Most manufacturers are chasing the visible 10% of the buying journey. The invisible 90% is where the leverage actually lives.
Walk into almost any manufacturing AI conversation today, and you’ll hear the same chorus: we’re exploring a configurator, evaluating a chatbot, looking at a recommendation engine for the website. Customer-facing, product-centric, digitally visible.
These are safe, defensible bets. They look good in a board deck and read well in a press release. And right next to them sits a bigger opportunity that most manufacturers walk straight past.
The instinct is understandable. AI vendors sell what’s easy to demo, and what’s easy to demo is what the customer sees. A configurator looks impressive when a prospect spins through 200 SKU variants in 90 seconds. A chatbot answers questions at 2 a.m. A recommendation engine surfaces accessories that the customer might have missed. Each earns its keep. The argument isn’t that they’re wrong. It’s that they address a smaller part of the revenue equation than most manufacturers realize.
Running the Math
A customer-facing tool lifts engagement among people who have already found your website. Useful, but the audience is self-selected, and the upside is bounded. A 10% conversion improvement gets you a 10% lift on a relatively small base. Real return, but it doesn’t change the trajectory of the business.
Now look at the other side. In consultative B2B manufacturing, lighting, electrical, and industrial channels, the rep is the product experience. A specifier may pull a cut sheet from your site. They may compare three manufacturers’ photometric files before lunch. The decision to specify, though, happens in a Tuesday morning meeting, on a project the agency principal has been working on for nine months, the last two spent outmaneuvering three competitors. The configurator never enters the room.
A senior rep covering a major metro is typically juggling 30 to 50 active opportunities at any given time — product knowledge across hundreds of SKUs, pricing across regions, lead times across factories, competitive dynamics that change quarterly.
The best reps win the way they always have: by walking into every meeting with depth. The ceiling on that kind of selling isn’t effort or talent. It’s hours in the day. Salesforce’s 2024 State of Sales report found that non-selling tasks consume 70% of a rep’s time — administration, meeting prep, data entry, and internal reporting. That’s the ceiling AI lifts.
In my experience working with rep networks, even a 15–25% improvement in effectiveness per interaction compounds fast. It shows up in faster proposals, fewer specification flips, better margin discipline, and tighter follow-up on stalled projects. Multiply that across every territory, every week, every quarter — and the numbers stop being interesting and start being structural. McKinsey’s research on gen AI in B2B sales backs this up: companies that empower sales teams through technology report consistent efficiency gains in the 10–15% range, with meeting prep tools alone freeing up more than 10% of seller time.
The customer-facing tool optimizes the visible 10% of the buying journey. The sales-facing tool optimizes the invisible 90.
Why This Is Easy to Miss
Three reasons, in roughly the order I see them.
Customer-facing AI is photogenic. It produces screenshots and case studies. Sales-facing AI lives in a CRM, an email thread, a call prep doc, and a margin analysis. None of that travels through a marketing channel.
Customer-facing AI doesn’t threaten anyone internally. The website is a low-stakes environment for experimentation. Sales-facing AI, done badly, looks like surveillance dressed up as enablement. Reps have learned to be skeptical of tools introduced from the top. The ones that stick are the ones that genuinely save them time.
The people who run digital transformation typically come from marketing, IT, or operations. They know the customer journey on a website. They don’t know what it feels like to sit in a specifier’s office at 4 p.m. on a Friday, trying to save a project that’s sliding sideways. People optimize what they understand.
What Sales-Facing AI Actually Does
Strip away the vendor language, and there are four real applications — all with measurable impact on commercial performance.
Preparation. Before a specifier meeting, every rep should know the last three projects from that firm, competitors considered, where things landed, and the current pipeline by application type. That information exists — scattered across CRM, project tracking, and email. AI assembles it in 90 seconds. No rep does this manually for every meeting. With AI, every rep does it for every meeting.
Response speed. A specifier sends a question on Wednesday afternoon. The rep is in another meeting. The clock starts. Whichever manufacturer answers first — with the most useful response — has a structural advantage on that project. AI drafts the response, pulls the relevant cut sheet, flags the right cross-reference, and surfaces pricing context. The rep edits and sends in five minutes instead of two days.
Consistency. The best rep in your network and the median rep are not equally good at quoting, follow-up, objection handling, or competitive positioning. They never will be. But AI can compress that gap. The median rep, equipped with the right tools, starts performing like the 70th percentile. Across a sales force, that shift is worth more than any single new hire.
Institutional memory. The lighting and electrical channel runs on relationships built across decades. When a senior rep retires, the manufacturer loses 25 years of project history, specifier relationships, and competitive intelligence. AI doesn’t replace that — but it captures and structures enough of it that the territory doesn’t start over when the rep does. Senior agency principals across much of the lighting and electrical channel are approaching, or past retirement age, and the succession challenge is real. This isn’t a hypothetical concern. It’s already happening in territories across the country.
None of this requires a moonshot. Most of it can be built on the data manufacturers already have, with tools that already exist.
The Adoption Problem Nobody Talks About
Customer-facing AI and sales-facing AI fail differently, and the failure modes have very different consequences.
Customer-facing AI fails quietly. The chatbot underperforms, engagement dips, the dashboard shows the dip, and someone schedules a working session to revise the prompts. The customer never knows it failed.
Sales-facing AI fails loudly. The rep tries the tool twice, finds it slower than what they were already doing, and stops opening it. Within three weeks, the deployment is dead. There’s no working session. There’s no second chance. Reps voted with their behavior, and the next time you ask them about AI, you’ll get a polite version of “we tried that.”
This is why the manufacturers who succeed with sales-facing AI treat the first 90 days of adoption as a separate discipline from the technology itself. The build is the easier half. The harder half is integration into daily rep behavior. The rule of thumb is simple: when a tool saves the rep meaningful time in the first week, it earns its way into the second month — and from there it becomes part of how they work.
Customer-facing AI can be bought and deployed. Sales-facing AI rewards a different posture: ongoing engagement with the field, calibration to actual workflow, and the patience to let it earn its place in how reps work. The leaders who stay close to the deployment are the ones whose deployments stick.
Where to Start
The contrarian position isn’t that customer-facing AI is wrong. It’s that sequence that matters.
Build rep capability first: deeper preparation, faster response, more consistent execution, and a way to hold institutional memory as senior reps move on. Then build toward the customer — with a much clearer sense of what they actually need to see, grounded in how your commercial process actually works.
The leaders who get the order right are rewarded twice. Better commercial performance now. Better customer-facing AI later, built on a clearer understanding of their own business.
The smartest commercial leaders I work with have stopped asking what AI can do for their customers. They’ve started asking what AI can do for the people who carry their brand into thousands of specifier offices a year.
That’s where the leverage is. That’s where the math works. That’s where the conversation should start.
