88% vs 6%: AI Adoption Meets CRM Reality

Summary

The gap between companies using AI and companies getting value from it is 82 points wide — and it’s not closing.

McKinsey’s 2025 State of AI survey found that 88% of organizations now use AI in at least one business function. But only 6% report meaningful bottom-line impact. In CRM environments, the gap is often even worse.

This post explores why AI deployments fail inside real-world sales systems, what the “88/6 gap” actually means, and why clean data, aligned processes, and operational readiness matter more than the AI tool itself.

I've been quiet on this blog for almost three months. That wasn't the plan. But the ground shifted, and I needed to spend time working on Optrua. The gap between what the AI market is selling and what mid-market companies actually need became impossible to ignore.

So, I stopped writing and started rebuilding. How Optrua thinks about AI deployment, how we position the work we do, and how we talk about the problems we solve. This post is the result. It's also the restart.

The stat that explains everything

The 88/6 gap would be interesting on its own. What makes it important is that every major player in the ecosystem independently arrived at the same conclusion this quarter. The problem isn't AI capability. It's deployment.

• Deloitte’s Tech Trends 2026 found that only 11% of enterprises are running agentic AI in production. 42% are still building a strategy. 35% have no formal agentic strategy at all.

  
  

• Gartner predicts 40% of agentic AI projects will fail by 2027. The reason isn’t technical. It’s that organizations automate broken processes instead of redesigning them.

  
  

• Microsoft’s own numbers show roughly 34% active usage of D365 Copilot among licensed users. 20 million paid enterprise seats deployed, but usage that actually improves productivity is a fraction of that.

  
  

• OpenAI made the most expensive admission of all. They spent $4 billion building a Deployment Company, acquired 150 forward-deployed engineers, and embedded them directly in enterprise environments. The biggest AI company on earth concluded that the model isn’t enough. Getting AI to work inside real organizations requires something the model can’t do on its own.

  
  

• Cisco and Omdia surveyed enterprise executives and found 80% say agentic AI is survival-critical by 2027. Legacy infrastructure, not budget or willingness, is the primary bottleneck for most of them.

The pattern is clear. The technology is awesome – it works. The systems underneath it don’t. And until that changes, the 88/6 gap stays open.

Why the gap exists in AI adoption

This isn’t a mystery to anyone who’s spent time inside a CRM implementation. The gap between “this tool is amazing” and “this tool is working in our environment” has always been the gap in AI adoption. AI didn’t create it. AI made it wider and more visible.

Here’s what happens in practice. A sales team gets access to Copilot for Sales, or Sales Agent, or any of the AI tools now shipping inside Dynamics 365. The demo is incredible. The tool pulls up account summaries, scores leads, drafts follow-up emails, suggests next actions. Everyone nods. This is exactly what we need.

Then someone asks about the process and the data.

“Joe and Mike don’t use CRM, they use spreadsheets.”

“Our pipeline stages haven’t been reviewed in three years.”

“We have 40,000 contacts, but half are from trade shows back in 2019.”

“Marketing and Sales don’t agree on what a qualified Lead looks like.”

“Our team logs activity only about 40% of the time.”

The room goes quiet.

AI reads everything. Every record, every pipeline stage, every gap in the data. And it doesn’t skip what’s broken. It acts on it. With confidence. Lead scores get calculated from incomplete records. Account summaries get generated from outdated contacts. Deal recommendations get made based on pipeline stages that don’t reflect how the team actually sells.

The result isn’t AI failure. It’s AI doing exactly what it’s designed to do, on a system that isn’t ready for it. The good get better. The bad get worse. The unaware get left behind faster than ever before.

That’s the 88/6 gap. Not a technology problem. It’s a systems problem.

What we built for this moment

This is the problem Optrua was built to solve. Not the AI part. The systems part.

Our framework is called SSBS, which stands for Smarter Systems. Better Sales. It’s a registered trademark (U.S. Registration No. 8242530), but the trademark isn’t the point. The point is the thesis underneath it:

Don’t deploy AI to fix a broken system. Deploy AI to scale a system that already works.

AI is a force multiplier. It compounds strength and chaos equally. Organizations with clean data, aligned workflows, and disciplined processes find AI accelerates their advantage. Organizations with broken systems find AI accelerates their dysfunction. SSBS builds the foundation that determines which outcome you get.

The framework rests on a simple lens: People x Systems = Outcomes.

Every revenue system problem we’ve seen in 20+ years of consulting maps to one of four pillars:

Strategy.

Is the revenue team aligned on how the business actually sells? Do pipeline stages reflect the real sales process, or the process someone configured four years ago? Is there a shared definition of a qualified lead, or does every rep have their own version?

Data.

Is the data clean, consistent, and trustworthy? Can a leader pull a forecast and believe it? Or does every board meeting start with “well, the numbers in CRM don’t really reflect reality ...” followed by a spreadsheet that does.

