The 30-60-90 Ramp Plan When New Reps Work Alongside Agents

The classic ramp plan is an information transfer schedule. Week one product, week two competitors, week three process, week four shadowing. It exists because information used to be scarce and slow to acquire.

That constraint is gone. A new rep can now produce an account brief in ninety seconds that would have taken a senior rep an afternoon in 2022. If the ramp plan does not change, you are spending thirty days teaching something the tooling already delivers, and skipping the thing that now decides performance: judgement about machine output.

What ramp is actually gating on now

Three capabilities separate a productive new rep from a fast-typing one.

1. Discrimination. Can they tell a good machine output from a plausible wrong one. This is the editor's skill, and it is learned by doing the work manually first. The reasoning is in the editor's mind.
2. Qualification judgement. Can they hold a disqualification decision under pressure from a manager who wants pipeline coverage.
3. Commercial construction. Can they build a deal shape, not just a proposal, including the multi-threading and the internal path to signature.

None of these are content problems. All three are practice problems, which means ramp becomes a rehearsal schedule rather than a curriculum.

Days 1 to 30: manual first, deliberately

The counterintuitive move is to withhold the agent for the first two weeks on the tasks the rep must be able to judge later. Not as hazing. As calibration.

• Write five account briefs by hand. Then read five machine-generated ones and mark the errors. The gap between the two is the lesson.
• Listen to ten recorded calls and write the summary yourself before comparing to the automated summary.
• Learn the qualification criteria as a decision, not a form. Practise disqualifying out loud.
• Learn the systems, the stage definitions and what each field means, because those definitions are now production dependencies rather than reporting preferences.

Gate one: the rep can find at least one material error in a machine-generated account brief, unprompted.

Gate two: the rep can state, in one sentence per criterion, why a named prospect is not qualified.

Days 31 to 60: supervised leverage

Now switch the tooling on and change the measure. The rep is no longer judged on output volume, which is now trivially high. They are judged on the intervention rate and the quality of what shipped.

• Every machine-drafted message is edited before send, and the edit is logged. Managers review the edits, not the drafts.
• Live call practice against the three objections that lose the most deals in your data, not the three that are easiest to script.
• First multi-threading exercise: name every person required for a decision in a real open deal, and the path to each.
• Weekly deal review where the rep presents evidence, not narrative. This is the same standard applied to vendors in the pipeline truth test.

Gate three: intervention rate is falling and shipped quality is holding, measured by manager review of a random sample.

Gate four: the rep has run one full discovery call end to end with a scoring sheet completed by an observer.

Days 61 to 90: unsupervised, with limits

Full tooling, own pipeline, defined limits. Discounting, contract language and anything irreversible still require a gate, exactly as they do for agents. The reversibility test applies to humans in ramp too.

• Own a defined territory or book with a named coach, not a shared one.
• Deliver one internal teach-back on a competitor or a segment. Teaching is the fastest test of understanding.
• Complete a pipeline review with a forecast call, and be scored on the accuracy of that call two weeks later.

Gate five: a forecast call that proved accurate within an agreed tolerance.

Gate six: one closed-won or closed-lost deal with a written post-mortem the team can reuse.

What enablement stops doing

Two categories of enablement work lose their value in this model. Static content libraries, because retrieval at the moment of need beats memorisation, and certification quizzes that test recall of information now available instantly. Replace both with delivery at the point of work and observed competency tests. The delivery model is in just-in-time enablement with AI, and the wider role change is in enablement in the age of agents.

Measuring the ramp itself

Track four numbers and compare cohorts, not individuals: days to first qualified opportunity, days to first closed deal, intervention rate at day sixty, and forecast-call accuracy at day ninety. If the tooling is working, the first two fall and the last two improve. If only the first two fall, you have built a faster path to bad pipeline.

Related: Just-in-time enablement · The editor's mind · Systems beat talent · Skill: building onboarding ramp plans