§ Playbook · Commercial teams
The AI playbook for Commercial teams
Sales has adopted AI faster than almost anyone and fired it faster too - AI SDR tools churn at twice the rate of the humans they replaced. This playbook covers where the commercial value is real (mostly not where the demos said), what happens to SDR and revenue-ops roles, and how to reprice a CRM stack whose own vendor keeps changing its mind.
Free · no form, no follow-up sequence
Companies churn their AI SDR tools at 50-70% a year - roughly double the turnover of human reps. Meanwhile Walmart's negotiation bot quietly closes deals with 68% of the suppliers it approaches. The difference is not the technology. It is choosing work machines are actually good at.
50-70%
annual churn on AI SDR tools - double the human reps they replaced.
B2B buyer analyses, 2026
36%
of B2B software firms cut SDR headcount in 2025 - the steepest of any sales role.
Emergence Capital, 2025
3% / 68%
Walmart's negotiation bot: average cost gain, and share of approached suppliers who closed.
Bloomberg / Pactum
§ Part one
What actually changes.
Commercial AI has run the full hype cycle faster than any other function. Four hundred million dollars of venture money went into AI SDR startups; buyers now churn those tools at 50-70% a year, and the companies that went full-replacement had mostly reverted to hybrid by early 2026. Autonomous selling converts meetings to opportunities at roughly 15% against a human's 25%. Customers, it turns out, notice.
What actually works is quieter. Walmart's negotiation bot handles tail suppliers no buyer had time for: 3% average cost improvement, payment terms extended, and 83% of suppliers rating it easy to deal with. Machines on the long tail; humans on the accounts that matter. That is the whole shape of it.
01
Prep and admin time comes back
Research, call summaries, CRM hygiene, forecast roll-ups - the evidenced win is hours, not headcount. HubSpot puts it at 1-5 hours a week per rep. Daily AI users are twice as likely to hit quota, though the causation is doing some work in that sentence.
02
The long tail gets negotiated at last
Tail suppliers, small renewals, standard terms - deals too small for human attention now get worked instead of waved through. That is found money, and it does not threaten a single key account.
03
Pricing gets an analyst, slowly
Only 5-10% of organisations have scaled any agentic pricing use case, per McKinsey. Where it works today: price recommendations and discount-leakage detection with a human deciding. The autonomous pricing engine remains a conference slide.
04
Forecasting stops being folklore
Roll-ups assembled by machine from actual pipeline data, with humans arguing about the judgement calls rather than the spreadsheet. Modest, unglamorous, and worth more than most of the tooling budget.
05
Full-cycle selling returns
As machines take top-of-funnel drudgery, AEs increasingly own the whole motion. The SDR-to-AE production line - itself only a couple of decades old - is quietly de-industrialising.
06
Autonomous outbound stays a trap
Volume outreach with machine-written sincerity damages the brand with every send. The maths is documented: more activity, fewer opportunities, and a market that has learned to smell it.
§ Part two - the honest bit
What it costs you.
The role that changes most is the SDR. Emergence Capital found 36% of B2B software companies cut SDR headcount in 2025 - the steepest reduction of any sales role, mostly through attrition. The role does not vanish; it becomes supervising machine-generated pipeline and rescuing the conversations worth a human. Fewer people, different work, and a recruitment pitch that needs rewriting.
Then there is the stack. Salesforce shipped three different Agentforce pricing models in eighteen months - per conversation, per action, per user - which tells you the biggest CRM vendor on earth does not know what an AI agent is worth either. Every commercial tool renewal you sign this year is a bet on a pricing model its own vendor may abandon.
01
SDR teams shrink and change shape
The cut is real and already happened at a third of B2B software firms. What remains is pipeline quality control - a more senior job than the one it replaces, which your comp plan and career ladder have not noticed yet.
02
RevOps becomes the load-bearing wall
Someone must configure, monitor and QA the agents, own the data they run on, and referee the pipeline they generate. That is RevOps now - promoted from spreadsheet janitor to operating system, without anyone updating the org chart.
03
Your CRM data is the constraint
Agents amplify whatever is in the system. If your CRM is a graveyard of stale contacts and optimistic close dates, AI will now generate confident nonsense from it at scale. Data hygiene is a commercial capability, not an admin chore.
04
Comp plans need rebuilding
When a machine sources the meeting, who gets paid for it? Activity-based comp collapses when activity is free. Rebuild around outcomes before your best people find the loophole - they are, after all, professionally incentivised to.
