How Fluxym achieved AI adoption through a bottom-up approach, and is now ready to scale
- Industry
- Consulting
- Company Size
- 51-200
- Departments
- Company-Wide AI AdoptionKnowledge Management
Key Highlights
- Six months of bottom-up adoption with training, coaching, and proven use cases
- Widespread AI adoption achieved: Teams now proactively share agent experiences and best practices
- Now launching Phase 3: Dedicated project team for top-down deployment across regions
This story was written by Wesype, a Dust Platinum Partner with strong expertise in the IT services industry. It shows how Fluxym, a global IT services company, spent six months building organic AI adoption before transitioning to a structured, top-down deployment approach. With Wesype's support, Fluxym is now entering the next phase of its AI journey.
About Fluxym
An expert in Source-to-Pay since 2002, Fluxym supports Procurement and Finance departments in their supplier relationship digitalization projects. With over 500 projects completed worldwide, Fluxym offers a comprehensive range of services combining technical and functional implementation, as well as support.
Like many technology companies, Fluxym recognized that AI could transform their operations. But rather than mandating adoption from the top, they chose a different path: let people discover AI's value for themselves first, then structure the deployment once the foundation was solid.
Six months later, that bet has paid off.
The challenge: scattered AI usage
Before Dust, Fluxym's teams had developed individual AI habits that worked, but didn't scale:
- Consultants had accumulated extensive personal histories in ChatGPT, Claude, and Gemini
- Knowledge remained siloed in individual AI conversations rather than shared across teams
- Document retrieval from SharePoint required manual searching across multiple folders
- No way to leverage company data through AI tools.
For pre-sales teams facing RFPs with 150 or 200 questions, being able to leverage historical responses would make all the difference, but individual ChatGPT subscriptions couldn't access that data.
Phase 1-2: six months of bottom-up adoption
Centralized platform: bringing all AI tools under one roof
The first win was simple but essential: getting everyone onto a single platform. Dust replaced fragmented individual subscriptions with a unified interface where teams could access ChatGPT, Claude, and Gemini, all in their pro versions, all in one place.
This wasn't just about convenience. It meant prompts could be shared, data could be centralized, and usage patterns could be understood across the organization.
Training and personalized coaching
Rather than organization-wide workshops, Fluxym invested in one-to-one coaching sessions. The Dust administrator spent roughly one day per week helping team members:
- Understand how to leverage multiple AI models for different tasks
- Generate prompts via LLMs, then customize them for agents
- Create agents connected to SharePoint and Salesforce
The marketing team developed a distinctive workflow that leverages each model's strengths. They use ChatGPT for building outlines, researching ideas, and SEO optimization. Then they switch to Claude for the actual writing, taking advantage of its superior prose quality. This ability to combine models within the same platform produces more polished results than relying on any single tool.
"Almost everyone has become autonomous in their usage."— Gowtham Thevaraj, Dust Implementation Lead
Knowledge access use cases with proven impact
The first agents focused on a clear pain point: accessing company knowledge. General-purpose agents connected to SharePoint now provide company-wide access to documentation, eliminating manual folder navigation. Pre-sales agents leverage historical RFP responses to accelerate proposal completion.
These weren't theoretical benefits: teams experienced the time savings firsthand.
The clearest sign of success: growing internal demand
Perhaps the strongest indicator that AI acculturation worked: employees now want to share their experiences.
"At our last quarterly meeting, one colleague shared their AI best practices for 20 minutes. There's a clear willingness to continue."— Gowtham Thevaraj, Dust Implementation Lead
The result: foundation for scaled adoption
After six months of bottom-up adoption, Fluxym has built a solid foundation:
- Centralized usage: No more scattered individual AI tools
- Trained and autonomous teams: Multiple training sessions and personalized coaching delivered
- Proven use cases: Knowledge access agents with demonstrated time savings impact
- Cultural shift: Growing internal demand to share AI experiences
- Ready to scale: Organization prepared for structured, top-down deployment
The organization is ready for the next step.
"The agents that are accessible to everyone have become like a second hand depending on the need."— Gowtham Thevaraj, Dust Implementation Lead
Phase 3: time to go top-down
A structured project team
With the foundation in place, Fluxym is now transitioning to a more structured approach. After six months of organic usage, leadership has decided to dedicate resources to a project team that will pilot internal AI initiatives with clearer ambitions.
This team includes Dust experts who will work systematically, deploying use cases one by one rather than waiting for bottom-up discovery.
From organic growth to structured rollout
The path forward involves scaling what's currently organic:
- Dedicated AI team: Resources specifically allocated to pilot internal AI projects
- Department-by-department deployment: Structured rollout of business agents tailored to specific team needs
- Documentation: Creating formal training materials based on early adopter experiences
- Cross-regional sharing: Extending best practices to all offices in France, North America and APAC.
What's next: business agents in production
Fluxym's experience provides realistic expectations for similar organizations: genuine, widespread AI adoption takes 6-8 months to truly take root. With that foundation now solid, the real work begins.
The next chapter of Fluxym's AI journey will focus on deploying purpose-built business agents across specific departments, moving from general knowledge access to targeted workflow transformation.
We'll be back in three months with the next chapter: business agents rolled out in production.
Interested in learning more about how Dust can help your team? Visit our solutions page or reach out to our sales team.
