A Framework for Redesigning SDR Compensation for Hybrid Teams

AI changed the SDR job before most comp plans noticed.

The old plan was built for a rep who spent the day finding accounts, writing emails, chasing follow-ups and updating the CRM by hand. Most of that work now runs on software. McKinsey estimates that roughly a fifth of today's sales-team functions can already be automated, and the prospecting grind is first in line.

That leaves a strange gap. The plan still pays for activity, the dials logged, the emails sent, the contacts touched, while the AI does most of that work. The part of the job only a person can do, the live conversation that turns interest into a booked meeting and eventually a closed deal, barely shows up in the math.

It is worth fixing properly. Even before AI arrived, reps spent less than 30% of their time actually selling, by Salesforce's count; the rest went to research, list-building and admin. AI's real promise is to hand those hours back. A comp plan that still rewards the admin will quietly fight that shift instead of funding it. Here is how to redesign SDR compensation for a hybrid team, where AI runs the outreach and your reps own the relationships.

Why the old plan fights the new reality

Most SDR comp plans run a 60/40 or 70/30 base-to-variable split, with the variable tied to activity and meetings set. The Bridge Group, which has benchmarked the SDR role for years, puts the average somewhere around $80,000 to $85,000 in on-target earnings, most of it in base pay. The fully loaded cost runs well past that once you add benefits, tools, recruiting and management. That structure made sense when the rep personally generated the activity. It makes much less sense when the software does.

Say your plan pays $4 a dial and $50 a meeting set. Two years ago, those numbers tracked effort fairly. Now AI builds the list and sends the first-touch email, so the rep did not make those touches happen, and paying for them rewards a number the rep no longer controls. Worse, it nudges reps toward volume the system already handles better, instead of the conversations only they can have. You end up paying for friction.

Here is what AI now owns in a well-run hybrid motion:

• Finding the right accounts and researching them
• Building and verifying the contact list
• Drafting personalized first-touch outreach
• Sending the email and running the follow-up cadence
• Logging activity and updating the CRM

None of that is where a human is uniquely valuable anymore, which is exactly why the traditional SDR model is straining. A plan that still prices the rep as an activity machine is paying for the wrong thing.

Where the human actually adds value

When AI takes the grind, the rep's job narrows to the part that resists automation: the live human exchange. Reading a prospect's real intent on a call. Handling the objection that was not in any script. Building enough trust that someone agrees to a meeting and shows up. Carrying context cleanly to the account executive who closes. That is higher-value work than logging fifty dials, and the SDR role has shifted far enough that the title no longer explains itself. Your comp plan should say so out loud.

The hybrid compensation framework

Keep base salary as the anchor, then rebuild the variable around outcomes the human controls. Four levers do the work: the meetings a rep books, the pipeline that results, the deals that close, and the rep's skill at working with AI. The variable portion, roughly 30 to 40% of OTE, gets divided across those four. Here is how each one earns its place.

Anchor the base (60 to 70% of OTE)

Base pay is stability, and a hybrid rep needs it more, not less. When AI handles the volume, the rep runs fewer but higher-stakes conversations, and a slow week should not put their rent at risk. Decades of Harvard Business Review research on sales pay make a related point: plan design drives behavior, and blunt moves like capping commissions tend to suppress performance rather than lift it. So hold base at 60 to 70% of OTE. Set the number against your market and the complexity of the sale: $50,000 to $60,000 is typical, higher in expensive metros and for harder products.

Pay for the meetings your reps book

Meetings still matter, but raw count is the wrong unit. When AI warms the contact and the rep walks into a conversation that is already interested, the skill is converting that interest into a real, qualified meeting, not booking anything that says yes. Weight the payout toward quality: a smaller amount when the rep books the meeting, a bonus when it actually happens, and the largest piece when it converts to an opportunity.

A simple version looks like $50 when the rep books the meeting, $100 more when it is held, and $150 more when it becomes an opportunity. The rep is paid most for the outcome they most influence.

Reward real pipeline, not meeting count

AI surfaces interested leads. Turning that interest into pipeline an AE will actually work takes human judgment, and that judgment deserves its own line in the plan. Pay on opportunities created from the rep's conversations, on pipeline dollars influenced, and, where you can measure it cleanly, on velocity from meeting to opportunity. This is the lever that keeps reps honest about quality instead of gaming meeting count.

