Arishi and Magnolia DXP Publish New Research on B2B Buyer Behaviour in the AI Era

We are delighted to announce the publication of a research spotlight on B2B buyer behaviour in the AI era, in partnership with Magnolia DXP.

Magnolia DXP is the Digital Experience Platform that global enterprises choose to grow faster and stay in control of every customer experience. Founded in Switzerland in 1997, Magnolia DXP goes well beyond content management: its composable platform brings together content, AI-driven visibility and optimisation, personalisation and experimentation, so brands can be found, trusted and chosen in the AI era, across every channel.

We are proud to be a Platinum Partner of Magnolia, having supported some of its industry-leading clients and co-marketed our companies at major international trade events. We are delighted to have the opportunity to publish this spotlight.

About the Research

The spotlight, entitled The AI-Era Customer Journey – B2B Edition: How Enterprise Buyers Use AI, and Why They Still Check the Source reports research commissioned by Magnolia DXP and conducted by Evidenza, an AI research platform that models audiences with synthetic respondents grounded in public data. It is published by Arishi in partnership and with the permission of Magnolia DXP.

The study reports on 1,000 synthetic respondents, AI-generated personas that model AI-using professionals who buy on behalf of their organisation across Europe, the United States, and APAC. They are modelled across five purchase categories: banking and insurance, pharma and health, durable goods, everyday retail, and travel and hospitality. The respondents are not real people, so the findings are modelled behaviour, not a poll.

The findings look at how enterprise buyers use AI while purchasing and why the supplier’s website still carries weight with the final purchasing decision.

AI Does the Groundwork, but the Website Is Where Buyers Check

Artificial intelligence has become a core utility for B2B buyers during early procurement stages. According to the research, 73% of enterprise buyers use AI for research and learning, 66% use it to answer specific technical questions, and 65% rely on it to compare competing options.

However, when discrepancies arise between AI’s guidance and a brand’s website, trust resides with the primary source. 57% of B2B buyers rely on the supplier’s website when it conflicts with AI guidance, compared to 17% who trust the AI tool. Only 4% rely entirely on AI.

Buyers visit supplier websites primarily to access authoritative information directly from the source (22%) and to verify detailed technical specifications and features (20%). Across the study, fewer than 2% of buyers stop at an AI answer – the supplier’s own website is the most common next step.

A Wider Evidence Stack and the Need for Accountability

Rather than replacing traditional channels, AI adoption encourages buyers to build a larger, more rigorous ‘evidence stack’. Synthetic B2B respondents in the study use search engines (78%), brand websites (75%), review platforms (61%), and analyst or trade publications (59%). A larger share of heavy AI users (daily or more) than light users (monthly or less) use analysts and trade press and email newsletters.

This caution is driven by organisational accountability. Enterprise purchases involve substantial budgets, regulatory oversight, and complex sign-off procedures. As one synthetic respondent highlighted, "Every order I sign for has my name on the audit trail, and I can't explain a machine's choice to a regulator."

Respondents described AI recommendations as worth researching when they were specific, verifiable and honest about limits: for example, citing companies of a similar size and sector, lining up against RFP criteria, or saying where a recommendation was weak.

Personalisation: Utility First

The study also provides findings on digital customer experience. B2B buyers welcome practical, context-driven personalisation – such as real-time answers to technical queries (scored +64 net helpful) and content tailored to stated project needs (scored +64 net helpful). Conversely, identity-based tactics – such as being greeted by name (−11) or unprompted tracking follow-ups (−13) – score negative.

The Arishi Perspective: Preparing for Governed, AI-Ready Content

In our view, if AI platforms source out-of-date specifications or conflicting prices from fragmented legacy systems, prospective buyers may eliminate a vendor before direct contact is ever made.

As a Magnolia Platinum Partner, Arishi helps complex enterprises turn fragmented data into a governed, single source of truth:

• Content Consolidation: Establishing unified, approved data layers for product specifications, regulatory standards, and pricing.

• Headless Architecture and API Integration: Ensuring that websites, customer portals, partner ecosystems, and external AI agents draw synchronously from the same master repository.

• Enterprise Governance and Audit Trails: Implementing granular approval workflows, access controls, and technical compliance to satisfy regulatory requirements.

• Multi-Market Synchronisation: Delivering localised digital experiences that stay aligned with core brand data across global jurisdictions.

Download the Full Research Spotlight

Gain complete access to the findings and our strategic playbook for managing enterprise content in the AI era. Complete the form below to receive your copy of The AI-Era Customer Journey (B2B Edition).