To view this email as a web page, click here

Our 20+ AI Agents and Their Moats: Real But Weak

Hey SaaStr Community,

We had a moment last month that should worry every AI agent startup. Or anyone building AI agents.

We were deploying Agentforce – our newest sales agent – to follow up with 1,000+ leads our team had ghosted. We needed to train it on our messaging, our tone, our process. The kind of thing that generally takes weeks or even months, with the help of a Forward Deployed Engineer.

Instead, we just copied the prompts from another outbound AI tool, changed a few details, and it just … worked.

Not “kind of worked.” Not “worked after heavy modification.” Just… worked. And very well.

Same quality. Same personalization. Same results (72% open rate, if you’re counting).

If you’re building or investing in AI agents, this should make you uncomfortable.

AI Agent Moats May Be Weaker Than Traditional SaaS. They Are For Us.

We now run 20+ AI agents across our entire GTM stack. Different vendors, different use cases, different price points. Some we coded ourselves. Some are specialized platforms. Some are from incumbents like Salesforce.

And here’s what six months of deploying agents has taught me: the differentiation between AI agents is narrower than you might think.

Yes, there are differences:

Some have better UX
Some integrate more natively with existing tools
Some have more specialized features
Some have better deliverability infrastructure

But the core intelligence? The actual AI work? If you train them well, the leaders all work well. They’re all good. They all use basically the same LLMs. And you can basically use the same prompts, and same training, with all similar AI agents.

AI AGENTS

SaaStr AI London is Starting Soon

SaaStr AI London is NEXT WEEK.

1-2 December:

Join us next Monday and Tuesday for 2 days of hands-on AI workshops & training, SaaStr's Who Do You Want to Meet Networking Program, Meet a VC and so much more!

Be in the front row for these top 5 sessions:

Tues at 2 pm: B2B in the AI Era: The Great, The Bad, and the Ugly with SaaStr's CEO & Founder
Tues at 3pm: From Startup Scrappiness to Enterprise Scale: Lessons on Selling, Hypergrowth, and Building Companies in the AI Era with OpenAI's GTM Leadership
Tues at 9am: Kick-off to SaaStr AI School: How to Use AI to Hyper-Customize Go-To-Market at Scale with SaaStr's CEO and Chief AI Officer
Tues at 1pm: The AI-Driven CRO: The Step-By-Step Playbook for Building AI-Powered GTM Teams with Personio's CRO
Tues at 10:30am: From $0 to $32B in 5 Years: Inside Wiz's Hypergrowth Secrets to Scale with Wiz's CMO

Some ticket categories have already SOLD OUT. Limited spots remain.

Grab $100 off final tickets until they're all gone for good.

HELLO LONDON

How Hard It Is to Be a Great VC? Vencap’s Real Hard Numbers from 1,900+ Funds

Every new VC fund pitches the same story: “We’re going to return 5x.”

Every LP hears it. Most nod politely. Few believe it.

And the data shows why they shouldn’t.

David Clark at VenCap International just dropped some of the most sobering VC performance data I’ve seen in years. He analyzed over 1,900 VC funds raised between 2000-2019 using Pitchbook data, and the results are brutal for anyone who thinks generating top-quartile returns is just about “working hard” and “adding value.”

Here’s what the data actually shows about the entire VC industry:

The Harsh Reality of VC Returns

The overall numbers are tough across VC funds:

22% of VC funds lose money. Not “underperform” — they literally return less than 1x. One in five funds handed back to LPs less than they put in.
64% of funds never hit 2x. Nearly two-thirds of all VC funds fail to even double capital. That’s the base case, not the exception.
Only 6% of funds return 5x or more. That target every GP pitches? It happens in 1 out of every 17 funds.
Just 2% hit 10x. The mega-outcome funds are a rounding error in the data.

Let me put this in perspective: A randomly selected VC fund is 3.7x more likely to lose money than return 5x.

That’s not a minor difference. That’s not noise in the data. That’s the fundamental reality of the asset class.

THE REALITY

This edition of the SaaStr Daily is sponsored in part by Wistia

Get more ROI from your team’s videos & webinars

Repurpose your videos and webinars into high-impact content with insights from marketing, content, and video production experts. Plus, learn how to produce content faster with AI, from clip creation for your social channels to dubbed content for your global audiences.

GET THE GUIDE

Is AI Support Perfect? No. But It’s A Lot Better Than No Support at All

Dear SaaStr: We’re thinking about deploying AI for customer support. Is it actually good enough yet, or should we wait?

Don’t wait.

AI support isn’t perfect. It still hallucinates sometimes. It has to be trained to be any good. It can’t handle every edge case. Your power users will still need to escalate to humans.

But here’s what I’ve learned deploying AI agents across SaaStr and watching hundreds of portfolio companies do the same:

AI support that’s available 24/7 and answers 60-70% of questions well is infinitely better than human support that doesn’t exist at all.

Think about it this way as a continuum … from Worst to Best:

No support at all.

Terrible. Your customers are stuck. They can’t get unstuck. They churn. You never even know why.

Support only from basic, untrained bots.