Put a slice on closed-won

Give the rep a small stake in deals that close from their sourced opportunities, either a flat bonus or 1 to 3% of value. The first human impression often sets up the relationship that eventually closes, so rewarding the outcome ties the rep's work to revenue and pulls their attention toward quality interactions over volume.

Pay reps to work well with AI

The newest lever, and the one most plans miss, rewards the human-AI collaboration itself. Pay for the things that make the partnership produce: sharp follow-up on AI-surfaced interest, useful feedback that improves targeting and messaging, fast adoption of new capabilities. A small SPIF (a sales performance incentive fund) tied to collaboration milestones signals that working well with the system is part of the job, not a favor.

An example: an $80,000 plan

For an $80,000 OTE position, the split might look like this:

The shape is the point. Less weight on activity the software now owns, more on quality and outcomes, a new line for AI collaboration, and a thread of closed-won that ties the rep to revenue. None of it cuts the base your reps live on.

Putting it in place

Redesigning a comp plan is a change-management project, not a spreadsheet exercise. Five steps keep it from blowing up.

1. Audit what you have. Map your current components and, beside each, write down whether a human or the AI now produces that result. Anywhere the plan pays for something the software does, you have found a line to change.
2. Define the hybrid role in writing. Spell out what AI handles, what the rep handles, where the handoffs sit, and the quality bar for the human touchpoints. The same goes for targets: when AI does the prospecting, the old activity quotas need rethinking too.
3. Model the money before you ship it. Run the new plan against last year's numbers for a top performer, an average rep and a laggard. Top performers should earn more, average reps should hold steady, and the plan should create real urgency at the bottom. If your best rep takes a pay cut under the new math, fix the math.
4. Communicate and train. Explain why the plan changed, show earning scenarios, and train reps on the AI tools that drive the metrics they are now paid on. A plan nobody understands is a plan nobody chases.
5. Monitor and adjust. Track achievement by component, watch for gamed metrics, gather rep feedback, and tune quarterly. The first version is a hypothesis, not a monument.

Mistakes that sink the redesign

A few failure modes show up again and again.

The most common is cutting base too hard. High-variable plans feel bold, but they punish the steadiness a hybrid rep needs and tend to drive performance down, not up. Keep base at 60 to 70% of OTE even as you reshuffle the variable.

The second is paying for work the AI does. If the software sends the email, do not pay per email sent. Run the audit, find those lines, and retire them.

The third is rewarding meetings with no quality bar. Pay for raw count and reps will book anyone who says yes, so attach qualification standards that make the payout follow real opportunities.

The fourth is forgetting the relationship entirely. AI does not build trust; people do. If nothing in the plan rewards the quality of the human connection, through closed-won, customer feedback or held-meeting rates, you are underpaying the one thing that is now most valuable.

How Pair Selling makes the math obvious

Pair Selling is AvairAI's name for this division of labor, and it happens to make compensation design straightforward. One side of the line is the work AI does, which you do not pay a person to repeat; the other side is the human work you reward.

AI runs the prospecting: finding accounts, building and verifying the contact list, personalizing and sending the email, orchestrating the multi-channel cadence and keeping the CRM current. Your reps own the parts that close revenue: the calls and LinkedIn touches, the discovery conversations, the objections, booking the meeting and the warm handoff to the AE. Pay for that second list. It is the cleanest way to put the principle into practice, and it is why a compensation model built specifically for Pair Selling maps so neatly onto the framework above.

From a plan to performance

A comp plan is a message. It tells your reps, in the only language that fully lands, what you actually value. Teams that rewrite the plan for the hybrid reality, rather than bolting AI onto a plan built for 2015, tend to see the same things: reps who engage instead of resist, cleaner collaboration with the tools, better lead quality, stronger pipeline and a lower cost per opportunity. That payoff is part of the broader move to an AI-assisted SDR motion, and the way you lead the team shifts alongside it, which is its own discipline.

The framework gives you the structure. What it asks for is the willingness to pay for human work like it is the scarce, valuable thing it has become.

Want to see the AI half of the partnership at work? Launch your first Pair Selling campaign from just your website, and put your reps on the conversations that close.