Sierra Raises $950M at $15B Valuation
sierra.ai
sierra.ai
in reality 50-80% of callers come in with easily answerable questions because they don't know how to nav the website and prefer to ask in natural language
The vast majority of callers call in to resolve their issue, and most don't care if they are speaking to a bot because they just want their issue fixed. Agents (if implemented well) are an order of magnitude more effective at resolving issues compared to a call centre worker who is reading off a script and churn within 9 months
There's also the 2nd order effs of making CX cheap. before, there is the perverse incentive of companies trying to keep you off support because each call costs them way more than the value they get. if your cost per call drops 100x you can invest in turning a cost centre into a revenue driver (+ a better experience)
I'm sure AI Support Agents will be implemented better, but so far in my experience, the humans I connect to far outperform the AI agents.
I understand some people call with trivial or easy questions, and those might be handled just fine by an AI agent (just as they would by a human). But if I'm calling it's because it's not easy or trivial, otherwise I would have figured it out. Calling is always a last resort. And it's also because I want something handled quickly and don't want to spend time trying to navigate a maze of questions to try to get the AI agent up to speed and do what I want. So all the agent does is make me upset and not want to do business with that company again. And the more difficult the agent makes it to get to speak with a human, the more unhappy I become.
Stop trying to cut costs and extract maximum value while making things worse for customers, and stop trying to tell yourself -- and us -- that this somehow provides a superior experience for customers. That's BS. I've seen the step by step decline in phone customer service over the past 30 years, it has clearly never been about the user experience.
I'm also bullish because AI coding agents give up easily if my problem is complicated.
I think it'll be easier to convince an AI to transfer me to level 2 support than a human.
> Agents (if implemented well) are an order of magnitude more effective at resolving issues compared to a call centre worker who is reading off a script and churn within 9 months
But it seems like companies don't want to do this part, possibly because of fears that someone will trick the agent into giving them a refund or something. Or because the actual goal is to optimize for fewer costly refunds/cancellations/policy exceptions etc.
So for whatever reason, they stay stuck in that useless local maxima while simultaneously making traditional help increasingly difficult to get ahold of when needed for an overall net worse experience as a customer.
Refunds could require approval. And, it could not be just the agent's sole decision.
There's a certain vendor that requires me to place same-day orders by a specific time. You can easily place an order from the website. If you need to cancel one, you have to ignore the grayed out cancel button and call their cancellation support line. There you'll talk to an agent that doesn't have access to cancel orders, so you have to convince it that it can't help you before you can transfer to a real employee with the ability to hit the "cancel" button.
I’d say the vast majority of callers absolutely hate talking to these things and spend most of the call trying to get to a human, often getting frustrated and hanging up (shows up positive in the metrics, call handled without transfer!).
Though I’m not sure the companies deploying them really care, they’re just happy they can fire call center employees.
See, this is how we get such useless chatbots.
Though cox will perpetually try to gaslight me so it doesn’t help anyways
I went through this whole song and dance the other day with Uber. I needed to change something and the “AI helper” kept trying to force me into the lost item tree. They snipe keywords and ignore everything else. If you say “reservation” or “cancel” that’s all it works with with none of the context.
If we are designing a thing is so terrible that it makes customer support necessary (other than the obvious corner cases that ai cannot solve) then sure, let a computer do it. We’ve already failed at every other step.
And even something as simple toaster might need some customer support. These things do fail quite often, sometimes dangerously.
Increase complexity even a little; or worse, deal with a service (phone, internet, subscriptions, etc) and you will need this safety net.
Nothing is 100% reliable in this world.
For the other 20-50% it's much worse, and that's the problem. And people on HN will tend not just to fall in that group but in the top 5% of "least solvable by just reading the website" questions.
Raises the floor at the expense of lowering the ceiling.
And another thing...
Sierra says:
> Transform phone support with AI agents that speak naturally, reinforce your brand, and take action—across inbound and outbound calls.
Reinforce one's branding? For the better? Really?
Seems unlikely.
I've LARPed in most roles around product development. Tech support, sales, QA/Test, tech writing, marketing, etc. Enough to appreciate that engineering the entire lifecycle is important.
A comment elsethread states customers HATE these support robots. I believe it; me too.
