Marketers around the globe are starting to assess whether agentic AI improves their output. The latest figures provide grounds for immediate concern.
TL;DR: retain agentic AI for internal tasks.
‘Fast adoption, weak accountability’: this seems to be the overarching theme of agentic AI’s use in marketing, according to Acalytica’s culmination of industry-wide research on the topic.
Evidentially, 51.1.% of marketing organisations were using such technologies by August 2026, about a 155% YoY increase. Roughly half of US agencies utilise agentic AI for execution.
The stark contrast is seen in firms’ discipline. A mere 6.5% benchmark AI investment against pre-defined metrics and 65.1% do not measure AI return, or know if anybody does. In Salesforce’s survey of over 4,000 marketers, only 13% of AI users run self-directed agents.
Around three-quarters of agents are used for reporting and analytics, while around half for building campaigns briefs or having a hand in content creation.
The reason why may point to data quality (or lack therefore). Again, just 13% of marketers see their first-party data as extensive or well structured. Efficiency gains are more greatly evidenced, reported at a third of firms, but it is still a minority figure.
Because of this, Gartner even expects over 40% of agentic AI projects to be cancelled by the end of next year.
Whether it is the lack of ROI, or stringent regulations – the EU’s AI Act Article 50 transparency duties can fine non-complaint firms up to €15 million or 3% of turnover – that contributes to this, agentic AI’s boom is looking rockier. It is likely a combination of both!
So, what does this mean for fund marketers?
As a compilation of multiple studies that use different samples and definitions, these figures are more an indication of agentic AI trends than hard fact. Yet they should spur fund marketers to assess if their AI projects are bearing fruit or crippling operations today.
Take stock of current metrics before scaling
Considering that reporting and analytics are prominent uses of AI, IR teams likely use it to summarise LP and prospect engagement (or plan to). If so, baseline hours and error rates should be recorded first, otherwise it is impossible to show the tech’s ROI.
Clean up your CRM data
Duplicate investor records and inconsistent consultant or allocator targets wreck the data processing work that an automated tool can help with. A simple data-clean is the cheapest way to improve workflows, and potentially returns, too.
Check against Article 50 transparency rules
Firms that serve EU investors should see if the AI Act’s duties apply to any content they produce using AI, while all draft letters, DDQ responses and emails need to have approvals logged for potential regulatory inspection.
Agentic AI does offer a realm of possibilities, and especially so if it is used for low-risk tasks such as internal reporting or list segmentation. Keeping investor-facing content separate is a good idea, while every marketing and IR function can determine if the hypeworthy technology actually leads to asset-raising activity once and for all.
Source
Acalytica, “State of Agentic Deployment in Marketing — October 2026”
