The short answer is: not yet — at least not without a governance layer between the agent and the customer. AI agents can draft effective emails, but every customer-facing message carries the potential to create a commitment, and commitments made without the SE's awareness become liabilities.
The Risk Isn't Bad Writing — It's Wrong Commitments
Modern AI agents produce well-structured, professional emails. The quality of the writing isn't the concern. The risk is what the email promises.
An agent drafting a follow-up might reference a timeline the SE hasn't confirmed. It might offer to share a document that doesn't exist yet. It might imply availability for a meeting during a week the SE is already overcommitted. Each of these is a commitment to the customer — and if the SE doesn't know it was made, they can't fulfill it.
In technical sales, where deals often hinge on the SE's credibility and reliability, a single wrong commitment can do more damage than a missed follow-up. The customer doesn't distinguish between "the SE promised this" and "the SE's AI tool promised this." It's all the same brand.
The Approval Spectrum
The question isn't binary. There's a spectrum between "agent sends everything autonomously" and "SE reviews every word before it goes out":
Full autonomy — agent sends without review. Highest risk. Appropriate only for zero-stakes messages with no commitment potential (e.g., meeting confirmations with pre-approved times).
Pre-send review — agent drafts, SE approves before sending. Moderate friction, high safety. Works for most customer-facing communications, especially those involving commitments.
Post-send monitoring — agent sends, system monitors for commitment conflicts or drift. Lower friction, requires robust monitoring infrastructure. Best suited for high-volume, lower-stakes communications where the SE trusts the agent's context but wants a safety net.
The Execution Intelligence Approach
An execution intelligence layer enables a smarter version of this spectrum. When an agent drafts an email, the system can automatically check whether any commitment in the draft conflicts with existing commitments, whether the timeline matches the SE's availability, and whether the content aligns with the deal's current context.
Low-risk messages — no new commitments, no deadline changes, no scope implications — can flow through with minimal review. High-risk messages — new commitments, timeline changes, or content that conflicts with existing deal context — get flagged for SE review before sending. The SE maintains control over what matters while the agent handles the volume.