Before adopting agentic CX (of whatever its called), I'd worry about alienating my current and future customers.
Penny wise, pound foolish.
I'm pretty sure every one commenting here has their own horror story about LLM support. Now that is what people are angry at.
What percentage of interactions having this result will cancel out your cost savings?
It's an odd use case - we have used language for a millionish years or so and it makes sense that that's the easiest way for us to get at information/do things.
But at the same time it's faster for me to read than listen, but it's often slower to type than to speak. It's faster to hit one button in a familiar place to do some predetermined thing, but much slower when the location of that button changes/gets hidden under submenus/I'm not familiar with an app or website.
On Android I constantly use the search function of the settings menu and I feel like this will be the golden UX going forward - a side by side UX + NL interface. So I can ask "how do I add a photo" and from there I get taken to the right place and can continue to add multiple photos in one go following the same pattern.
Though I suppose the nicer alternative is just "add all the photos I took near the waterfall from today".
The future is extremely dystopian and sad right now. The corporations are not going to use this the way you think they are. They are going to use it to maximize their profits, not help their customers.
> Bret is Co-Founder of Sierra. Most recently, he served as Co-CEO of Salesforce. Prior to Salesforce, Bret founded Quip and was CTO of Facebook. He started his career at Google, where he co-created Google Maps. Bret serves on the board of OpenAI.
1. There are already apps/websites as an alternative for CSAs. Most of the time I have to call someone its because I couldn't do what I wanted through those portals, so adding an AI agent to the chain is unlikely to prevent an immediate escalation to a human.
2. How much money are you really going to save this way? CSAs aren't high salary employees. Sure you might need a bunch of them, but we've already seen that brand loyalty erodes quickly when you remove the human touch. United/Spirit airlines offer opposing views on the cut your way to profitability perspective.
3. "Pay only for good outcomes" isn't going to last.
4. Are agents good enough to even do this? Yes, the cherrypicked examples sound good, but... I just know how well coding agents really work and my only experiences with voice agents in the wild have been very poor so far.
This is funding for established tech businessmen; what the business claims to do doesn't matter beyond having "AI" in it.
A lot of money. Managing a large group of people needs structure. It comes with tons of headaches and cost.
In terms of the investment:
It's Bret Taylor who has one of the most impressive background in tech. He can raise any amount he wants. VC bets on the person, not the business.
If Bret Taylor allowed me to invest, I would have invested too.
1. most people (average, non-tech people) reach for the phone to call in for easily solvable problems. Plus, if the agent is integrated deep enough & has tools to interact with crms, you can raise the ceiling on the types of problems it can solve.
You're trying to avoid the bad customer experience of human 1 reading off their script, then they transfer you to some other department who may or may not know how to solve your problem, and the entire interaction cost the company way more than the value created, so the company is disincentivized to help customers.
2. All the companies in this space start with the outsourced BPO market for cx (multi billion market still) but the next market is going to be in revenue generation and churn prevention at scale, i.e. how do you proactively avoid customer issues, how do you upsell and generate revenue instead of reducing cost, how do you keep customers happy?
3. I think more companies will pivot to outcome based pricing on the contrary, makes it so much more measurable than seat-based and protects margins better than usage based. Plus cx is one of the few industries with very well known metrics
4. Kind of? Most companies in this space don't use native voice models which are noticeably dumber, they use transcription + a stronger text model + TTS. The majority of customers can be handled with the latest SOTA text model and you need smart context engineering to handle the long tail of more complicated asks
3. Cool so the user didn’t indicate if they were satisfied. What then?
4. You can’t use a SOTA model right now for reasoning, there’s too much latency for a conversation. So you’re either using an older, but significantly less capable model, or you’re paying out the nose for fast mode. If the former then you can’t trust the agent to do the right thing (see points 1&2). If the latter, there’s no cost savings over a human. So which is it?
3 depends on how companies want to measure it, but lack of user submitting satisfaction score is not a good thing
you can use a model w/o reasoning, + use various tricks to simulate low latency
You didn't address my concern, non-reasoning models are so, so variable in their output.
