Everyone talks about AI taking over sales. They show you a slick demo where a chatbot writes a perfect email. It’s a nice story, but it misses the point.
The real problem in technical sales isn’t writing emails. It’s data.
A Sales Engineer (SE) is supposed to be a translator. They stand between the people who speak in tolerances and specs (the engineers) and the people who speak in money and risk (the buyers). It’s a hard job, and the current system makes it worse. They spend two-thirds of their time on administrative work, such as looking up old specs, digging through CRM notes, and typing up RFP responses [1]. This is a system bug. It forces your best people to act like slow, expensive robots.
AI promises to fix this. But if you try to plug a general-purpose AI into your sales process, it will fail. Why? Because the AI doesn’t know your business. It doesn’t know the difference between your sensor’s MTBF and your competitor’s. It doesn’t know the subtle pain points of the vertical markets you serve.
The problem isn’t the AI model. The problem is the input.
The First Step is Not AI
The polite thing to say is that you need a “data strategy.” What that really means is that most companies, especially in specialised industrial B2B, have no good data at all. They have a mess. They have a pile of old documents, tribal knowledge, and half-written notes. It’s a goldmine, sure, but it’s a goldmine you can’t dig in.
So, the first thing you build isn’t an AI. It’s a data machine.
This is the unglamorous, human work that makes the AI possible. You have to map out how to get the data. This might mean:
- Writing things down: Formalising the knowledge that only lives in the head of your senior engineer.
- Scraping and connecting: Building small tools to pull in market data or connect old systems.
- Changing the product or process: Making small tweaks to how your product logs information so the right data is captured at the source.
This is the hard part. This is where you build the context layer that turns a generic LLM into a specialised tool. Without this human-driven data fix, your AI will be generic and useless. The real data science challenge is not the algorithm; it’s the data pipeline.
To get this right, you have to understand the specific trade-offs and pain points in your customer’s world. This is what separates a generic sales pitch from a winning one. You need to know the market better than the customer knows themselves. Experts who focus on this – on understanding the vertical markets for technical components or subassemblies – are the ones who can help you build this context layer.
The Augmentation Hack
Once you have the data, the AI becomes a powerful tool for augmentation, not replacement.
Think of the SE’s job as a piece of software. The AI is a compiler for the technical specs.
The RFP Bug:
RFPs are a considerable time sink. The AI can read the RFP, search your clean data machine (the graph database of specs and past answers), and write a 90% complete draft. The SE’s 20-hour task becomes a 2-hour review. The AI does the robot’s work.
The Translation Bug:
Engineers struggle to speak to the CEO. The AI acts as a real-time translator. You feed it a technical fact (“We improved the API latency by 40ms”) and tell it to output “VP of Operations.” It helps you to address the needs of different stakeholders and gives you, e.g., “Our system now processes orders faster, ensuring your assembly line never waits for data, potentially saving $10,000 per hour in idle time.”
The AI takes the technical truth and wraps it in the language of value. It gives the SE empathy at scale.
The Human Guardrail
The AI is fast, but it can also lie. We call this a “hallucination,” but in technical sales, it’s a disaster. An AI that invents a safety certification or exaggerates a load capacity can get you sued.
So, you can’t let the AI talk to the customer on its own. You need a Human-in-the-Loop.
The AI is an exoskeleton for the SE. It handles the heavy lifting of data retrieval and formatting. The SE handles the strategy, the validation, and the trust. The AI gives the SE back the time they need to be human – to build relationships, navigate politics, and close the deal.
The future of technical sales isn’t about replacing humans. It’s about giving humans the tools to process data at the speed of software, so they can focus on the one thing AI can’t do: being the human.
References
McKinsey & Company. An unconstrained future: How generative AI could reshape B2B sales. [URL: https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights/an-unconstrained-future-how-generative-ai-could-reshape-b2b-sales]