01.2 Set the strategic frame

The future of AI in customer support for technical fields

By 2028, every manufacturer's technical library is an answer engine. The question is whose answer engine your customers use when they have a question about your products.

Chapter 1.25 min readSet the strategic frame

AI customer support is not a future innovation. It's a current category. The leaders for B2C and SaaS support (Fin, Decagon, Sierra) made AI the default expectation in those segments. Built-environment manufacturers are next, and the transition window is short.

The manufacturer library trajectory
2020 PDF directory "here are the manuals" 2022 Keyword search "search by part number" 2024 Search results 8 PDFs to scroll 2026 · NOW Answer engine "here is your answer" cited & version-aware 2028 Agent-readable built for AI consumers too From document directory, to answer engine, to agent-readable system
Every phase to date took about two years to reach the technical-fields majority. The next two are compressing.
01

Five forces compressing the timeline

1

Techs brought consumer AI to work

A 35-year-old installer who uses ChatGPT to write a permit at 7 a.m. has zero patience for a search page that returns 8 PDFs at 11 a.m. The default failure mode is no longer "call support." It's "switch to a public AI." Silent, and you never see it.

2

Public AI is guessing about your products

An installer asks ChatGPT how to wire your relay. Sometimes right, sometimes plausibly wrong, with no version awareness, no citation, and no refusal behavior. The wrong answer gets installed, and the warranty claim traces back to a query that never touched your library.

3

The agent-majority web has arrived

More AI agents read your technical library today than humans do: competitive-intel bots, customer assistants doing pre-purchase research, ChatGPT and Perplexity ingesting your pages. (Full treatment: 3.2.)

4

Veterans are aging out

Senior techs who built relationships with your support line over decades are retiring. The techs behind them came up on smartphones and conversational AI. Your library is now competing with their expectations.

5

Your competitors are deploying

Category leaders in security, fire, HVAC, and electrical are already piloting. Early deflection numbers are strong enough that "do nothing" stops being a viable board answer.

02

What customers expect by 2028

Forming now

A direct answer in a sentence. A citation to a page or video moment. Available 24/7. Aware of which revision the tech is on. Speaks the trade vocabulary. Works on a phone in the field.

Table stakes by 2028

All of the above, plus integration with ordering, certification training, warranty registration, and the product-update feed. The agent isn't a chat widget. It's your primary customer-facing surface.

03

What happens to brands that don't adapt

Questions answered without you

Questions about your products get answered, just not by you: ChatGPT, Perplexity, a competitor's portal. You lose the relationship at the moment of urgency, and the data that would have informed your roadmap leaks with it.

Brand erosion among brand-agnostic techs

Installers form brand opinions on library experience. A library that requires PDF-scrolling becomes a reason to recommend a different brand on the next bid. Invisible in your analytics. Visible 18 months later in declining specification rates.

Competitor interception

Public search and AI are advertiser-influenced. When a buyer asks about your part number, a competitor is bidding for that prompt. Every query that doesn't reach your library is monetized by the open market. (Modeled in 3.3.)

The compounding documentation gap

Without AI surfacing where your technical library is failing, the gaps stay invisible. Thirty unanswered queries a week compounds into a thousand-question shadow library nobody ever closes. By 2028 the early adopters have closed two years of documentation gaps; the waiters are two years behind.

The trajectory isn't a prediction. It's a description of what happened in adjacent categories (B2C, SaaS, retail) arriving on technical fields with a 24-month lag.

By 2028, the answer to every question about your product already exists. The only question is whether you own the narrative, and the surface where that conversation happens.

What good looks like
A CS director who has internalized the trajectory:
  • Has framed the project to leadership as catching the inflection, not chasing a feature
  • Can articulate the five forces and pick the two most acute for their company
  • Knows which trade competitors are already piloting
  • Has identified the documentation gap producing the most lost questions today
  • Has a credible answer to "what happens if we wait 18 months"
Next · Chapter 1.3
Five users, one product, thirty years
Get started

Want to see what category leaders are deploying?

We'll walk you through what manufacturers in your trade are piloting today, and where the strongest deflection is coming from.

Talk to us →