Momentum, call insight, and what the acquisition does not prove

A sales leader asks for the business case on Monday. They want a percentage. Record the calls, they say, and the win rate moves. The number is already in the slide, with a footnote that says "expected". Nobody in the room has read the announcement. The footnote is doing the work the evidence did not do.

Take the percentage off the slide before you do anything else. Then decide whether you are allowed to record, who may hear the recording, and which sentence, if any, is allowed to land on an opportunity.

#What closed on 2 March 2026

Salesforce's close notice says: "Salesforce completed its acquisition of Momentum on March 2, 2026." Momentum is conversational insights from video and calls, aimed at Agentforce and Slack. People talk. The product side of the deal is about drawing insight from that talk, and about pointing that work at Agentforce and at Slack. The sentence that closes the deal does not say your org already has the feature, and it does not say a forecast field will update itself.

Aimed at is not the same as installed. If a seller asks whether their Sales Cloud org gained call summaries on 2 March 2026, this source does not say yes. Check the contract and the release you are actually on. A completion date is a fact about Salesforce and Momentum. It is not a fact about your tenant.

Keep the aim in view when you draw the picture. Insight from calls and video. A path towards Agentforce, where an agent might use that insight. A path towards Slack, where a person might read it. Two paths, two rooms, two permission problems. Do not draw one arrow labelled "AI" and call the design finished.

#A recording is not a win rate

Do not invent a win-rate change from call recording. The announcement does not give you one. A before-and-after on your own pipeline is a study you have not run. Putting a borrowed percentage on a slide, or a round number that "feels fair", is how a programme gets funded on a claim you cannot defend when finance asks for the working.

What a recording can be, if you do the rest of the work, is a source. A person spoke. A summary might say they asked about timing, or named a competitor, or agreed a next step. That summary is a claim about words. It becomes a CRM fact only if you choose a field, an owner, and a rule for when the write is allowed. Until then it is a note. Notes do not move win rate by existing.

There is a second temptation: to treat the summary as coaching proof. A manager reads three sentences and concludes the seller pitched badly. Maybe they did. The summary is not a verdict, and it is not a score you should drop into a pay calculation because a model sounded sure. If coaching is the goal, say so, and decide who may read the call. If pay is the goal, you need a policy a person can apply, not a paragraph generated from audio.

Refuse the slide that says win rate in the title and "Momentum" in the footer. Those are unrelated objects until you have evidence, and this page is not that evidence.

#Consent, retention, and the room

Before anyone records a customer, write the lawful basis and the words the customer will hear. This page cannot tell you which basis applies in your country. It can tell you that a project which skips the question will record first and ask legal later, and later is too late for the calls you already stored. Name the person who signs the decision. A shared channel is not a sign-off.

Retention is the next line. How long the audio or video is kept. How long the transcript is kept. How long the summary is kept. Whether a deleted customer request reaches the recording, the transcript, and the CRM note, or only one of them. If you cannot answer, you are not ready to turn recording on for a pilot of "a few friendly accounts". Friendly accounts are still customers.

Who may hear it. The seller on the call. Their manager. A revenue-operations analyst. An agent running as a broad integration user. A Slack channel for the whole region. Each of those is a different publication. The region channel is the one teams pick because it feels collaborative. It is also the one that puts a customer's words in front of people who were not on the call and who may not need them. Default to the smallest room. Widen it with a decision, not with a habit.

Slack as a destination needs the same care as any other destination. A summary posted to a channel will be forwarded. Write the rule: link, or text, or nothing. If the text includes a price the customer floated, that price is now in chat history. Decide that before the first pilot call, not after a screenshot circulates.

#What may be written on the opportunity

Insight that never lands on a record cannot break a forecast, and it also cannot help a colleague who was not on the call. So you do have to decide the write. Keep it small.

Pick the fields. Next step. Competitor mentioned. Date of the next meeting. A short note. Do not start with amount, stage, or close date. Those fields already have owners, and a summary that overwrites them will create a fight you have seen before, between a seller's judgement and an automated sentence.

