Every prospect gets one conversation, one path and one next move.

Follow-ups always reply inside the original thread. Workloom carries the reply headers and the provider thread reference forward with each send. That changes what the prospect sees: one continuing conversation instead of five cold emails that happen to come from the same sender. The mechanics matter because a follow-up that starts a new thread looks like a new interruption, even when the copy says it is continuing an old one.

The failure mode is familiar. A prospect replies to the first message, a rep misses it, and the next automated step lands anyway. Or a later message arrives as a separate thread and makes the sender look disconnected from their own outreach. Here, the per-prospect record holds the thread context alongside the sequence position. A reply can move the prospect forward without asking a rep to repair the sequence by hand.

That same record covers the campaign path around the conversation. A campaign can run as tiers, with each tier carrying its own sequence, or as one flat sequence when every prospect should follow the same route. Use tiers when the message path changes by account or segment. Use a flat sequence when the operating rule is simple. Either way, the thread and the path stay attached to the prospect.

Workloom puts the channel plan beside the sequence and the prospect pipeline. Email can carry the main thread. Other channels can play their assigned part without forcing the rep to reconstruct what happened elsewhere. The point isn't to spray the same message across every available surface. It is to give each step a place in the campaign while keeping enrollment, timing and status visible in one record.

A prospect may move through a campaign that combines several channels, or through a single-channel sequence with no extra handoffs. The operating choice belongs to the campaign. Workloom keeps the composition method, sequence step and prospect position together, so a change in channel doesn't hide the next action or create a second version of the record. This is where most stitched-together outbound stacks lose the plot.

The channel layer also connects to the rest of the operating surface. Use voice when a campaign needs a spoken step. Use ai for the reasoning behind message composition. The platform holds the campaign state around both. You should be able to inspect why a prospect is in a step, what comes next and which conversation carries the history without opening several vendor tabs.

Workloom supports four composition methods, spanning fully personalized messages through templated messages with a written opener. That gives a campaign a controlled range instead of forcing every prospect into the same copy pattern. Choose the method per sequence or step, then test the actual composition in the place where it will run. The goal is to compare working message paths, not produce a report disconnected from delivery.

A/B testing works at sequence level and per step. Sequence-level testing lets you compare two complete paths. Per-step testing isolates the part of the path that needs scrutiny, such as an opener or follow-up. A campaign can therefore test the whole motion or one decision inside it without rebuilding the surrounding sequence. Keep the test scoped to the thing you want to learn. Otherwise the result tells you several conflicting stories at once.

The test doesn't sit open forever waiting for someone to declare a winner. Workloom promotes a winner automatically once the sample is large enough to mean something. That gate matters. A tiny early lead shouldn't rewrite a campaign before there is enough evidence to support the change. Review the composition methods, choose the level of testing, then let the campaign move the better path into the active sequence when the sample threshold is met.

Every prospect carries their own pipeline position inside the campaign. Replies move them along that pipeline without a rep updating a field. That means the record reflects the conversation as it develops, rather than waiting for someone to remember which status to change after a busy call block. The sequence, thread and position travel together. A rep can act on the current state instead of reconstructing it from activity logs.

There is a campaign-level switch to turn reply-driven movement off. Use it when the team wants replies visible but prefers manual control over progression. That setting belongs to the campaign, not a global rule that surprises every motion. One campaign can advance prospects from replies while another leaves position changes to the team. The important part is that the choice is explicit before launch and remains visible in the campaign setup.

Before anything sends, Workloom runs suppression rules, a 30-day send projection and an eight-category pre-flight. These checks happen before launch, where they can still change the campaign instead of becoming a postmortem. If a campaign has the wrong exclusions, an overloaded send plan or a missing setup detail, the issue should surface before enrollment. The infrastructure underneath this is built to keep those handoffs in the same operating path.