The challenging economics of making AI actually work today in the real world in B2B — and why the hottest job in enterprise might be a luxury SMBs can’t afford.
Why SaaStr’s Own AIs for Sales + GTM Actually Work (And Most Don’t Today)
Before we dive into the FDE discussion, let’s start with a reality check from our own experience. SaaStr’s AI SDR is the #1 performing AI with our outbound vendor. Our own SaaStr.AI has helped over 40,000 founders with their B2B questions. Why do our AIs actually deliver results while most companies struggle?
Because we’ve trained them. A lot. Far more than 95% of SMBs will ever train their AIs. And not just weeks of training up front, but daily training every day thereafter.
Can most small businesses really do this?
Most SMBs don’t have IT people, or extra technical resources. They’re running lean, focused on core business operations, and hoping AI will “just work” out of the box. At SaaStr, we are our own forward deployed engineers. But that’s not practical for 98% of SMBs.
This is the dirty secret of AI success that nobody talks about: the companies winning with AI aren’t winning just because they bought better tools. They’re winning because they invested in making those tools actually work for their specific use case. And that brings us to the heart of today’s discussion — the role that’s become essential for enterprise AI success, but might be economically impossible for the businesses that need it most. At least for now, at least for today.
The TL;DR
Forward Deployed Engineers (FDEs) have become the secret weapon of enterprise AI success, commanding $135K-$200K+ salaries to embed with customers and turn AI promises into production reality. Forward Deployed Engineers work directly with customers owning Gen AI strategy and implementation, with responsibilities similar to those of a hands-on AI startup CTO, while the estimated salary range is $135,000 – $200,000/year. But while FDEs are scaling enterprise AI deployments at companies from Palantir to OpenAI to newcomers like Artisan and Clay and Qualified (albeit sometimes with different names), the economics may not work for SMBs. The very businesses that need AI transformation most may be priced out of the implementation model that actually delivers results.
Even basic deployments often come up $25k+ up front charges, $60k+ annual deal sizes, and a month or two of “forward deployed” help getting the models and apps trained and going,
What Forward Deployed Engineers Actually Do (And Why They’re Different)
Forget what you think you know about solutions engineers or technical consultants. Forward deployed engineers have become one of the most strategic assets in enterprise AI companies, and here’s why they’re fundamentally different:
FDEs don’t just implement — they build. As an FDE, you partner closely with customers who are using your company’s products, working side-by-side to help them maximize the value they derive from your platform. Need to optimize a model’s architecture for performance? Develop a speech-to-text pipeline? Consult on a CI/CD pipeline for a novel AI system? They’re writing production code, not just configuring existing systems.
They embed with customers, not just visit them. Forward Deployed AI Engineers work directly with customers owning Gen AI strategy and implementation. On a daily basis, you will build end-to-end workflows, take them to production, and solve real world problems at the largest scale. This isn’t about quarterly business reviews — it’s about being in the trenches daily.
They own outcomes, not features. B2B businesses are not delivering software, but outcomes, and “how you integrate, embed, and operate becomes the moat.” The best FDEs measure success by customer business metrics, not technical milestones.
Think of them as “technical co-founders for your AI projects,” as one Baseten FDE put it. I like to think of my role as being akin to a technical co-founder for our customers’ AI projects.
The Economic Reality: Why Enterprise AI Needs FDEs
The dirty secret of enterprise AI? Enterprises buying AI are like your grandma getting an iPhone: they want to use it, but they need you to set it up.
Here’s what the market has discovered: tooling alone doesn’t transform businesses. Just examine the products and technical docs put out by the leading AI application companies. They have more in common than their landing pages indicate. Much of the product differentiation comes from how the same underlying technology is implemented and applied differently across customer sets.
The numbers tell the story:
- 22 of the 311 open roles on OpenAI’s career page right now fall into these categories — that’s 7% of all hiring focused on forward deployment
- Companies report faster implementation cycles, higher adoption rates, and measurable productivity gains with embedded engineers
- Hands-on support can also enable young companies to accept larger contracts, potentially powering faster and more durable topline growth than lighter-touch approaches
But here’s the catch: Individual deployments can have terrible margins, because it’s really R&D, not COGS. The customer implementation is primarily an opportunity to build and learn, rather than harvest short-term cash.
The SMB Challenge: Where the Model Breaks Down
Now let’s talk about the tougher challenge: SMB AI training. Listen, if you can train an AI, it’s easy. HubSpot just put out a new report on AI with SMBs. It found 69% of VC-backed start-ups have an AI team to do this. Most SMBs don’t have a team though.