One owner per field. If the seller owns next step, the summary may propose a next step in a place the seller accepts or rejects. It may not overwrite the field because the model was confident. Confidence is not an owner. Field history should show a person if the value changes.

A refusal list sits beside the field list. Do not write a medical comment a customer made in passing. Do not write a card number someone read aloud. Do not write a guess about whether the deal will close. Do not write a personality judgement. If the summary contains one of these, the write does not run. A person can still read the transcript under the access rule. The opportunity stays clean.

Agentforce, as the aim, does not change this. An agent that can read the summary can also repeat it. If the summary is wrong, the agent will be wrong in a helpful voice. Fix the write. Do not add a paragraph of instructions asking the agent to ignore fields you allowed it to see.

#The migration keyword, and the cost list

"Momentum data migration to salesforce quote" assumes a store of Momentum data you load into Salesforce. You might have one, if you used the product before the close: recordings, transcripts, summaries, user maps. You might not. Ask. If an export exists, the quote is about format, identity, retention, and which fields you will accept. If it does not exist, say so, and do not let a migration line survive in the proposal out of habit.

Price these items. Do not price a win-rate outcome.

• A written decision on consent, with the person who owns it.
• Retention for audio or video, transcript, and summary, including how a deletion request is carried out.
• The list of people and rooms that may hear or read a call.
• The fields a summary may propose, and the person who must accept a write.
• The refusal list, tested on a real transcript, not on a polished example.
• What happens to pre-close exports, if any exist, and who may open them.
• A pilot boundary: which teams, which call types, and the date you review the writes.
• Explicitly out of scope: any claim that recording will change win rate.

A partner who opens with a percentage has skipped the list. Send the percentage back. The list is the cost. Meeting time to settle consent and fields is part of that cost, and it is cheaper than a quarter of recordings you are then told to delete.

#How to test a partner

The keyword says "best momentum salesforce integration agency". There is no such ranking that knows your consent rules. Use a test.

The firm asks where consent is written down, and stops if it is not. They will not put win rate on a slide. They name the fields they will not write: amount, stage, and anything on the refusal list. They can show a transcript that contains a card number or a health aside, and show that the CRM write did not take it. They treat a Slack channel as a publication. They can separate "aimed at Agentforce and Slack" from "already on in this org".

Hiring a Momentum Salesforce consultant is the same test. Implementation services that start by connecting a recorder to every user's calendar have started at the wrong end. The calendar is easy. The field owner is the work. A partner who cannot name the field owner in the first workshop is not cheaper because their day rate is lower. They are early, and they are early at your expense.

#What to settle before the pilot

#A pilot that stays small enough to reverse

Pick one team, one call type, and a date six weeks out when you will read what was written. The call type should be ordinary. A renewal conversation is a better test than a launch webinar, because a webinar is a broadcast and a renewal is a customer talking about their own contract. If you cannot get consent language agreed for that one team, you cannot get it agreed for the region, and the region is the wrong place to discover that.

During the six weeks, read the field history. You are looking for writes the seller did not accept, summaries that mention money, and notes that include a remark the refusal list was meant to catch. Count them. A count of defects is allowed. A win-rate comparison is not, not on this sample, and not as a reason to expand the pilot early. Expansion is a new audience decision. The people who were not in the first team do not inherit the first team's consent conversation by being in the same business unit.

Write down what you will turn off if the review is poor. The recorder. The CRM write. The Slack post. Those are separate switches. Teams that only have "the integration" as a single switch will leave the write running because turning it off also turns off a report someone likes. Split the switches before the first call, while turning them off is still cheap.

Search across company knowledge, with Slack as the interface, is a different acquisition and a different permission problem. Do not brief them as one "Slack AI" project. The search side is Doti, Slack, and enterprise search.

#The other searches

The project quote is on momentum implementation cost. That page does not print a figure. Who bought the company is on what happened to momentum. The connector is on integrating momentum with Salesforce. Whether to move is on should I migrate from momentum. The customer fee is on momentum licence cost. Who to hire is on hiring for momentum. This page does not rank an agency.