Why is support now a brand function?

When AI handles the routine — FAQs, order tracking, basic queries — the interactions that remain are the high-stakes, high-emotion ones. How a brand handles what doesn't resolve cleanly is now the most visible expression of its values. Support isn't overhead anymore; it's the brand's public face in moments of stress.

What does 'support is the new billboard' mean?

Billboards used to broadcast brand values at scale. Now, at AI scale, the support experience — especially escalations and edge cases — broadcasts what the brand actually values. Every unresolved ticket, every handoff from bot to human, every edge case that gets handled badly is seen, shared, and cited. Support is now brand performance at scale.

What are the Agentic Brand articles?

The Agentic Brand Thinking series is a 14-module framework that translates brand strategy into machine-readable operating logic — covering values, voice, proof, policies, metrics, governance, and identity — so AI systems can represent a brand accurately without constant human intervention.

Support is the new billboard

Journeys are compressing. AI assistants, recommendation engines, and zero-click results handle more of the decision-making before a human ever touches your brand. Marketing gets less contact surface than it used to.

The first real brand experience, for a growing share of customers, will happen when something goes wrong.

Lowe’s recently credited its “Mylow” AI assistant with a two-point lift in in-store satisfaction and doubled online conversion. The CEO framed it as “agentic commerce.” The implication underneath: buying gets easier, the human moment shifts later, and concentrates in exceptions.

Support is where those exceptions land. Where people decide whether you’re solid or whether they should warn others.

Why support became the primary brand surface

Three structural reasons, not one philosophical argument.

It’s often the first real human touchpoint.

Salesforce’s 2025 State of Service report shows AI currently handles 30% of service cases, projected to reach 50% by 2027, pushing human reps progressively toward complex and contested interactions. Gartner projects that by 2029, agentic AI will autonomously resolve around 80% of common service issues without human intervention.

What’s left skews toward exceptions, higher stakes, more emotion, more accountability. The first human interaction is disproportionately brand-defining precisely because it’s rare.

It’s always the first risk moment.

Returns, cancellations, billing disputes, delivery failures, warranty edge cases. In those moments, customers don’t need storytelling. They need a clear outcome, a clear timeline, and a sense they aren’t being played.

This is the proof surfaces argument made operational: risk removal and responsiveness are literally part of the interface now. If your policy is buried and your escalation path is unclear, the product experience is incomplete regardless of how the rest performs.

It’s increasingly the first proof moment

Support behavior is portable: screenshots, chat logs, review summaries, forum threads. Consistency and ownership become evidence. If answers vary by agent, shift, or channel, you don’t just create extra contacts, you create doubt that travels.

Summaries are built from the same proof surfaces. When they’re inconsistent, the system can’t cite them cleanly. Small contradictions become big promises when the journey compresses to one answer.

Automation and the human pull-back

Klarna ran one of the more public experiments here. They pushed hard on AI efficiency in service: 2.3 million conversations, resolution times down from 11 minutes to under 2, a 25% drop in repeat inquiries.

Then they reversed part of the posture. Customers should always have the option to speak to a real person. They started recruiting for it explicitly.

The two moves aren’t contradictory. Automation compresses the routine path. The human option holds the edge case, the moment where someone needs a person to take responsibility, make the call, and communicate it directly. Remove that option and you’ve optimized past the point of accountability.

What it looks like under pressure

Same policy range. Completely different brand.

Customer:

“My package says delivered. It’s not here. I need this tomorrow.”

Response A, classic

“Thanks for contacting support. We’re sorry for the inconvenience. Please allow 3–5 business days while we investigate with the carrier.”

No ownership. No options. No meaningful timeline. The inconvenience is acknowledged; the problem is not.

Response B, brand under pressure

“You’re right to flag this. ‘Delivered’ without delivery is not acceptable. I’m owning this now. Two options: We reship today with express, or we refund immediately. If you prefer reship, reply with your preferred delivery window and I’ll confirm within 30 minutes.”

Ownership named. Options clear. Timeline specific. Control returned to the customer.

The difference isn’t voice guidelines or tone workshops. It’s whether the org has decided what to do when the promise meets reality under pressure. That’s a governance question dressed as a support question.

A pattern that shows up fast

Most brand people have never read their own support tickets.

Pull 10 from the last two weeks that include one of these words: cancel, refund, broken, not delivered, charged, doesn’t work, disappointed. Read them with three questions:

Did we restore control with clear options, clear timestamp?
Did we contradict ourselves: website or policy vs what support actually said?
Did we own the outcome, or hand the customer a process?

The pattern, when it shows up, is faster than any audit.

Where this sits in the series

Support is where the promise meets reality under stress. Article 06 asked whether you’re solving a real job. Article 07 asked whether you’re measuring what survives. Support is where those questions become visible.

Closing thought

It also marks the boundary of automation. Routine can be systematized. Trust moments need someone who can take responsibility and communicate it cleanly. That’s not a sentiment; it’s the constraint Klarna discovered in public.

Next: Brand has a missing time horizon. This is how you fill it.

Support spikes are usually symptoms of drift in claims, policy, product, or proof. The next module shifts from present-state signals to the time horizon brands need in order to stay coherent before pressure compounds.

Read

CX Dive: Lowe’s / Mylow performance data + “agentic commerce” framing
Salesforce: 7th State of Service Report
Gartner: Agentic AI will autonomously resolve 80% of common customer service issues without human intervention by 2029
Klarna Press Release: original AI results (2.3M conversations, 2 min resolution, 25% drop in repeat inquiries)
CX Dive: Klarna reversal reintroducing human support, admitting quality tradeoff
Earlier in this series: Proof is the new interface: What gets believed in AI-shaped journeys
Earlier in this series: What to measure when touchpoints collapse into one answer