How this track works
Six sessions, and every one of them ends with a decision you have to actually make rather than a slide you have to nod at. a1 is the honest picture of what changed and what did not. a2 is the evidence, so you can read a vendor deck without being read by it. a3 is the forecast, and how to keep it believable when a model starts disagreeing with your reps. a4 is enablement, coaching and ramp on a searchable record. a5 is the compliance rails, because volume that used to take a week now takes a minute. a6 is rollout, adoption and comp, where all of it either lands or quietly does not.
There is a parallel rep track running b1 to b10, built on one running territory: an AE at Cadence carrying eight named accounts, anchored on Northwind Logistics. You do not need to sit through it. You should skim b1 (the deal loop, the trust line, the five levers of an email) and b3 (what your team is actually taught to send), because a leader who has not seen the material their reps were given will make policy for a job that no longer exists.
The productivity paradox 7 min live
Here is the finding that should reorganise your quarter. Gartner reports AI now gives sellers back nearly five hours a week, and that 72% of sales organisations fail to reinvest that time in high-value activities. The hours arrive automatically. They then evaporate into more tabs, more admin, and more sends to the same tired list, and at the end of the quarter the number looks exactly like it did before you bought anything. Nobody decided that. It is what happens when time is returned and no one is accountable for spending it.
LiveThe four numbers that should shape your operating plan4 min▶
Four findings, all from Gartner, all pointing the same direction: the licence is not the intervention.
| Finding | What it means on Monday |
|---|---|
| 3.7x quota attainment | Sellers who effectively partner with AI, from a survey of 1,026 B2B sellers. "Partner" is a working habit, not a login. You cannot buy it, you can only coach it. |
| ~5 hours a week returned | The time genuinely comes back. It is real, it is measurable, and it is the only budget you got for free this year. |
| 72% reinvest none of it | The default outcome. If nobody owns where the hours go, they go into activity that looks like work and produces nothing. |
| 2.6x and 2.4x | Organisations giving sellers AI-enabled next best actions are 2.6x more likely to achieve commercial growth. Those emphasising upskilling on AI are 2.4x more likely to see good revenue growth. |
Read those last two together and the shape of the intervention is obvious. Growth correlates with what you tell the seller to do next and whether you taught them how. It does not correlate with how many seats you bought.
The five hours are already gone somewhere. Pick any rep on your team and ask what they stopped doing this year. Most cannot answer, because nothing stopped - the drafting got faster and the freed minutes went straight back into sending more of the same. That is the 72% in one conversation, and you can run it this afternoon for the cost of a coffee.
Self-studyWhy "adoption" is the wrong metric to celebrate3 min read▶
Adoption dashboards measure whether a tool was opened. That was a useful question in 2023 and it is a nearly useless one now, because usage is close to universal and tells you nothing about whether the job got better.
- Seats used tells you procurement worked. It does not tell you a single deal moved.
- Drafts generated tells you your team can type a prompt. Volume of drafts is an input that gets cheaper every quarter, which makes it a terrible metric to reward.
- Hours saved is real but incomplete on its own. The honest version of the metric is hours saved and redeployed into a named activity. Without the second half you are measuring the 72%.
The metric that survives is the one a rep cannot game by doing more of something cheap. a6 goes at this properly, including what happens to these numbers the moment you attach them to comp. For now, notice how many of your current AI slides answer "did they use it?" rather than "did it change what they did?"
Three places AI destroys trust 6 min live
Buyers did not sign up for any of this, and they have adapted faster than most sales orgs have. Platform-average cold reply rates fell from about 5.1% to about 3.4% in two years while send volume went vertical, and Gartner predicts that by 2030, 75% of B2B buyers will prefer sales experiences that prioritise human interaction over AI. That is not nostalgia. It is a rational response to an inbox where the default assumption is that nobody read anything. Three parts of selling carry the whole relationship, and automating any of them costs you more than it saves.
LiveThe read, the promise, the price4 min▶
Your reps learn this in b1 as the trust line. It is worth having the leader version, because two of the three are things only you can authorise them to protect.