Technology.

Is the tech stack configured to support the actual workflow? Are integrations working? Is server-side sync running? Is there a Dataverse auditing trail? Or has the system accumulated five years of workarounds, custom fields nobody uses, and integrations that technically function but practically confuse?

People.

Do the humans who use the system trust it? Do they use it consistently, or only when someone’s watching? Is CRM a tool people rely on, or a reporting obligation they resent?

AI capabilities sit on top of all four pillars. Copilot for Sales needs configured server-side sync and clean data to pull real account intelligence. Copilot Studio extensions need a governed data model and well-scoped security roles or they surface the wrong data to the wrong people. Sales Agent and autonomous outreach agents don't need accurate pipeline stages and trustworthy contact records to run. They'll run either way. They'll just act on whatever's there — confidently, at scale, whether it's right or not.

Every AI capability is a force multiplier on the underlying system. SSBS builds the system that makes the multiplier work.

How we deliver it

SSBS isn’t a philosophy deck. It’s a delivery methodology with four phases:

Diagnose.

Understand where the system stands before anything changes. This is where the AI Readiness Scorecard lives. More on that in a moment.

Design.

Map the gap between where the system is and where it needs to be. Not a 200-page requirements document. A practical design that accounts for the team’s real workflow, the data model’s actual state, and the technology stack’s genuine capabilities. The design phase is where “automate broken processes” gets replaced by “redesign, then automate.”

Build.

Execute the design. Configure Dynamics 365. Clean the data. Rebuild pipeline stages. Set up integrations. Deploy AI tools on a foundation that’s ready for them. This is where the tangible deliverables live: Sales Agent configured and running, Copilot for Sales pulling real account intelligence, custom Copilot Studio extensions built for the specific workflow.

Evolve.

The system doesn’t stop changing after deployment. New AI capabilities ship quarterly (or more often). Data quality drifts. Team behavior shifts. The Evolve phase is the ongoing relationship: managing the AI layer, tuning agent configurations, monitoring Copilot Studio credit consumption, and keeping the system aligned as the business grows.

Diagnose, Design, Build, Evolve. Each phase maps to a specific engagement type, from free self-serve diagnostics to fixed-fee deployments to ongoing subscription management. The system meets organizations where they are and moves them forward at their pace.

Where you Start: the AI Readiness Scorecard

The Diagnose phase starts with the AI Readiness Scorecard.

It’s free. It’s self-serve. It takes five to seven minutes. And it gives you something most AI assessments don’t: a real answer.

The Scorecard measures your system across the four SSBS pillars: Strategy, Data, Technology, People. It doesn’t ask generic maturity questions. It asks the specific questions that determine whether AI tools like Sales Agent, Copilot for Sales, and autonomous agents will work in your environment or surface your existing problems faster.

What you get

A real score.

Not a range, not a color, not a maturity tier. A number. Concrete, measurable, defensible. Something you can track over time and benchmark against.

A named starting point.

Not “you need better data.” A specific gap with a specific path. So the deployment conversation doesn’t open with a blank slate.

Language for the leadership conversation.

Something you can take to your CRO, your CIO, or your CEO and have it land. Not a 30-slide deck. A score, a gap, and a clear next step.

The Scorecard is the entry point to the SSBS framework. It’s also the fastest way to find out which side of the 88/6 gap your system is on.

The Restart

Three months ago I stopped writing because the ground shifted. AI sales tools have matured (some) and moved from roadmap items to shipping products. The webinar series we ran confirmed that the market had moved past “should we do AI?” and into “which tool, how fast, and is our system ready?” DynamicsCon confirmed it at scale: packed rooms for deployment sessions, IT Directors asking about agent governance, sales leaders comparing notes on rollout plans.

The old content plan was built for a market that was still deciding. That market is gone. The new one is deploying. And the organizations that succeed won’t be the ones that move fastest. They’ll be the ones that move fastest with the right foundation underneath them.

That’s the work. That’s what SSBS is for. And that’s why this blog is back.

We’ll be writing about implementation work — how to get your system ready, how to deploy AI tools that actually produce results, and how to close the gap between "we have AI" and "AI is working."

If you want to know where your system stands, take the Scorecard. Five minutes, a real score, and a named starting point for the conversation that matters.

Take the AI Readiness Scorecard → optrua.com/ai-readiness

About the Author

Ryan Redmond is the founder of Optrua, specializing in CRM strategy and business process optimization. He brings a practical, execution-focused mindset shaped by lessons learned in the Navy and refined through more than two decades of consulting.

Ryan helps organizations align sales, marketing, and technology systems to improve visibility, efficiency, and accountability. His focus is simple: build smarter systems that empower teams to work better, serve customers more effectively, and grow revenue without unnecessary complexity.

Connect with Ryan on LinkedIn.