This perfectly captures the core problem. While enterprise customers can justify $200K+ FDEs because they’re deploying AI across thousands of employees with million-dollar+ contracts, SMBs face brutal unit economics:
The Training Problem.
SMBs can’t access the benefits of FDEs:
- They need AI to compete with larger companies
- AI systems learn from human-fed data sets, including the questions and prompts that users type in. SMBs don’t have the skills or team to leverage this.
- Training is expensive. 55% of SMBs reported that the most significant barrier to the adoption of AI-enabled tools was cost
The Economics Don’t Work — Today. Here’s the brutal math: if an FDE costs $200K annually and can handle 3-5 enterprise accounts, that’s $40K-$67K per customer per year just in engineering costs. Add travel, overhead, and margin, and you’re looking at $75K+ annual cost per deployment.
Most SMBs can’t justify that math. The cost of the programs, their integration, and the training of employees are all concerns. When your entire company has 20-50 employees and $2M-10M in revenue, spending $75K+ on AI implementation represents 1-4% of total revenue — before you even buy the actual AI tools.
What SMBs Actually Get Instead
So what happens when the FDE model doesn’t work for SMBs? They get what one industry observer called “sparkling Sales Engineering” — the cargo cult version of forward deployment without the commitment or results.
Slapping a title on your field team because it sounds cool is one thing – building a truly “Forward Deployed” culture is another. It doesn’t work without complete commitment to the benefits and costs, and unless you’re willing to accept all of them, you’re engaging in thin imitation – just sparkling Sales Engineering.
Instead of embedded engineers, SMBs typically get:
- Pre-packaged solutions with limited customization
- Remote support instead of on-site implementation
- Training sessions instead of embedded knowledge transfer
- Documentation instead of hands-on problem-solving
As one investor noted: it may well be it takes a year or two longer because if you think about even the HubSpot journey is that what tends to happen is the big companies with loads of money fart around kind of mentally defining these apps and figuring out what it should be cuz they can afford to right and then of the features lock in the SMB guys go, “Oh, we’d like that.”
The Pre-Baked AI Solution: Hope or Hype?
The promise for SMBs is that AI will become more “pre-baked” .
This approach has some promise: for things like phone answering, simple order dispatch, that there’ll be a whole whole bunch of pre-anned, pre-baked, this is how it works, Mr. SMB. Just turn it on and you two can sound like a big call. So I’m I’m optimistic. It won’t be trained on a company by company level, but I think it will deliver big ass value.
But we need to be realistic about the limitations. Pre-baked solutions work for:
- Standard workflows (customer service, basic automation)
- Common use cases (email marketing, scheduling)
- Generic integrations (CRM, accounting software)
They struggle with:
- Custom business processes
- Industry-specific requirements
- Complex data integrations
- Unique competitive advantages
What This Means for the Market
The FDE model is creating a two-tiered AI economy:
Tier 1: Enterprise (The FDE Advantage)
- Custom AI implementations
- Embedded engineering support
- Business-specific training and optimization
- Competitive moats through deep integration
- Higher success rates and ROI
Tier 2: SMB (The Self-Service Reality)
- Pre-packaged solutions
- Self-implementation with documentation
- Generic training and support
- Lower barriers to entry but also lower differentiation
- Higher failure rates but lower costs
The question is whether this gap will narrow or widen. As copilots and agents that generate bespoke software become more ingrained in daily workflows, many startups will need to embrace a more tailored approach or face the prospect of a customer churning to build something themselves.
The Bottom Line for Founders
If you’re building AI for enterprise, FDEs aren’t optional — they’re table stakes. The more durable and integrated models consistently have some form of hands-on support, such as a forward deployed or solutions engineer.
But if SMBs are your target market, you need a different playbook:
For SMB-focused AI startups:
- Design for self-implementation from day one
- Create industry-specific templates rather than custom builds
- Invest heavily in onboarding UX and automated training
- Build community-driven support instead of 1:1 relationships
For enterprises considering the FDE model:
- Budget for the real costs — it’s not just salary
- Commit fully or don’t do it at all
- Measure business outcomes, not technical metrics
- Plan for scaling beyond individual deployments
The Forward Deployed Engineer isn’t just a job title — it’s a strategic choice about how you deliver AI value. Make sure you’re choosing the right model for your market reality.