2. another part of the value prop of these companies is figuring out how to construct the proper harness to take advantage of the lower latency of faster models while shoring up the weaker intelligence, how you blend deterministic and non-deterministic behaviors, compliance etc.
its a hard problem which is why f500 is willing to pay up
Ideally, businesses will escalate to an empowered human for all undefined parts of the flowchart. In practice, I truly hope it will be better than the current pre-recorded phone tree system that leads to a human following a script.
I personally only call support because a fix is not available through an organization's website.
They seem to be a "for pricing, let's go play C-level golf" type of company.
For myself - and admittedly maybe I’m just far out on the long tail of customers - I think these need to be treated like self driving cars, where 98% of the way there just isn’t good enough to cut it for me.
I think of support channels are just there to deflect customers and not really support anything. An AI bot will have infinite patience for that kind of interaction. Empowerment is never part of the equation.
> Ensure you only pay for the value Sierra delivers with outcome-based pricing.
Yeah... that won't last.
That said, my life hack for these things to get escalated to a human is to just keep saying or typing curse words. Usually that triggers a "connect to human" flow. I can't promise it will always work, but I can say it has worked every time I have tried it.
The tree is structured and gives me an immediate sense of how to map my task to the support offering. If I’m calling, I probably have an issue that I can’t self-serve resolve via the customer portal or whatever, so walking the tree lets me get an idea of who can help.
The “voice assistant” gives me no sense of what the system is capable of or how to take advantage of those capabilities. So I’m left guessing at phrases or functions based off of the assumption that there’s still some kind of tree-like structure that’s been abstracted away. Same outcome, more cognitive overhead, plus I usually have to shout in my best William … Shatner … impression to get it to understand me.
E.g. instead of waiting for the IVR tree to be read out to find out you needed to press 4 for the shipping department the AI asks "Please state the department you wish to connect to or reason for calling" and you just say "shipping" (or however much of a life story you want to give it) and it's the call system's job to figure out where in the menu that is instead. For repeat calls once you know its AI call routing you can just say "shipping" right as the call starts, the same as you'd known press "4" before the 2nd time around an IVR tree, except you don't have to remember the random digits.
but even a simple impl to answer questions can knock out like 50% of callers who are tech-illiterate at 100x cheaper cost, it's just strictly better economics and better for those customers
their moat is distribution
obviously the product needs to deliver and nrr needs to be good in the long run
It's a moat from a defensive perspective. It's a firehose from an offensive one. Outside state capture, most moats are both.
It is trust.
Everyone in the valley knows Bret Taylor and will back any project he does, even if the product has no distribution.
The same way everyone in the valley knows Naval Ravikant for example, angels and VCs will back any project he does even if his product has no distribution.
> OpenAI has appointed Paul M. Nakasone, a retired general of the US Army and a former head of the National Security Agency (NSA), to its board of directors, the company announced on Thursday.
and the money quote:
> “Artificial intelligence has the potential to have huge positive impacts on people’s lives, but it can only meet this potential if these innovations are securely built and deployed,“ board chair Bret Taylor said in a statement. “General Nakasone’s unparalleled experience in areas like cybersecurity will help guide OpenAI in achieving its mission of ensuring artificial general intelligence benefits all of humanity.”
[1] https://old.reddit.com/r/anime_titties/comments/1dh4wx4/form...
Well worth a read even if you are generally anti-AI.
Their implementation is rather cumbersome, requiring implementation fees and AI configuration that is rather bespoke to Sierra. Anyone rolling off of Sierra will find there is nothing they can take with them.
In general, I think CX ought to disappear as a vertical in an AI world. If I'm talking to a product AI and need support, why should I switch to another AI to do that? Even if that second AI is invoked by the first as a tool, how much am I gaining?
Interestingly, the first and best implementers of AI support so far have been at companies that roll their own.
There is nothing unique to CX about AI, as far as I can tell. Sierra is still just the same AI infra people are putting in products. Granted, you can make good money positioning yourself this way, but I expect on some time horizon they will need to reposition.
I don't know much about their product offerings, but I was doing some speech-to-text work and came across https://research.sierra.ai/mubench/ for comparing current models. It felt fairly thoughtful, particularly in regards to coming up with better benchmarking metrics than word error rate.
