At large technology companies, sales teams often serve as the bridge between customers and deeply technical engineering organizations. But when sellers need highly technical answers to customer questions, things can get complicated.

 

At one such company, sellers routinely spent days or even weeks locating the right engineer who might know the answer. Success depended heavily on internal networks — a system that favored tenured employees and left newer sellers at a disadvantage. Engineers, meanwhile, were inundated with requests that pulled them away from their core work.

 

In short:

• Knowledge was fragmented

• Response times were slow

• Documentation was often out of date

• Sellers relied on relationships instead of systems

• Engineers bore an unsustainable support burden

 

The organization needed a new way to deliver fast, accurate, consistent information to technical sellers, without requiring armies of engineers to manually triage questions.

 

The Solution: An Internal AI-Powered Knowledge Assistant

The company’s Sales AI team began building a retrieval-augmented generation (RAG) chatbot designed to pull from trusted internal sources and provide fast, consistent answers. But the team recognized that technology alone wouldn’t solve the problem; adoption would hinge on trust, usability, clarity, and cultural alignment.

 

That’s where I came in.

 

My Role: Leading the Human Side of Adoption

I served as the overall change lead, with a team supporting communications, training, and structured change management. Crucially, I was brought in early, while the tool was still being developed — allowing human factors to shape the solution instead of being retrofitted later.

 

My team partnered with the Sales AI technical team to:

 

1. Ground the solution in real user pain

We activated an existing global change champion network (established for a broader sales transformation) to gather firsthand stories, use cases, and frustrations. This early input sharpened the tool’s design and ensured sellers saw themselves in the solution.

 

2. Build a compelling narrative for why the change mattered

We crafted a clear, consistent story about:

• Why this tool was being built

• How it addressed real seller pain

• What would improve in their day-to-day workflows

• What the tool could do, and what it could not

 

This narrative helped set realistic expectations for both sellers and leadership.

 

3. Create trust-building guardrails

Because accuracy was essential, we worked with subject-matter experts to:

• Establish accuracy thresholds before expanding access

• Define when human validation was required

• Clarify what could be shared publicly vs. only with NDA-protected clients

• Educate teams on RAG vs. GPT — specifically, why hallucinations were not a risk with this system

 

These guardrails became foundational to responsible use.

 

4. Prepare both sellers and engineers

A key insight: even though engineers weren’t the primary users, their behavior could undermine adoption. If engineers kept directly answering questions as they always had, sellers would never build trust in the new tool.

We worked with engineering teams to shift their behavior so they rerouted sellers to the AI assistant and reinforced its validity.

 

5. Enable confidence through communication and training

We launched a full enablement ecosystem:

• Intranet articles and videos

• Executive communications

• Live demos and office hours

• On-demand training and step-by-step guides

• Regular feedback loops with champions

• “Rules of the road” for appropriate use

 

This combination helped sellers adopt the tool gradually and confidently.

 

6. Advocate for a phased rollout

Instead of trying to be everything, everywhere, all at once, we launched first for data center sellers, then expanded into other segments (client, network, etc.). This allowed us to build trust, improve accuracy, and demonstrate quick wins before scaling.

The Impact

Within six months of launch:

• The user base grew from a few hundred early testers to ~5,000 global users

• The AI assistant resolved more than 90% of queries

• The organization saw an estimated $4M in annual savings

• Sellers reported finding answers in minutes instead of days or weeks

• Onboarding time for new sellers dropped significantly

• Engineers reclaimed time to focus on their core work

• The successful rollout paved the way for a broader, company-wide AI platform (“OneAI”)

 

As one seller put it: “This used to take me the better part of a week. Now I get answers in real time while I’m still in front of the customer.”

Lessons Learned

 

Two insights stand out:

 

1. Change and culture must be addressed early — not as an afterthought.

By shaping the strategy to address human factors during development, we prevented misalignment, unrealistic expectations, and resistance. This is where most AI pilots fail.

 

2. Adoption requires influence far beyond the core user.

Engineers, leaders, and operations teams all played a role in whether the change would stick. Adoption is a system, not a silo.

 

And a final leadership insight: In the AI era, leaders must get closer to the technology. They need to understand its strengths, limitations, and risks — not just approve the vision.

Why This Matters

This case study illustrates a universal truth: AI success lives or dies in the human layer.

 

Not in the model.
Not in the code.

In the culture, trust, behaviors, expectations, and mindsets that surround it.

 

For this company, addressing those human elements early meant unlocking enormous value: faster sales cycles, empowered sellers, and a blueprint for scaling AI responsibly.

Return to case studies page →