Still awful. You know the ones — they can barely parse what you’re asking, loop you through the same 3 FAQ articles, then give up. Frustration compounds. At least customers know someone theoretically works at your company. But that’s about it.

No live support — email only with 12+ hour response times.

Getting an email back tomorrow? A fundamentally suboptimal experience. Sure, you eventually get an answer. But your customer’s problem is happening right now. They’re blocked right now. They’re evaluating your competitor right now.

The data backs this up. When we looked across Gorgias’s 13,000 ecommerce customers, the average response time was an abysmal 12 hours. Even worse? 44% of those companies thought they provided excellent customer support — but only 1.3% actually did based on objective metrics.

Well-trained AI support available 24/7/365

Now we’re getting somewhere. Here’s what modern AI support actually delivers when properly implemented:

Instant response. Not in 12 hours. Not in 1 hour. In seconds. At 2am on Sunday. On Christmas. Always.
60-70% resolution rate for well-trained systems. Is that 100%? No. But it’s 10x better than the 6% resolution rate of traditional chatbots. And it’s handling the bulk of tier-1 questions that used to clog your support queue.
Consistent quality. AI doesn’t have bad days. It doesn’t get tired on the 47th ticket. It doesn’t ghost you because it’s overworked and looking for a new job.
Perfect memory. It knows every conversation this customer has ever had. Every ticket they’ve opened. Every feature they use. Every pain point they’ve mentioned.

The game-changer: You can actually provide support now.

NOBODY'S PERFECT

This edition of the SaaStr Daily is sponsored in part by Replit

Build Anything with Replit

Trusted by millions of users and Enterprise customers like Duolingo, Whoop, Coinbase, HubSpot, Google, and SaaStr's own CEO & Founder, Jason Lemkin.

Go to replit.com/signup and enter the code SAASTR to get two months free.

GET 2 MONTHS FREE REPLIT

Dear SaaStr: How Much Vacation is Normal for an Early Stage Startup CEO?

Hey SaaStr Community,

I’ve gotten worse at this. I think many of us have.

996 matters as founders. The pace of change, the scale of competition, is like nothing we’ve ever seen before now, in the Age of AI.

But I can tell you I never skimped on my own vacations as CEO — except in the first year of my second start-up, when for a while I was sort of on my own.

In fact, I probably took more true vacations as CEO than I do now because I needed them more, and I insisted everyone on the team take them.

But … the key is (1) having a strong management team, (2) having a right-hand person as part of that team, and (3) that you won’t really 100% disconnect.

Take ’em once you have a few good VPs, at least. They’ll know how to execute. They might even be OK with you not being around the office that week. You actually often see things seemingly work even better when you take a week off as CEO, if you have a very strong team under you.

And you’ll get time to think about the business more strategically than you can with 50 distractions a day.

OUT OF OFFICE

Dear SaaStr: How Long Should a CEO Continue Being the Head of Product?

Roughly — only for as long as it’s about a 10-hour a week job.

Or often, maybe only up to $3m-$5m in ARR or so. These days, especially, I see way too many B2B companies at $8m-$10m ARR or beyond without a true VP of Product.

Too many CEOs especially keep winging it for far too long.

Three mistakes I see so many B2B founders make:

First, they wait too long to hire a full-time “head of product” even after the product gets complex. Once your product has 50–100 workflows, dozens of use cases, 100s or 1000s of customers, etc. … even just 15-20+ really big customers, all using your product differently … it’s just too much to hack and juggle informally. My rough rule of thumb is somewhere between $3m and $5m in ARR in B2B most products just get too complicated to hack product anymore. You’ll ship better-prioritized features, with happier customers, with better trade-offs, with a full-time head of product. CEOs that hang on too long here start shipping features and products that are no longer force-ranked the way they should be. And the key is to hire a true head of product. Someone that can own it, and in the past, has owned a product. If you hire someone too junior, you get a project manager but not enough of an owner here.
And second, later, CEOs wait too long to give up on being the product visionary. Hopefully, the CEO is always the strategic visionary, forever. But at some point, the product itself gets so deep and rich, and the markets evolve so much, and the customer needs are so varied and nuanced once you dig in deep … that the founders’ original product vision becomes too simplistic, too dated and stale. Every 5–6 years or so, in most cases, everything in the market has changed, and the product has evolved so far that the Day 0 vision is very dated. At this stage, founders need to let go and have the team own the product vision — not them. The founders still are the heart of the corporate strategy. But the team is now spending 1000x more time with customers, with feature enhancements, with nuances on product lines with 1000s of features … you have to let the team define 95% of The Future. And your job becomes more to set the direction, but not provide the directions.
And finally, too many CEOs just hire too junior of a product person. That never works. You need someone seasoned enough to own the product and the product roadmap. Someone that has truly put features into production and managed some sort of team. Not someone just to make charts and hope someone in engineering might listen to them.

So bring in more help, and let it go a bit earlier than planned. It’s your idea, and your vision, but both of those need to evolve in the coming years.

HOW LONG

644 Emerson St. Suite 200, Palo Alto, CA, 94301

If you'd like to stop receiving the SaaStr Daily, click here

SaaStr

Our 20 AI Agents and Their Moats