- 1. The read. Outreach that proves nobody looked. A merge field is not a read - it is a machine filling a slot, and the day the field is empty your buyer sees exactly what she is to you. The read costs four to six minutes per prospect and it is the single largest lever on reply rate. If your team's activity targets make four minutes per prospect impossible, you have already decided this one and you decided it wrong.
- 2. The promise. Anything auto-drafted that commits your company: a scope line in a proposal, a timeline in a follow-up, a "yes we support that" in a reply. A model will happily generate a commitment nobody in your company has agreed to deliver, in fluent, confident prose, at 11pm. This is a policy question, not a taste question.
- 3. The price. A number no human owns. The moment a buyer senses the price came out of a machine, every discount conversation becomes a negotiation with a machine, and your rep has nothing left to trade.
Notice what is not on this list: research, list building, note-taking, transcript search, first drafts, CRM field population. That is a lot of the week, and it is genuinely yours to automate hard.
Self-studyWhy the reply rate fell and what that costs you3 min read▶
Sending got free. That is the whole mechanism, and it has a predictable consequence: when the cost of producing a message collapses, the scarce thing stops being the message and starts being the attention.
- Average is falling. Platform-average B2B cold reply rates sit around 3.4%, down from roughly 5.1% two years earlier, across published 2025-2026 benchmark reports.
- Top quartile held. The best senders still land 15-25%. The spread between average and good got wider, not narrower, which is the most useful fact in this session: the market rewarded discipline and punished volume.
- Your domain is an asset with a balance. Every send to a bad-fit row draws it down. A team hitting activity targets by spraying is spending a capital account nobody put on your P&L.
Treat these as the shape of the curve rather than gospel - vendor samples skew heavily to their own users, which is exactly the reading skill a2 is built around.
What this actually puts on your desk 20 min · four cards
Everything above is context. These four cards are the part that changes what you do this quarter: what a rep's week genuinely looks like now, what has not moved an inch, how to read the agent forecast without either panicking or dismissing it, and the six questions that make a vendor conversation short.
LiveWhat actually changed for a rep's week4 min▶
The change is not "reps write emails faster". It is that the preparation half of every stage collapsed in cost while the judgment half did not move at all. Your team's week now looks like this:
| Stage | Collapsed in cost | Unchanged, still theirs |
|---|---|---|
| Target | Enriching, ranking and deduping a list | Killing the bad-fit rows on purpose |
| Reach | Drafting, tightening, checking for tells | Finding the trigger event and choosing the ask |
| Discover | Account briefs, question banks, prep | Asking, listening, following the thread |
| Advance | Recaps, first drafts, chasing the chase | The promise and the price |
| Record | Notes, fields, next steps captured | The honest call on whether the deal is real |
That is the deal loop from b1. Read the right-hand column again: none of it got cheaper, and all of it got more valuable, because it is now the only part that is scarce. Your enablement budget belongs in that column.
LiveWhat did not change at all3 min▶
Short card, and the one worth repeating to your board.
- Buyers still buy from people. Gartner's 2030 prediction is that 75% of B2B buyers will actively prefer sales experiences that prioritise human interaction over AI. The preference is strengthening, not weakening, and it strengthens fastest in exactly the segments that write the biggest cheques.
- Trust is still built at human speed. You cannot parallelise it, and a machine cannot hold it on your behalf.
- Targeting still beats copy. A perfect email to the wrong seat is still a no. AI made it trivially cheap to send that no at scale, which is why the discipline matters more now than when a list took a week to build.
- Someone still has to decide. A model can rank your pipeline. It cannot tell your CFO which deals you are standing behind, and it will never carry the consequence of being wrong.
Self-studyThe agent forecast, read honestly4 min read▶
Gartner predicts AI agents will outnumber sellers 10 to 1 by 2028, and in the same breath that fewer than 40% of sellers will say agents improved their productivity. Both halves matter, and the second half is the one your vendors will not put on the slide.
