Right now the AI for revenue teams is being engineered for a specific kind of seller — someone with a manageable product catalog, digital-first buyers, and the luxury of time between buyer interactions. For that world, conversational intelligence, lead scoring, and predictive analytics are genuinely useful.
But that’s not the world a lot of sellers live in.
I started my career as an account manager selling into manufacturing, with goods like graphics chips and some of the first PCs. And I’ve spent the years since working alongside many other organizations that also sell tangible products. Medical device companies, CPG giants, beverage distributors, and industrial suppliers. Their reality looks nothing like a typical SaaS sales motion. And the gap between what AI promises them and what it delivers is getting wider.
Selling complex physical products isn’t simple
These organizations don’t sell four SKUs with a tidy feature comparison chart. They manage thousands of SKUs with fine variations between models. Their catalogs change constantly — new specs, new configurations, new competitive pricing. And the sales cycles range wildly.
One company might close a quick deal selling ingredients to a local bakery. Another manages a multi-year capital equipment contract with services attached. The range is enormous, and the margin for error across all of it is razor-thin. Getting a product spec wrong, missing a compliance requirement, or misquoting a certification doesn’t just stall a deal. In regulated industries, it creates liability.
Field-selling environments are unforgiving
These sellers aren’t behind a laptop with a calendar full of second chances. They’re in the field: going into hospitals, factories, and construction sites. They’re managing limited buyer attention, articulating price and value in mere minutes, and often doing it without reliable connectivity to pull up info or call HQ for support.
They’re stretched thin from travel and back-to-back meetings. They’re juggling service options, warranty terms, delivery timelines, and localized partner dynamics. And when the meeting ends, there’s often no option to follow up. You land it then and there or you don’t.
I was recently talking to a consumer goods company sales leader. Their reps walk into retail locations with a few minutes to convince a store buyer that their mayo deserves better shelf placement over a competitor’s. In that window, they need to know their margins, their promotional calendar, and their sell-through data versus every other brand fighting for that same space. And they need to read whether they’re pitching a premium grocer who cares about brand story or a discount chain that only wants to talk volume and price. Because those are completely different conversations about completely different priorities.
The diversity of buyers and buying motions come with challenges
But that CPG rep is just one example. Across every industry we work with, the buyer landscape is just as varied as the selling environment. Some deals go through regulated tender processes that take months. Others are quick contracts signed on the spot. Buying committees include technical evaluators, procurement teams, and end users — each needing different information at different moments. And in many industries, localized buyer networks mean one bad experience doesn’t just lose a single account; it ripples through the entire regional market.
When you have four SKUs, AI is a nice-to-have vitamin. When you operate in this world — thousands of SKUs, high-stakes field environments, and buyers who demand precision — AI is an essential painkiller.
Where the AI industry is getting it wrong
Most AI being shipped today wasn’t built for this scope. Not because the technology isn’t powerful enough, but because it isn’t grounded in your specific product knowledge, compliance rules, or field reality. A hallucinated product spec or a non-compliant claim creates consequences that go far beyond inconvenience — damaged trust, regulatory exposure, and lost credibility across an entire account.
When this happens, sellers do what I would have done early in my career; they stop trusting the tools. They check every output manually. Or they stop using them altogether. And the efficiency AI was supposed to deliver disappears.
Different selling realities demand different AI solutions
Instead, field sellers need AI that can navigate a large catalog of SKUs in real time — surfacing the right spec, competitive comparison, and pricing for the exact conversation they’re in. Not a generic recommendation. A verified one that reflects any recent product changes, so a seller never has to second-guess whether what they’re sharing is still accurate.
Field sellers need AI that understands the environment they’re operating in. That means helping sellers prep for a morning with a technical evaluator and an afternoon with a procurement lead, adapting to the context of each decision-maker. Sellers on the road don’t have the comfort of sitting down to write notes after every visit. AI has to capture next steps in the background.
Field sellers need AI that recognizes how a six-month regulated tender is a completely different motion than a same-day contract signed over a handshake. Effective, context-aware AI must not give one-size-fits all answers but adjust the content, messaging, and compliance guardrails accordingly.
And finally, these field sellers need AI that accounts for the full scope of what’s actually being sold — not just the product, but the service agreement, warranty, delivery timeline, and coordination with partners. I’ve seen this firsthand across every sales organization I’ve led. The deal doesn’t end at the signature. It’s what happens after that determines whether you have a lifetime customer or a one-time transaction.
Why reliability is the foundation of useful AI
AI that operates in this world — high-pressure, regulated, and exacting — can’t be built on generic models, with revenue leaders and their field sellers hoping for the best. It has to:
- Be grounded in governed, accurate, and up-to-date product knowledge
- Reflect seller behavior in the field
- Have compliance rules, local regulations, and cross-border controls built into its foundation
Trust isn’t an “add-on.” It’s the layer everything else has to be built on. Without it, AI stays on the sidelines — observing, suggesting, and summarizing — but never trusted enough to help in the deal moments that matter.
This isn’t a problem any one company solves overnight. But I believe revenue leaders need to demand a change from industry innovators. To not just ask “How fast can we ship AI for sellers?” But, “Have we built an AI that’s reliable enough for the sellers who need it most?”
For more on my team’s AI vision, check out a recent article from our VP of Global Revenue Enablement on how to make AI reliable and scalable for field-selling organizations.
Frequently asked questions
Most sales AI on the market today was designed for a pretty specific scenario — a seller with a small product catalog, digital-first buyers, and plenty of time between interactions. Field sellers live in a completely different reality. They’re walking into hospitals, factories, and retail locations with minutes to make their case, often without reliable internet connectivity. They’re managing thousands of SKUs, complex compliance requirements, and buyers who demand precision on the spot. Generic AI just isn’t built for that kind of pressure.
The consequences go way beyond a minor inconvenience. A hallucinated product spec, an outdated price, or a non-compliant claim can create real liability — especially in regulated industries like medical devices or manufacturing. It damages trust with the buyer, creates regulatory exposure, and can torpedo credibility across an entire account or region. And once sellers get burned by bad AI output, they stop trusting the tool altogether. They either check every answer manually or abandon it entirely, and all the efficiency AI promised disappears.
Field sellers need AI that’s grounded in their specific, up-to-date product knowledge — not generic recommendations. That means AI capable of:
- Navigating large catalogs in real time: Surfacing the right spec, competitive comparison, and pricing for the exact conversation they’re in.
- Adapting to different buyers and motions: Recognizing that a morning meeting with a technical evaluator requires completely different content than an afternoon with a procurement lead.
- Handling the full scope of a deal: Going beyond the product itself to account for service agreements, warranties, delivery timelines, and partner coordination.
The differences are massive. SaaS sellers typically work with a manageable number of products, tidy feature comparisons, and digital-first buying cycles. Organizations selling physical products — think medical devices, industrial equipment, consumer goods — manage thousands of SKUs with fine variations between models. Their catalogs change constantly with new specs, configurations, and competitive pricing. Sales cycles range from quick same-day contracts to multi-year regulated tenders. And the margin for error is razor-thin because getting a product spec wrong doesn’t just stall a deal — in regulated industries, it creates legal liability.
Stop leading with “How fast can we ship AI for sellers?” and start asking “Have we built an AI that’s reliable enough for the sellers who need it most?” That means demanding AI that is grounded in governed, accurate, and current product knowledge, reflects how sellers actually behave in the field, and has compliance rules, local regulations, and cross-border controls built into its foundation. Trust isn’t a nice-to-have add-on. It’s the layer that supports everything else — and without it, AI stays on the sidelines in the deal moments that matter.
