↓ Skip to main content
  1. Articles/

A Few Experts, Hundreds of Sales Engineers: Scaling Security Expertise With AI and Just-in-Time Enablement

▶ Present this article as 11 slides you can share, or browse all decks.

Working on something like this? Ask me. Questions about your setup, or anything this article didn't cover. No account needed, private by default.

Only I read it. It isn't published unless you tick the box, and your email is never shown or added to any list. Privacy

Specialists who join every call become the bottleneck. Specialists who help everyone else shine become the multiplier.

Executive summary
#

The business problem. In enterprise sales, the hard security questions often decide the deal: how is our data encrypted, who can see it, where does AI processing happen, what does the regulator say. The people who can answer them well are always too few. If every tough question ends in “let me bring in our expert”, a handful of calendars becomes the speed limit for the whole sales organisation, and the sales engineers never get the chance to own the answer themselves.

What works instead. A small group of specialists can support hundreds of sales engineers (SEs) on tough security questions, but not by being in all of their meetings. Four practices do the scaling:

  1. An AI skill that works the way the specialists do. It holds their knowledge, researches the way they research, and writes answers the way they would. SEs use it first.
  2. A champions team with a clear deal. SEs who want to grow in security get guidance on how the role helps their career, and clear expectations of what it takes.
  3. Writing first, call second. Whenever possible, the customer gets the answer in writing, with an offer to clarify on a call once they’ve read it.
  4. The SE presents, the specialists enable. The SE gets prepared for the session and supported during it. Nobody spoon-feeds, and nobody takes over the meeting.

The idea underneath. This is bite-sized, just-in-time learning with AI, and just-in-time enablement from people. The SE learns exactly what the deal in front of them needs, at the moment they need it, instead of sitting through training they’ll have forgotten by the time a customer asks.

The payoff for the business: expertise that grows with the organisation instead of with a few people’s calendars, SEs who own their customer conversations, and customers who get considered answers they can pass around internally.

The bottleneck nobody plans for
#

Specialist expertise usually scales in one of three ways, and none of them works for long.

  • The specialist joins every call. It feels like great service. In practice the specialists run out of hours, the SE turns into a meeting organiser, and the customer learns to ask for the specialist every time.
  • Write a wiki. Useful, but it goes stale, nobody reads it end to end, and it rarely answers the specific question a specific customer just asked.
  • Train everyone in security. Days of training months before anyone needs it. By the time a customer asks the question, the answer is long forgotten.

What works is the opposite of all three: help at the moment of need, tied to a real customer, small enough to absorb, delivered in a way that leaves the SE more capable than before.

1. An AI skill that works the way the specialists do
#

The first line of support isn’t a person. It’s an AI skill built for the SEs. It’s not a chatbot pointed at a pile of documents. It captures three things:

  • Knowledge. What the specialists know about the topics customers ask about, and the positions they’d take.
  • A way of researching. Where to look, which sources count, and how to check that an answer is still current, instead of answering from memory.
  • A way of working and formulating. How to structure an answer and which tone to use, including saying clearly what something does not do, and what “it depends” depends on.

An SE with a tough question starts there. They get a well-structured draft, built the way a specialist would approach it, and they learn something about the topic while reading it. When the skill isn’t enough, they reach out to the specialists, and that’s then a better conversation for both sides, because the groundwork is done.

For leaders, the point isn’t the technology. Your best people’s method, not only their knowledge, can now be packaged and put in front of everyone who needs it, at the moment they need it.

What to watch. The SE stays accountable for what goes to the customer, so a draft from the skill gets read and checked before it’s sent. Answers should point to their sources so they can be verified. And what goes into the skill, and what SEs put into it, has to follow your company’s rules for customer and confidential data.

2. A champions team with a clear deal
#

AI covers a lot, but some questions need a person who knows the topic and the customer. That’s what champions are for: SEs who build deeper security expertise and become the first port of call for their colleagues.

A champions programme lives or dies on one thing: does the champion know what’s in it for them, and what’s expected? Without that, it becomes extra work that nobody asked for and nobody recognises. So both sides are written down clearly:

  • How it helps their career. A recognised specialism, visibility beyond their own team, closer work with the specialists, and a track record of wins and enablement they can point to when they go for their next role.
  • What’s expected. What a champion does, how much of their time it takes, and how they keep their own knowledge current.

For managers, this is the cheapest lever in the model. The guidance costs almost nothing to write, and it’s the difference between a list of names and a team that actually carries load together.

3. Writing first, call second
#

When a tough question comes in, the instinct is to set up a call. Whenever possible, answer in writing first, and offer a call to clarify once the customer has read it.

Writing first is better for everyone involved:

  • The customer can read it at their own pace, check it with their own experts, and forward it internally. Security answers almost always end up in front of people who weren’t on the call.
  • The answer is more precise. Writing forces the exact wording that a security review needs, and it can be checked before it’s sent.
  • The call gets better. The customer arrives with specific follow-up questions instead of a first explanation, so the time goes to the parts that are genuinely unclear.
  • The answer lasts. A written answer can be reused by the next SE who meets the same question, and teaches them in the process. A call is gone when it ends.

4. The SE presents, the specialists enable
#

When a session with the customer is needed, the SE runs it. The specialists help them prepare beforehand and are there to support during the session. What they don’t do is spoon-feed every line or take over the meeting.

That’s a deliberate choice, and it can feel slower the first few times:

  • The customer relationship stays with the SE, who owns the account long after the session.
  • The SE’s credibility grows in front of their customer, and the whole team looks stronger for it.
  • Each session leaves the SE more capable, so the next similar meeting needs less support, or none.

The specialist in the room is the safety net, not the presenter, and steps in when the SE asks for it.

Just-in-time, not just-in-case
#

Put together, the four practices are one idea. Learning happens in small pieces, at the moment of need, on a real customer question:

  • Just-in-time learning with AI: the skill gives the SE an answer built the way a specialist would build it, when the question comes up.
  • Just-in-time enablement from people: champions and specialists prepare and back up the SE for the specific session in front of them.

Every question answered this way leaves something behind: a better-prepared SE, a written answer the next person can reuse, and a champion with one more win. It’s a team sport, and everyone on the team gets better at it.

What it means for leaders
#

The pattern isn’t specific to sales engineering or to security. It applies wherever scarce expertise sits in front of a large organisation: security architecture, privacy, legal, or any specialist team that’s always asked to “just join the call”.

Questions to ask your organisation:

  • Where is a small team of specialists the speed limit for many others?
  • Do your specialists mostly join meetings, or mostly make other people able to run them?
  • Is your specialists’ method captured anywhere an AI can use it, or only in their heads?
  • Do your champions know, in writing, what the role does for their career and what it expects of them?
  • When a hard question comes in, does the customer get a written answer they can pass around?

Decisions for leadership:

  1. Treat specialist time as a scarce resource: measure it by how many people it makes more capable, not by how many meetings it attends.
  2. Fund the capture of specialist method in AI skills, with clear rules for what data goes in.
  3. Give champions a written career deal and real recognition, not just a title.
  4. Set “the account team presents, the specialists support” as the default for customer sessions.
Working on something like this? Ask me. Questions about your setup, or anything this article didn't cover. No account needed, private by default.

Only I read it. It isn't published unless you tick the box, and your email is never shown or added to any list. Privacy

Kevin Keller
Author
Kevin Keller
Personal blog: the business case for secure AI and data platforms, and the architecture blueprints to build them. Snowflake-related articles explore the art of the possible and are not official Snowflake solutions or endorsed by Snowflake unless explicitly stated. Opinions are my own. Content is meant as educational inspiration, not production guidance.
Share this article