Human analysts still handle 46% of brand interactions

ICUC has published data showing that human analysts handled 46% of 2.7 million brand interactions without automation support. The figures cover client work across the first five months of 2026.

The remaining 54% involved some form of automation, largely for support tasks such as classification, tagging, routing and initial processing before an analyst reviewed the outcome.

The data offers a snapshot of how brands are managing social community operations across platforms and sectors, at a time when many companies are weighing how far to rely on artificial intelligence in customer-facing work.

According to the findings, 90% of tracked interactions were in English, while the remaining 10% spanned 45 other languages. That points to the range of markets and languages community teams are expected to monitor.

Nicole van Zanten, Chief Growth Officer at ICUC, said brands should pay closer attention to the role of direct engagement in social channels. "We're entering a time in social marketing where every comment section is a focus group," she said.

She added that volume alone should not determine where people intervene. "The volume of conversations isn't what matters most. What really matters is understanding the conversations that require human intervention and why," van Zanten said.

Platform mix

Instagram accounted for just over half of all interactions in the dataset, making it the busiest platform overall. Direct messages on Instagram alone represented more than one in five community interactions, more than any other single message type.

That pattern suggests a heavy tilt toward private exchanges rather than public comment threads on the platform. By contrast, moderation pressure was higher elsewhere.

YouTube recorded the highest moderation rate of any channel in the dataset, at 14%, ahead of Facebook. Instagram public interactions were hidden at a much lower rate of 2.6%, despite the platform carrying the largest share of total volume.

The figures suggest that scale and moderation burden do not necessarily move together. A platform may host the most customer activity while generating fewer incidents that require content to be hidden.

Industry gaps

The data also showed marked differences by sector. Pharmaceutical brands required the highest level of moderation of any industry analysed, well above the average rate of around 4% of public interactions hidden across most sectors.

That pattern reflects the compliance, safety, misinformation and adverse-event reporting demands tied to regulated marketing. Consumer electronics and not-for-profit organisations also showed elevated moderation rates, at close to twice the overall average.

These differences suggest that moderation requirements are shaped less by broad industry labels than by a brand's regulatory exposure, reputational risk and public trust obligations. Sectors under tighter scrutiny appear to need more active oversight of public conversation.

Response times also varied sharply. Finance and insurance brands posted the fastest median response times of any industry measured, at more than four times faster than the overall benchmark of 8 hours and 56 minutes.

It was the only sector in the dataset to keep median response times below two hours across all message types. That sets it apart from other industries, where response speed varied more widely by platform and format.

Human role

The central finding is that human involvement still sits at the core of community management work, even where software is part of the process. In practice, automation appears to support workflow management rather than replace judgement in live customer conversations.

Van Zanten said that distinction matters as companies assess where to invest in social operations. "That's where experience, governance, and strong, experienced community management become differentiators for brands," she said.

She also argued that technology alone would not determine which teams perform best. "The organizations that win won't necessarily have the most advanced tech. They'll be the ones that combine AI with human expertise to respond faster, identify risks earlier, and create more meaningful customer relationships," van Zanten said.

The dataset was compiled from ICUC's community management systems and covers client work across pharmaceutical, finance and insurance, retail, restaurant, entertainment, consumer electronics and automotive sectors, among others.

The figures were aggregated and anonymised across the client base and are positioned as operational benchmarks for an area of marketing that has often lacked standard cross-industry measures.