Salesforce has agreed to buy Listen Labs, an AI customer research company. Neither side has disclosed the price. Press reports put it at around $2bn, roughly 67 times Listen Labs’ reported annual revenue.
That is a lot of money to pay to listen.
Talking about CultureFlow in the same breath feels ambitious. But the market is moving in the direction we are already heading, and I think it is worth saying where.
What Salesforce actually bought
The clearest understanding of the deal I have seen came from Matt Gretczko at PwC. He calls Listen Labs the new “sense” layer of the stack.
His frame is a loop: sense, decide, act, and orchestrate. For years, companies put their money into deciding and acting: routing, resolution, and tracking. Then they ran all of it on stale inputs. Lagging surveys. Fragmented insight. Matt’s point is that you can have the best decision logic in the world and still get the customer wrong if the signal underneath it is out of date.
Listen Labs fixes the input. Its AI agents design studies, recruit from a panel of more than 50 million people, run interviews in more than 120 languages, and build digital twins, so a team can test a message on a simulated customer before a real one ever sees it.
Salesforce is not the only one betting on this. In July, Simile raised $200m at a $2bn valuation to simulate human behaviour. The market has now spent billions confirming one idea: simulate before you make contact. It saves money, time, and effort.
I agree with all of it. But I think there is a piece missing.
The third of your base
Matt gives an example worth focusing on. You build an agent that handles a billing dispute, then find out in production “that your script reads as tone-deaf to a third of your base.”
Which third? And why?
Very often, the answer is cultural.
Same words. Different interpretation. Directness reads as honesty in one room and as aggression in the next. Warmth reads as care to one buyer and as a lack of seriousness to another. A question meant as curiosity lands as a challenge to someone’s competence.
I have run workshops for more than 200 professionals, and this is the pattern they kept bringing me. You can see that it isn’t working. You can’t explain why.
A sense layer built on interviews and behaviour is very good at the first part. It can tell you that the script is broken. It is much harder for it to tell you why, and the why is what you need to fix it before you ship.
Scripted AI interviewing in 120 languages cannot carry the cultural signals underneath them.
Nationality isn’t the answer either
The obvious fix is to segment by country. Most cultural tools stop there, and the research says it isn’t enough.
A 2016 meta-analysis by Taras, Steel and Kirkman, covering 558 studies, found that about 80% of the variation in cultural values sits within countries, not between them.
So “German buyer” is not a profile. A Millennial German woman who is an SVP of Operations in Hardware with 15+ years’ experience reads a cold email differently from a Gen Z German man three years into a junior sales role. The country is the same. The response is not.
CultureFlow’s version of the sense layer
At CultureFlow, our sense layer is focused on how people work together, globally. We call it Cultural Intelligence.
Underneath it is our Cultural Intelligence Engine: a graph of peer-reviewed research in which six inputs (nationality, gender, industry, age, experience, and seniority) combine into a Cultural Signature. Those six inputs compose 469,200 distinct Signatures, and around 110 billion possible pairings between them. Each pairing shows where friction between two people is likely to sit, and why. Each pairing is queryable.
It’s applied AI on a research-backed graph, not an LLM’s best guess.
Here is what that looks like in practice. Before you write to that German SVP, COMPASS reads her Cultural Signature against yours. It shows you where the friction is likely to be, then drafts the email, the LinkedIn message, and the call script, each with the cultural reason it is written that way. You can see why it should land before you send it.
If you would rather have that read inside your own tools and platforms, the engine is queryable directly. How the engine works.
Culture is the sense layer no one has built for
Matt is correct in saying that the sense layer is becoming the missing layer of the enterprise technology stack. Culture is one of the sense layers no one has thought to build for.
Our Cultural Intelligence Engine can tell you how and why a customer is likely to respond, and help shape any interaction to make it smoother, so that meetings, emails, calls, and messages land more successfully.
It makes sense, for any business whose pipeline crosses borders, or that wants to win deeper within existing markets to improve conversion and deal flow, to understand how to work with their potential buyers.
If you want to see the difference, message me. It will change how you understand your ICP and how you work with your next potential customer.