Salesforce to Acquire Fin (formerly Intercom) for $3.6B

Every automated service I ever interacted with was built in a way that it would not help you with anything that incurs a cost to the company that built it.

I think frontier models are really good at solving technical problems. So if your support call is "how do I do this with the product" and there is a way to do that then I agree it can be a great experience. At the point where you need to involve the engineering team and maybe get bugs fixed I imagine the AI will push back because the entire reason the AI is there is to solve cost and if it ends up creating bugs or bugging the engineering team it'll start consuming money instead of saving it. It probably could be engineered to do this but the incentives are pretty strong not do that.

But what you shouldn’t do is try to dress up adversarial policies behind a friendly customer service bot and then fire your entire support staff. That is immediately obvious and will drive people bonkers.

its surprising to HN tech crowd but a well implemented support agent gets higher reviews and successful resolutions than humans

I find that 99%+ of support LLMs are just an extra hurdle that cant provide me with what I want.

I used Kogans chat LLM the other day, I gave it the details of my enquiry and it escalated me to a human. The human reviewed the case and then raised a ticket for a technical contact. I am not particularly angry at the LLM or the first human, but one of them is definitely redundant and wasting my time.

My main takeaway from that interview was that tier1 CS is a low-wage, high-turnover industry, so the amount of training agents can realistically get is limited. AI support is cheaper, smarter, and can take in more context faster. With AI, it makes sense to have a "support engineer" who identifies the edge cases and writes Markdown documents instructing agents on what to do in those cases. This actually lets you make agents much, much more helpful than human employees ever could be.

[1] (There's a high-quality transcript available) https://www.complexsystemspodcast.com/episodes/des-traynor/

This is the key, IMHO. I have not been able to access my Coinbase account for some time due to a login issue. (No error is displayed, but a successful login returns me to the login screen). The AI driven phone agent was actually very good at working through the problem, but there was no escalation to a human after that. When the agent couldn't solve my problem, it told me to mail a physical letter describing my issue to an address on 5th Ave in NYC.

Containment from the bot alone shot up past 50%. Customer satisfaction with support went through the roof. And this was basic knowledge support with account awareness. We had not wired up any tools yet.

So instead of companies working on their AI support execution, I wish they would just work on their website.

Can someone give me examples of the kind of support they want from a company that could be provided by AI, but could not be better provided by a well designed app and/or site?

Seems like a catchup play by Marc

> We replaced our helpdesk with Hermes

You didn't just put a computer somewhere and have a customer go up and type on the computer right? You have surrounding infrastructure on that right? That is the value add.

So many companies have such failed cultures they are just getting by delegating all serious matters to younger companies with people who actually care. If your staff never benefit from any of their work, nobody has any reason to care about how well you build your own in house Support / CRM / Chatbot / SaaS.

Not sure if this has been coined as a term, but its some form of "effort arbitrage"

demand for ai support vendors is going vertical this year

This is the same reasoning people use to say SaaS is dead, but it makes no sense. Rolling things yourself is often 10x more costly and not worth it, even with agents you need to pay 5-10 guys 150k-250k a year to build and train your own agent, why not pay fin 250k flat and not deal with any of it? Same goes with basically all other software that has nothing to do with your core product.

SaaS is alive and well and will continue to be.

You want to sell stuff? Don't mind my existence, let me look first at everything you have on display, and I'll initiate the conversations when I feel ready.

So this isn't as much of a financial engineering cost cutting move as it feels like to the type of person who truly calls because the require a human. It truly provides better service to the majority of people because they get their answer faster and more efficiently.

This is also demonstrated in the pricing of these systems at a per "open cases resolved"- they're putting their money where their mouth is.

Of course I'm also personally in the group where I call because I can already read a support page and I really need a human.... It could conceivably put true human support into another tier higher of perceived value.

Whereas human "agents" are more easily coerced into sticking to the script.

I’ll note this failure mode generally applies on tier 1/2 support with humans as well.

In one case I was literally repeating back to the human what I’d just been told, and getting them to confirm that what I said was correct. First bill arrives and I find out the truth.

Second case I was told I’d have to cancel and create a new account to add a service. I decided to keep my existing account and learned that there is a web page where I can easily self-serve and add the additional service in one or two clicks. (I assume like the human actually made more money for “new account signups”.

My point is that the feeling of being a valued customer is really independent of whether you’re interacting with a human.

Hopefully Salesforce did their due diligence, because the "AI agent" story here on Intercom (Fin) seems highly inflated. The product seems to be a a hybrid of RAG, some post trained models, curated help center content, custom answers, workflows, a bunch of if-else rules, API connectors, escalation logic, and specially generous resolution accounting.

Calling every solved interaction with the "AI did it" is misleading unless they separate confirmed resolutions from assumed resolutions, and disclose how much came from rules and workflows or custom answers versus LLM reasoning...

From their own docs, it seems a Fin "outcome" can be counted on, not only when a customer confirms resolution, but also ...when the customer simply does not ask for more help after Fin responds...A very soft resolution metric...

However, in the case of support agents. If it worked, and it was painless that would be something.

For example… On the company side, if it could reduce human support to the customers that actually need support, that’s cool. Your support agents aren’t spending all day with the three common issues or replacing stickers.

On the customer side; if I could call in and immediately get support without being on hold with their shit repeating audio script, didn’t have to spend 10 minutes “looking up my account” to an accent I can’t understand and repeating my name and address multiple times.

That said… AT&T is already using the absolute worst case scenario - they are currently using AI with a slight Indian accent and pretending it’s real peoples. It seems to be 90% automated, and if you question it about being AI or have a question it can’t understand a human pops in on the other side, interacts, then hops off and it goes back to being full-AI.

It could be great but it’s already awful.

But many people will contact support instantly when they think of something, no matter what. Even if the website and other customer-facing material is crystal clear and has all information necessary.

AI chatbots is the way a company deals with the latter, because these customers most of all want a conversation. The question is if they will be satisfied with a robot, or still demand to talk to a person.

https://www.intercom.com/blog/today-intercom-becomes-fin/

It's the AI model that scores higher than Anthropics on customer support tasks.

It was a super hopeful time: JQuery meant programming for the web was less painful, sockets and HTML5 SSE meant realtime was just starting.

Intercom and Olark were one of the first two "install this <script> tag" based customer support apps. They made websites a way to talk to companies rather than just read see and buy.

I can't believe the exit took so long, especially in a field so crowded. But it looks like it was well worth it.

Everything around you was made by someone just like you.

The very first interaction I have with Heroku is a two-factor sign-in, and …it's this horrible page hosted on salesforce.com, which doesn't have retina graphics (i.e. is blurry on screens Apple have used for 15+ years), doesn't work properly with automatic one-time-code generators (because the login form is heroku.com, but the two-factor is salesforce.com), and …gah! What a mess. Thankfully you fall through into the pretty, well thought-out Heroku dashboard of yesteryear.

AI is definitely capable of taking on customer support work at the moment, and to a high standard as well. Sure, it's not perfect. But it's not a grift either.

I’m not sure where you’re getting this from but their customers find Fin to be a hugely impactful tool

But businesses will always chase that dream of reduced customer contact, so Salesforce will keep selling it to them.

An AI secretary seems perfectly acceptable for both sides. The expectation is that a real human comes in soon after but this seems like a way to free up the most tedious parts of the process for both sides.

While I think that this is a bad move for Intercom, it's actually brilliant for Dublin and Ireland that they have finally exited.

https://investor.salesforce.com/financials/quarterly-results...