How Adomik proved that useful AI tools adopt themselves

Industry: B2B SaaS
Company Size: 51-200
Department: Company-Wide AI Adoption

Key Highlights

Instant adoption validation: Users alert the team within hours when agents or data sources are temporarily unavailable
Multi-team deployment: Customer Success, Sales, and Product teams using specialized agents daily
Champion-led governance: Weekly team meetings + bi-weekly company-wide showcases driving sustained engagement
Bottom-up philosophy: No mandates, no forced adoption, no obligation. Just tools that simplify work.

This story was written by Wesype, a Dust Platinum Partner with strong expertise in the AdTech industry. It shows how Adomik turned Dust into an essential daily tool by letting teams discover its value organically, without any obligation to use it. With Wesype's support, Adomik created agents so useful that adoption happened naturally and users notice immediately when they're unavailable.

About Adomik

Adomik is a Paris-based AdTech company providing a programmatic advertising analytics platform for publishers and advertising professionals. The platform helps advertising and media teams (publishers, broadcasters, ad platforms) make sense of their revenue data. Adomik connects to all their ad sources, unifies the data, and powers AI agents that answer performance questions, flag issues, and automate reporting workflows.

In a market where every major tech company is launching AI agents and MCP servers, Adomik recognized that AI adoption wasn't optional for the business. It was the new standard. The company is now developing its own customer-facing AI product called Jean-Pierre, alongside its internal AI transformation with Dust.

💡 Interested in learning more about how Dust can help your team? Reach out to our sales team.

Challenge: When data lives everywhere but insights live nowhere

The multi-platform juggling act

Before Dust, Adomik's teams faced a daily reality that many B2B SaaS companies know well: data scattered across multiple systems with no unified way to access it.

Customer Success managers needed to monitor client usage patterns, identify at-risk accounts, and prepare for reviews. Each task required connecting to different analytics platforms, generating reports, manually filtering by client portfolio, then cross-checking with HubSpot. The workflow was functional, but fragmented.

Sales teams experienced similar friction. Preparing for a client meeting meant pulling data from the CRM, reviewing call transcriptions, checking email threads, and synthesizing meeting summaries. All from different sources.

The competitive imperative

"We turned to Dust because in tech, all the major companies are launching their own MCP servers, moving towards GenAI, creating their agents. It proves this is not just a trend. It's the new standard." — Kim, Project Lead

The AdTech industry moves fast. When competitors started building AI capabilities, Adomik knew the question wasn't whether to adopt AI internally, but how quickly they could move.

Solution: Demonstrate value and let adoption follow

The philosophy that changed everything

Rather than imposing AI tools from the top down, Adomik chose a radically different path: build something genuinely useful and let teams discover it themselves. No mandates. No requirements. No obligation to use the tool.

"We chose not to impose AI but rather to integrate it where there's real, immediate, measurable value. If a tool is truly useful and simplifies life, teams adopt it naturally. There's no need to force it." — Kim, Project Lead

This bottom-up philosophy meant that every user who adopted Dust did so because they saw personal benefit, not because management told them to. The result? Authentic engagement instead of reluctant compliance.

Building a champion network

The deployment followed a deliberate approach:

Weekly champion meetings: One champion per team identified concrete needs and collaborated on agent development
Bi-weekly company showcases: New agents were presented to the entire organization, demonstrating tangible value
Centralized building, distributed usage: Kim built agents for business teams; Product team members created their own for specific needs

"We started with champions. Weekly meetings with a project lead in each team, focused on very concrete needs. Then we demonstrated value through more open meetings with the whole company." — Kim, Project Lead

The key distinction: champions were evangelists, not enforcers. Their role was to show what was possible, not to mandate adoption.

Specialized agents that solve real problems

Client Usage Agent (Customer Success)

The most-used agent at Adomik consolidates everything CS teams need: platform analytics, client usage patterns, power user identification, and portfolio filtering. What previously required navigating multiple analytics platforms now happens in a single Dust query.

"Previously, they had to connect to other analytics platforms, generate their reports, manually filter by client, filter on their portfolio, and cross-check with HubSpot. Now they don't have to. Everything is in Dust." — Kim, Project Lead

Risk Alerting Agent (Customer Success)

Connected to usage data, this agent sends automated Slack alerts when clients show risk indicators. The system monitors for usage drops and other warning signs, enabling proactive churn prevention.

Sales Preparation Agents (Sales)

Connected to CRM, call transcriptions, meeting summaries, and email, these agents give sales reps complete account recaps before every client meeting.

QA and FAQ Agents (Product)

The product team built multiple agents to automate manual, repetitive QA tasks. Each team member creating agents tailored to their specific workflows.

Designing for immediate usefulness

Kim's approach to agent design reinforces the bottom-up philosophy. Users simply launch the agent, sometimes with no input at all, and get results. For more complex agents, a brief disclaimer message appears at launch, providing just enough context without requiring formal training.

When the CS team initially tried to clone and customize agents for their own portfolios, they quickly realized effective customization requires prompt engineering skills. The solution: Kim rebuilt the Client Usage agent with built-in portfolio filtering, so one well-designed agent serves the entire team without modification.

Results: When users prove adoption by missing what's gone

The ultimate adoption test

Without access to deep usage analytics, Adomik discovered an unexpected way to measure success: what happens when agents become temporarily unavailable.

"Every time a data source isn't updated right away, or an agent is pulled from public use, I immediately get messages from my users. That proves they're using it. It's well adopted." — Kim, Project Lead

When Kim temporarily unpublishes an agent for improvements, or when a data source falls behind schedule, messages arrive within hours. These immediate reactions became the clearest signal that Dust had moved from optional tool to essential infrastructure.

Daily integration across teams

The proof is in the daily usage patterns:

Customer Success: Uses the Client Usage agent multiple times per week for portfolio monitoring and risk identification
Sales: Queries account recap agents before every client meeting
Product: Individual agents automate repetitive QA workflows
Kim: Uses Dust integration in everything she does, every day

"For me now, it's the essential tool to have every day. I have my Dust integration. It's really something I use in everything I do." — Kim, Project Lead

Based on Adomik's experience, agents that achieve sustained usage share common characteristics:

Immediate value: Users see results from day one
Minimal friction: Launch and get answers, no training required
Real problem-solving: Addresses actual daily pain points, not theoretical improvements
Visible demonstration: Regular showcases keep new capabilities on everyone's radar

What's next: Expanding the AI foundation

Building Jean-Pierre

Adomik isn't just using AI internally. They're building it into their product. The team is developing Jean-Pierre, a customer-facing AI assistant, alongside their own MCP server. The goal: eventually connect Adomik's MCP to Dust for seamless integration.

Maintaining momentum

The biggest challenge isn't launching AI. It's sustaining attention over time.

"The biggest challenge is to always give a little reminder and show that it's still there. Keeping the same momentum every week when there aren't always new things to say about Dust. That's the hard part." — Kim, Project Lead

Weekly champion meetings and bi-weekly company showcases keep AI visible and relevant, even during quieter periods. The cadence proves investment and commitment. And it works.

Interested in learning more about how Dust can help your team? Visit our solutions page or reach out to our sales team.