05
Seat economics wobble under you
Per-seat CRM pricing shrinks its own market as agents replace seats; Gartner expects 40% of enterprise SaaS spend on usage or outcome models by 2030. Negotiate this renewal knowing the vendor is more frightened than you are.
06
Key accounts stay human - deliberately
The evidence says relationship selling holds value precisely because it cannot be automated. Protect senior selling time as the scarce asset it is; the saving from automating it would be the most expensive money you ever made.
§ Part three
What good looks like.
A good commercial function on the other side runs two motions honestly: machines working the long tail - small deals, tail suppliers, routine renewals, admin - and humans working judgement, relationships and the accounts that decide the year. The pipeline is smaller and truer, the forecast survives contact with the quarter, and nobody pays twice for the same seat.
01
Machines on the tail, humans on the head
Every deal, supplier and renewal is classified by whether it deserves human attention. The tail gets worked by agents with clear mandates; the head gets more senior time than before. Both numbers are reported.
02
A pipeline you would bet on
Fewer, better-qualified opportunities with honest stages - because machine-generated volume was measured on conversion, not activity, and trimmed accordingly.
03
RevOps with real authority
Owns the data, the agents, the tooling budget and the measurement. Staffed and paid as the operating function it now is.
04
Commercial terms fit for agents
CRM and sales-tool contracts on pricing models that survive the shift, comp plans built on outcomes, and partner agreements that name who benefits when the machine does the work.
The change, across all five pillars
People
SDR work becomes pipeline quality control; senior selling time is protected as the scarce asset it is.
Processes
Machines on the tail, humans on the head - with comp rebuilt around outcomes before reps find the loophole.
Data
CRM hygiene treated as a commercial capability, because agents amplify whatever they find in there.
Platforms
CRM and agent pricing negotiated knowing the vendor has changed its own model three times already.
Partners
Key relationships stay human; partner terms name who benefits when the machine does the work.
§ Part four
The first 90 days.
Resist the AI SDR demo. Start where the evidence is: give reps their hours back, put an agent on the tail, and find out what your pipeline is really made of.
Ⅰ
Weeks 1-2
Audit the funnel and the data
Conversion by stage, cost per opportunity, CRM data quality, where rep hours actually go. Decide which segment of deals or suppliers is genuinely tail.
Ⅱ
Weeks 3-6
Ship assist first, tail second
Roll out call summaries, research and CRM hygiene to every rep - the boring, evidenced win. In parallel, put one bounded agent on the tail segment with clear rules and a human escape hatch.
Ⅲ
Weeks 7-10
Measure conversion, not activity
Meetings-to-opportunity, tail-deal outcomes, forecast accuracy, hours returned. Kill anything that produced volume without conversion. Publish the numbers to the team.
Ⅳ
Weeks 11-13
Reprice and reshape
Open the CRM renewal conversation armed with your data. Redraw the SDR and RevOps roles honestly, and rebuild the comp plan before the machine-sourced pipeline breaks it for you.
The day-90 scorecard · what you should be able to show
§ Part five
The traps, so you can skip them.
Trap 01
The autonomous SDR fantasy
Half the market bought it; half of those returned it within a year. If a vendor's case study has no conversion numbers, that is itself the conversion number.
Trap 02
Confusing activity with pipeline
Agents generate outreach beautifully. Opportunities, less so. Any metric a machine can inflate for free is no longer a metric - it is decoration.
Trap 03
Running agents on dirty data
AI on a neglected CRM automates your worst habits at scale. Clean the data first or enjoy confident fiction in the forecast.
Trap 04
Paying seats and agents both
The vendor would love the uplift and the seat count. Make one fund the other, in the same negotiation, in writing.
Trap 05
Automating the relationship
Your key accounts can tell when the sincerity was generated. The market has receipts now. Spend the saving on more human time for the accounts that matter, not less.
Trap 06
Leaving comp for later
Reps optimise for the plan, not the strategy - that is their job. If the plan pays for activity machines now do, expect creative behaviour by the second quarter.
§ Before you go
Six questions worth an honest hour.
Score yourself before someone with a rate card does. If more than two answers are uncomfortable, that is not a crisis - it is an agenda.
The self-assessment
§ Start with the workshop
Book your
Catalyst Workshop.
Bring the strategic change you have been circling. In 1-2 executive days, we will turn it into the first move, the case for it, and the mobilisation plan.