- Plan for arrival, not for benefit. The agents are coming into your stack whether or not they help, because they will arrive bundled inside tools you already pay for. Budget for the integration and the governance work regardless of the productivity case.
- Do not headcount-plan off the ratio. A 10:1 agent-to-seller ratio is a statement about software instances, not about work replaced. Anyone converting that ratio into a reduction plan is doing arithmetic on a metaphor.
- Under 40% is a design signal. If most sellers will not feel a gain, the gain is not automatic, which puts you straight back to the 2.6x and 2.4x findings: next best actions and upskilling are the difference between an agent that helps and one that adds a tab.
- Governance is the constraint that binds. An agent that can send on your behalf sits on the wrong side of the trust line by default. a5 covers what has to be true before that is allowed.
The honest leader position: assume the tools land, assume the average seller feels nothing, and make your money on being in the minority that designed for it.
Self-studySix questions to ask your vendor3 min read▶
Paste these into the next demo invitation. Sending them in advance is the point: the answers you get back tell you as much as the demo does, and a vendor who cannot answer question 4 in writing has told you everything.
1. Which specific task does this remove from my rep's week, and how many minutes per rep per week, measured how? 2. Can it send, commit or price anything without a human clicking approve? Show me where that switch lives and who can flip it. 3. Where does our customer data go, who can see it, and what is your retention and deletion policy in writing? 4. Your headline number: who was measured, against what baseline, over how long, and who paid for the study? 5. What does the record look like if we leave? Export format, and does the call and note history come with us? 6. What happens on day 90 if adoption is 30%? Name the intervention you run, not the dashboard you show me.
Question 1 separates a product from a feature. Question 2 is the trust line as a procurement clause. Questions 3 and 5 are the ones your legal and RevOps teams will otherwise discover in month seven. Question 4 is the whole of a2 compressed into a sentence, and question 6 is the only one that predicts renewal.
Before session 2 ◐ 40-50 min total
- Audit the hours. Take last quarter and write down where your team's saved time actually went. Not where you hoped - where it went. If the honest answer is "more sends", you have found your 72% and you have also found your first intervention.
- Write your trust line. Three to five sentences, in your words, naming what a buyer of yours must believe came from a person. Then name one task your team is doing today that sits on the wrong side of it, and decide this week whether you are moving the task or moving the line.
- Ask three reps. Separately, ask each one what AI does for them today. Compare their three answers to each other, and then to what you believed before you asked. The gap between those two comparisons is your enablement plan for a4.
- Skim the rep material. Read b1 and b3 so you have seen what your team is being taught to send. Fifteen minutes, and it will change at least one policy you were about to write.
- Bring one vendor deck to a2. Any deck currently being sold to you, with the headline number still in it. Session 2 takes it apart.
Official sources covered
The leader track teaches the working core of the current sales-AI evidence base: Gartner's seller and buyer research, published cold-email benchmark reports, and the compliance regimes that actually govern outreach. Vendor certifications stay with the vendors. This page covers:
Three questions before you go 🎯 ◐ 90 seconds
1 · AI returns nearly five hours a week to a seller. What does Gartner find happens to that time in most organisations?
The hours come back automatically and then evaporate into more tabs, more admin and more sends to the same list. The findings that do correlate with growth are the deliberate ones: 2.6x for AI-enabled next best actions and 2.4x for emphasising upskilling. The tool is not the intervention, the habit is.
2 · Which three things must a buyer believe came from a person?
The read proves someone looked before writing. The promise is anything that commits your company. The price is a number a human owns. Everything else - list building, briefs, note capture, first drafts, CRM fields - is genuinely yours to automate hard, and that is most of the week.
3 · Gartner predicts agents will outnumber sellers 10:1 by 2028 while fewer than 40% of sellers report a productivity gain. What should a leader plan for?
A 10:1 ratio counts software instances, not work replaced, so converting it into a headcount plan is arithmetic on a metaphor. The under-40% figure is the real signal: the gain is not automatic. Plan for arrival, then earn the benefit through next best actions and upskilling, and keep agents on the correct side of the trust line.