Key points
- Invoca's 2026 Broken Funnel Report found 44% of B2C marketers have conversational AI agents live in at least one channel, with another 27% testing them.
- Only 19% can reliably feed agent conversion data back to ad platforms, and just 26% can link agent chats to the originating campaign, keyword or ad.
- Only 27% make revenue or booked outcomes their main yardstick for agents; most lean on satisfaction, lead counts, volume or cost to serve.
- The survey covered 451 U.S. marketers at companies with 100 or more employees, in home services, healthcare, finance, auto and telecom.
Why it matters for marketers: Agents may be answering customers while ad platforms never learn which campaigns produced the sale, so budgets get optimized on incomplete data and agent ROI stays unproven.
Marketers are putting conversational AI agents in front of customers faster than they can measure what those agents deliver. According to Invoca's 2026 Broken Funnel Report, 44% of B2C marketers have an agent live in at least one channel, and 27% more are testing one. Yet only 19% can reliably return agent conversion data to advertising platforms.
The problem sits inside a wider attribution gap. Seventy-four percent of respondents cannot fully follow leads from first digital touch to conversion, and the biggest losses come at dealers, franchises and local locations, where 56% lose visibility. Text and chat and phone calls also break the trail for many. Agents add one more handoff to a path that was already hard to connect.
For marketing teams, this means success metrics may be misleading. Only 27% treat revenue or booked outcomes as their primary measure. Others rely on satisfaction, qualified leads, conversation volume, containment or cost to serve, which show whether an agent is busy or efficient but not whether it earns its keep. Only 27% also say their agents can use a customer's earlier digital journey, so a returning prospect may be treated as a stranger.
The report also points to AI answer engines as a lead source. Invoca says 44% of respondents rated them as their most qualified source, ahead of Google Local Services Ads at 40%. Meanwhile, 86% can identify at least some leads that began there.
The practical step is to audit whether agent conversations carry campaign, keyword or ad identifiers and whether outcomes flow back to media buying. Teams should also consider judging agents on booked revenue rather than activity alone.
Sources
This brief is our own summary. Read the full reporting at the source.