The challenge
MaxContact was launching Conversation Analytics, its AI-powered quality assurance solution, into an increasingly crowded AI market.
Capability alone would not set it apart, so Conversation Analytics needed a compelling point of difference and a short, engaging pre-demo story. That story had to work for Conversation Analytics on its own and as part of the wider MaxContact suite.
Delivering it consistently was the harder problem. Sales teams were already using Claude to tailor presentations from discovery notes, which made each pitch more relevant. But the early AI-led approach gave Claude too much freedom. It was deciding what each deck should say, and also how it should look: layouts, backgrounds, image use, typography, spacing and structure.
The sales logic was often good. The PowerPoint itself changed from one generation to the next, and important messages, particularly who MaxContact is and why buyers should choose it, could slip out along the way.
The challenge was keeping AI’s intelligence for the customer story, without letting it reinterpret the brand and design system every time.
Our solution
We started with the narrative.
We created a concise launch narrative that positioned Conversation Analytics around powerful technology backed by genuine partnership, differentiating MaxContact through simplicity, customer outcomes and hands-on support. We translated it into a repeatable pre-demo presentation built around three themes: performance, protection and productivity.
Then we treated the presentation problem as an AI engineering problem. We separated the work into two Claude Skills with distinct jobs. The Sales Deck Skill decides what each presentation should say: the meeting type, the right product route, the customer’s priorities, which capabilities and case studies matter, and the speaker notes. The PPT Creator Skill decides how it looks.
The guiding principle was simple: Claude chooses the content, and a deterministic builder owns the design. Claude produces a structured plan for each deck. The builder already knows every position, size, font, colour and background, and turns that plan into slides the same way every time.
Testing shaped the final build. An early version locked layouts by turning finished slides into images, which kept them consistent but made them hard to edit. So we locked the rules that construct each slide, and rebuilt the output as native, fully editable PowerPoint.
We built a library of standard layouts for the slides that matter most, including What We’ve Heard, Who Is MaxContact, four-slide Platform and Conversation Analytics stories, product slides with approved image treatments, case studies, a Performance Promise and Next Steps. Each demo topic follows a Tell, Show, Tell sequence, with a relevant case study straight after each topic, so proof lands next to the argument it supports. Customer-specific talking points and live demo actions sit in the speaker notes, which keeps the slides clean and on brand while still giving sellers what they need.
Throughout, the approved original Skill stayed the baseline. Only changes MaxContact explicitly asked for were made, and every version was checked for anything that had shifted unintentionally. That is the discipline that turns AI presentations into something a sales team can rely on.
The outcome
MaxContact now has a clear launch narrative, a consistent pre-demo story and a system that turns discovery notes into tailored, editable PowerPoint decks.
Every deck uses the same approved layouts, brand and core message. What changes is the content that matters to each prospect: their priorities, the capabilities that fit, the right proof and the next steps.
MaxContact’s sales team tested the system and the feedback was strongly positive. The final round of changes was minor, and MaxContact is ready to roll it out to the wider sales team as part of its sales enablement.
The team also has a shared vocabulary for its slide layouts and a clear process for future changes, so the system can grow without drifting from what was approved.
The result is a sales presentation system built on a simple division of labour. AI decides the customer story. The Skill controls the sales logic. The builder protects the design.
