depend on it and those who direct it. Dependers ask AI how to do their job.

His rule: if you can't do something well by hand, don't use AI to do it. You can't judge an answer when you don't know what a good one looks like.

Quota attainment keeps sliding while sellers have more tools than ever. With AI, the first 80% of a task is fast, and he sees too many teams ship at 80%.

↳ The last 20% still takes time and skill.

Study your best sellers first

He calls the starting point WGLL, short for what good looks like.

Upload 40 calls from one seller and 40 from another. Then ask what the first does differently in discovery, objections, follow-up and pricing.

At his last company, he gave his SDR leaders one day to analyze more than 4,000 calls. They learned which openers worked, how the best handled objections and how they asked for the meeting.

➤ By his account, the conversation-to-meeting rate doubled within 60 days.

↳ Look for patterns by persona. Problems, objections and buying steps repeat more than you'd expect.

↳ His agents also collect what each persona says in online reviews and forums. That language goes into emails, ads and the pitch deck.

Do the work in six steps

→ Diagnose what works and why.
→ Define what good looks like for that skill.
→ Document it in writing.
→ Demonstrate it with recorded calls or demos.
→ Drill it. AI role play removes the old limit, the manager's time.
→ Deploy it with AI.

↳ Teams that start with AI and wonder why it fails have the order backwards.

Write scorecards AI can score

A scorecard holds the 7 to 10 things in a seller's control that raise the odds of success on that type of call. For each one, write what it is and add examples, so AI knows what to look for.

Every item has to be a behavior:

✘ Build trust
✔ Paused before answering

✘ Understand pain
✔ Asked three problem questions and put a dollar value on the problem

Build one scorecard each for discovery, demo, cold call and pricing call. Sellers should read it before and during the call, like the answers to a test.

↳ With every call scored, Kevin sees skills move before results do. Rising call scores tell him the close rate will be up 30 days later.

↳ He sums up the levers behind any metric as BIPSY: behavior, individual, process, skill and you.

Diagnose the metric before you coach

Telling a team to raise its show rate changes nothing. Kevin keeps a checklist for each metric, from meetings set and deal size to connect rate and email response rate.

For a low show rate, the checklist asks:

→ Do reminders go out, and are they any good?
→ How far out is the meeting booked?
→ Did the first call find a problem?
→ Do the reminders mention that problem?

Connected to the CRM and call recordings, an agent can then say why one seller's show rate is low. The answer goes to the seller and the manager.

It becomes a coaching plan and an AI role play for that one skill. Then it comes back in the next 1:1.

↳ One of his checklists covers the person: sleep, breaks, recent changes in behavior and the last growth conversation.

↳ His point about managers: too few coach enough, and even more coach the wrong thing.

Pick one metric per seller

The leader's job is to name the number one metric for each seller, each team and the org. Chase every metric at once, and none of them moves.

Ask which metric has the biggest effect on revenue. His example: a 27% close rate you'd like at 35%, next to a 48% show rate. The show rate is the bigger lever.

AI then tracks the trend, splits the metric by channel and looks for causes.

➤ His last tip: talk to AI instead of typing. You describe the problem in far more detail when you speak.

Start with these five

✔ Have AI compare calls from your best seller with calls from an average one.
✔ Write one scorecard with 7 to 10 behaviors and examples.
✔ Build a diagnosis checklist for your weakest metric.
✔ Name one metric per seller and the skill behind it.
✔ Set up an AI role play for that skill and review it in the 1:1.

Recommended reading

Keith M. Laughner: Give your AI agents an org chart

Keith runs his AI bots like a team. A lead bot routes each request to a specialist with one defined job, every bot keeps its own memory under shared writing rules, and anything sent under his name comes back to him first.

Mike Groeneveld: Use AI to save deals before the forecast call

Flagging a stalled deal only makes the forecast more accurate. Have AI help sellers act while there is still time: where engagement dropped, who to add to a single-threaded deal and what really stalled a deal after pricing.

Rory Sadler: Train for the AE job AI has created

Buyers now arrive with AI-made vendor comparisons, CTOs ask why they can’t build it themselves, and finance and security question every AI claim. Before you answer a missed target with more calls, check whether training, support and metrics changed as much as the job.

Todd Busler: Keep backing the seller who works the basics daily

Todd’s team runs several prospecting agents, yet his Q3 leader started the quarter with 6x pipeline coverage and finished above $1M. That seller cold calls, works LinkedIn and asks for intros every day.

Quote of the week

Humor

If a fellow sales leader would get value from this issue, forward it to them.

Talk soon,
Haris