Had to check my assumptions though so I looked up what the lower end of GDP for a country is and sure enough they have American Samoa, Dominica, and Tonga beat. Now that money is probably meant to last 16 months so its not quite apples to apples but kind of wild regardless.
Ref: https://en.wikipedia.org/wiki/List_of_countries_by_GDP_(nomi...
You are basically comparing a nation's actual hard work and output to a speculative pile of cash. Even comparing corporate revenue to GDP is frowned upon by economists. A company's revenue includes the cost of all the inputs it bought from other companies (such as servers, electricity, and software licenses). GDP, by definition, strips out intermediate costs to avoid double-counting and measures only final value added.
You're also picking nations that do not produce much or have a very low population (Tonga only has 100k people), which pulls down the GDP due to how it's calculated.
I recently had the best customer support experience I've ever had, and it was an AI chatbot, helping my in ways I would not expect from a first-line support human.
Recently I opened the chatbot window on the site of a supplier of electric vehicle chargers, and asked some obscure things about how I could access and modify the charger in a non-standard way. I explained I was doing some R&D testing of a novel way of using the chargers, and I would need to temporarily reconfigure the public unit we had already installed in an unusual, non-standard way, rewiring its network to intercept traffic via transparent proxy, make it behave a little differently than it normally does, and that I didn't have the credentials.
I was expecting the first-line person or bot on the chat to be unable to handle my non-standard request. What I was hoping for was this would start the process of getting me to someone who could help. I expected to take a few days or weeks to reach someone appropriate in the company, perhaps a sympathetic technician who could understand what I was trying to do.
To my amazement, in the first response it one-shotted a nearly complete answer to every part of my question, including the detailed and implied parts I'd left out for a later interaction. It figured out what I was doing and what that needed, told me exactly what to do and why, and showed me that it understood every aspect of my request. There were things in this I doubt anyone ever asked about before.
Because the answer so complete I didn't really need more. I asked it some things, e.g. about electrical safety, just to confirm my understanding of the one-shotted answer. In every reply, the chatbot was insightful, informative, helpful and correct, and I didn't need to explain things twice, or wait on hold, or be transferred to anyone else.
I don't know what service they were using, but whoever implemented it, hats off to them for using good technology and providing it with high quality data about the products they are supporting.
Honestly, if AI customer support calls or chats can be made that good consistently, that sounds great.
(That was a stark contrast to my awful expereinces with support at banks, where many humans made mistake after mistake, contradicted each other, and seemed to struggle to understand simple things. Particularly the one where a bank required me to fill out a form with basic company info, then it took 14 hours of phone calls and hold time until I got to the one person who understood straight away that the entries on the form were correct, had to be that way, and they could tick the "it's done" box. My execellent AI bot experience suggests a good one will always understand things like that on the first iteration, as if they are the best-trained and best-informed humans available. But other, worse AI bot experiences suggests the 14 hour calls would become infinitely long, never able to resolve the problem, leading to things like unnecessary account closures and locked funds needing a court to resolve.)
I'd also love to see another point-and-click KQ game in the style of KQ 5/6 - as they are two of my favorite games of all time.
I don't think these games are better than the modern ones. But I wonder if the next generation's memories will be as strong.
We clearly do not live in the same universe.
As an aside, my favorite Sierra Entertainment logo version is probably the 1983-1993 version [1]. I think the design still holds up even today.
[0] https://velvetshark.com/ai-company-logos-that-look-like-butt...
There are 26 letters and millions of words; people should choose other ones.
1. https://preview.redd.it/remember-sierra-games-1979-2008-they...
One of the most beautiful game logos, going back to the early nineties.
https://sierrachest.com/gfx/Publications/IA/IA_8_1/023_Inter...
Since I also barely spoke English at the time, I got stuck in the game itself pretty soon anyway. Didn't manage to figure out how to say some things the right way. "Ken sent me" is the last thing I remember from it... and I never had any idea that the game was rather dirty until much later.
EDIT: holy shit I stand corrected: https://en.wikipedia.org/wiki/Lode_Runner
Gabriel Knight was awesome, I'd love a new one.
But there was one idea in QfG that I wish more games would use. Namely, designing three different solutions for every problem the player is facing. This idea works so well to create a sense of possibility in a game, I don't know why it got forgotten.