Part 3 of the Systemic Shift to Scaling AI from Pilot to Scale series, examining how organizations can transform AI initiatives into enterprise-wide value
In our previous post we covered Phase 1, the mental shifts required for AI and the steps involved in Setting the Foundations for scalable AI (getting your executives aligned and your data clean).
Now, we pivot to Phase 2: Creating AI Fluency.
At Ateko, we often come across companies spending millions on internal AI products built on Google Cloud, or they flip the switch on Agentforce in Salesforce or Now Assist in ServiceNow. Then… nothing happens. Adoption stalls. The tech is there, but the people aren’t.
If you want to bridge the gap between “we bought AI” and “we are using AI,” you have to treat AI as a skill that must be learned, reinforced, and rewarded.
Here are the four steps to building an AI-fluent workforce.
1. One Training Deck Does Not Fit All
The Reality: Most companies do a single “Intro to AI” lunch-and-learn and call it a day. That is not enough.
The Fix: You need a two-tier approach. First, yes, everyone needs the basics (e.g., “Don’t put confidential customer data into public ChatGPT”). But after that, you must tailor by role.
- Salesforce Reps: They don’t need to know how LLMs work; they need to know how to use Agentforce to draft a renewal email or find high priority deals in Sales Cloud that need action.
- ServiceNow Agents: They need to know how to use the AI to triage tickets faster and find SOPs relating to specific problems.
- Cloud Teams: They need deep technical training on model orchestration and data pipelines along with enablement content for AI first dev tools like Cursor or Claude Code.
Context is king. If the training doesn’t help them do their specific job today, they will ignore it.
Questions to ask yourself:
- Is somebody monitoring agent logs to refine models and data based on customer prompts that don’t work as expected?
- Do your sales reps know how to prompt for revenue, or just for fun?
- Have you defined what “AI Literacy” actually looks like for a non-technical person?
2. The “Too Busy” Paradox
The Reality: This is the biggest hurdle we see both internally at Ateko and externally with our clients. People know AI will save them time eventually, but they are too busy right now to learn how to use it. It’s a paradox: “I’m too busy drowning to learn how to swim.”
The Fix: You cannot just mandate AI usage on top of a 100% workload. You have to create Rituals. You need to officially allocate time, away from the daily grind, to learn. Maybe it’s “No-Meeting Friday Afternoons” dedicated to testing new workflows.
If you don’t carve out the time officially, the “urgent” will always kill the “strategic.” You have to pay the upfront cost of time to get the long-term dividend of speed.
Questions to ask yourself:
- Have you officially removed something from your team’s plate so they can learn AI?
- Is “playing with the tools” seen as working, or slacking off?
- Do you have a regular time slot where teams share what they learned?
3. Find the People Who Are Already Doing It (Champions)
The Reality: You can hire expensive external consultants, but they don’t know your business. They don’t know that “Project X” is actually code for “The Merger.”
The Fix: You don’t need external experts; you need internal Champions. In every company, there are people already using AI. They are your “Shadow AI” users. Instead of punishing them, turn them into your faculty.
A Champion is someone who translates general lessons into Ateko context. A platform admin who figures out a great way to clean data using AI is worth ten external consultants because they can explain it to their peers in their own language. Find these people, give them a badge, and let them teach.
Questions to ask yourself:
- Do you know who your “Power Users” are?
- Are you giving your internal champions a platform to teach others?
- Are you relying too much on outside voices rather than internal context?
4. Don’t Just High-Five; Copy-Paste
The Reality: Usually, one smart person figures out a brilliant use case, gets a pat on the back, and then… nobody else finds out about it. The win stays isolated.
The Fix: You need a Distribution Mechanism. When a team in Marketing figures out how to automate reporting, how does the Finance team find out? When a Customer Service Agent writes a perfect prompt to de-escalate an angry customer, how do the other 50 agents get that prompt?
You need to systematically “Copy and Paste” success. Whether it’s a shared Prompt Library in Salesforce or a weekly “Best Practices” newsletter, you need to design the pipes that let good ideas flow to the rest of the company. Don’t just reward the innovation; reward the sharing of the innovation.
Questions to ask yourself:
- If someone discovers a huge time-saver today, how long until everyone else knows?
- Do you have a central library for prompts and use cases?
- Are you rewarding the people who share their secrets?
Next Steps
By completing Phase 2, your human capital is finally ready to use the systems built in Phase 1. Now, comes the fun part.
In our next post, we will cover Phase 3: Scoping and Prioritization, how to decide which AI projects to actually build first.
So if you are struggling with AI adoption today? Connect with us.
