AI-Powered ABM Content for 57 Financial Services Accounts

We started by building a formula for balancing 80% account-agnostic with about 20% account-specific content. Our writers and designers created an 11-page whitepaper and a nine-slide playbook pegged to a unique thought leadership concept along with product details and differentiators. The account-specific sections sit in the middle of each piece, bookended by the approved material.

Once Samsung signed off on the templates, we turned those sections into placeholders built to receive copy programmatically.

The account research arrived as a dashboard export from Periphas and Ei Advisory, formatted for human readers. Specifically, it covered company backgrounds, key purchase drivers and dated account signals, spread across many tabs. First, we used a large language model (LLM) to clean and restructure that data into consistent fields for every company. Next, we loaded the cleaned records into a database that our custom RAG system could query.

From there, the RAG system did the heavy lifting. It read each company’s context, wrote up its top three priorities, paired each one with a Samsung solution and dropped the copy straight into the template.

Still, our team stayed in the loop the whole way. We hand-edited the first samples before scaling up, and we reviewed every draft before it reached the sales team.

The finished set gave the sales team a tailored pair of sales enablement assets for every account. Each whitepaper works as both a pre-read and a follow-up, built around one company’s goals. For example, Mastercard’s version focuses on diversifying its payment solutions, while Cigna’s centers on whole-person health. Meanwhile, each playbook anchors the meeting itself and walks through the account’s technology opportunity and tailored solutions.

We also wrote and designed the Knowledge Worker solution page on samsung.com, which tells the same story to every visitor.

57

Financial services accounts, each with a tailored whitepaper and playbook

114

Account-specific sales assets generated from one approved framework

80%

Standard content approved once and reused in every version

AI handled the repetitive tailoring, while our team set the framework, shaped the inputs and reviewed every draft. This approach also shapes our other AI-powered content systems and our ABM work for J.P. Morgan Payments. Four practices made the biggest difference:

• Lock the framework first, because every version inherits its quality.
• Clean the research before you generate anything, since the output can only be as specific as the inputs.
• Keep the product story in the standard content so every version tells it the same way.
• Hand-edit the first samples, then review every draft before it ships.