Automate the recording, the routing and the reminding. Leave the qualifying, the pricing and the first human sentence alone. The test is simple: when it gets this wrong, does it cost a correction or a deal.
Most teams automate in the wrong order, starting with the outreach because that is the part that feels like volume.
Automate these
Recording what happened. Calls logged, emails attached, meetings written to the record without anyone typing. This is pure admin, it is where reps lose hours a week, and getting it wrong costs a correction.
Routing. A new enquiry reaching the right person by territory, size or product, immediately rather than in the morning. Speed to first response is one of the few things in sales that reliably correlates with winning, and a routing rule is the cheapest way to buy it.
Reminding. A deal that has not moved in twice its usual stage time should surface itself. This is the single highest-value automation most teams do not have, because the alternative is remembering, and nobody remembers forty deals.
Preparing the draft. Assembling the follow-up from what actually happened, ready for a person to read and send. Note the ending: ready for a person.
Leave these alone
Qualification. It is a judgement about a specific situation with incomplete information. Scoring models approximate it, and the approximation is why the pipeline fills with deals that look qualified and are not. A person decides, the system records the decision.
Pricing and terms. Obvious, and it still happens through discount rules that quietly give away margin on deals that never needed it.
The first human sentence. The opening line of the first real message. Automate the sending, automate the timing, automate the follow-up schedule. The sentence that says a person looked at this company before writing has to be written by that person, and buyers can tell within a line.
| Step | Automate | What it costs when wrong |
|---|---|---|
| Logging activity | Yes | A correction |
| Routing a lead | Yes | A short delay |
| Chasing a quiet deal | Yes | An unnecessary nudge |
| Drafting a follow-up | Draft only | A rewrite |
| Qualifying | No | A quarter of wasted pipeline |
| Pricing | No | Margin, permanently |
| The opening line | No | The deal |
Where automations actually break
Not in the logic. At the seams.
A sequence that does not know the deal closed, so the prospect gets a chasing email the day after signing. A booking link that creates a meeting the CRM never hears about. An enrichment tool that updates a record the sequencer already copied.
Every one of those is two systems disagreeing about what is true, and the more tools in the stack the more seams there are to disagree at. Teams usually respond by adding another automation to reconcile them, which adds another seam.
That is worth knowing before buying the next tool: the failure mode of a six-tool stack is not any tool being bad, it is that nothing owns the truth.
The order worth doing it in
Start with recording, because it buys back the most time and risks the least. Then reminding, because it recovers deals you are currently losing to forgetting. Then routing. Drafting last, and keep a person on the send button.
Qualification and pricing stay with people. That is not caution, it is where the judgement you are paid for actually sits.
Where Empiraa sits
Empiraa Signal is a CRM and more. It is also the B2B prospect database, the enrichment, the email sequences, the booking link and the proposal tool, in one system with unlimited users on every plan. The reason that matters to this subject is the seams: the record, the sequence and the booking are the same system, so there is one less thing to reconcile and one fewer place for the embarrassing follow-up to come from.
Related reading: sales pipeline stages, and what probability to put on each, and the hidden cost of a disconnected sales stack.
How to write a sales email that gets a reply covers the sentence you should not automate.
