The technology to deploy AI in channel programs exists today — so why aren’t more organizations using it? In a candid Channel Focus Community webinar, four industry leaders identified the two forces that are actually stalling AI adoption: data readiness and human resistance to change.
The panel included Harbinder Khera (Founder & CEO, Mindmatrix), Raegan Wilson (VP of Ecosystem Consulting & Solutions Innovation, Spur Reply), Laura Cooling Braasch (VP of Client Partnership, Ansira), and Anthony D’Angelo (CRO, Unifier), moderated by Carlos Blanco.
Watch the Full Webinar
The #1 Blocker: Internal Security Sign-Off
Harbinder Khera was direct about where AI projects get stuck:
“Where things are stalled from an AI perspective — it’s not like we don’t have the technology. It’s that the internal AI security team has to sign off on it. That’s where things are stalling.”
— Harbinder Khera, Founder & CEO, Mindmatrix
His recommendation: start the security vetting process now, even before you’ve finalized which use cases to implement. Make sure any vendor you evaluate can demonstrate data governance, isolation between partner data, and compliance with your internal security requirements.
“Make sure they can give you a proper audit — how they’re managing the data, the governance and control, the security process. Make sure that your data will be protected and also separation between your partner data, your data, and individual data. One partner cannot access any other partner’s data via an agent. Everything is sealed.”
— Harbinder Khera, Founder & CEO, Mindmatrix
Laura Cooling Braasch offered a practical shortcut: start with vendors you already trust.
“If you have trusted technology stack today that has already gone through the ringers for you from a security standpoint, a data governance standpoint — that’s a really great starting point. Go to them and say, ‘Hey, what AI do you have?’ And then kick the tires of that AI that you could potentially light up that hasn’t maybe been lit up already.”
— Laura Cooling Braasch, VP of Client Partnership, Ansira
Change Management: Why AI Projects Fail
Harbinder Khera cited a recent MIT study that pinpointed the primary reason agentic AI solutions are failing — and it’s not the technology:
“One of the biggest reasons why the agentic solutions are failing is mostly it has come down to not the technology issue, it’s the people issue. People start losing trust, or people actually just don’t want to change. Change management is the number one reason why AI systems are failing.”
— Harbinder Khera, Founder & CEO, Mindmatrix
Raegan Wilson described the dynamic from a leadership perspective:
“The humans are being asked the question — what are you doing with the rest of your time? Because we’re giving you the tools. We’re enabling you with AI so that you can optimize your day. So, we expect a higher level of productivity out of you as a result.”
— Raegan Wilson, VP of Ecosystem Consulting & Solutions Innovation, Spur Reply
She shared a pointed observation she’d heard at a recent event:
“It’s not that AI is going to take your job. It’s that AI is going to take the jobs of people who aren’t using AI. And I truly believe that.”
— Raegan Wilson, VP of Ecosystem Consulting & Solutions Innovation, Spur Reply
The Rise of the AI Administrator
Harbinder Khera drew a parallel to an earlier era — the rise of sales enablement tools 10-15 years ago — and the lesson that was learned then:
“If you don’t have a sales enablement manager that’s making sure what goes into the drive, where you’re tagging your documents and assets — after a while, sales guys start using their own PowerPoint. We have the exact same problem with AI.”
— Harbinder Khera, Founder & CEO, Mindmatrix
His solution: a new role he calls the AI Administrator. The responsibilities span multiple areas:
The Data Readiness Problem
Raegan Wilson closed the webinar with a reality check she’s seen repeatedly in client engagements:
“We work with a lot of you and we know your data is in a bad place. So, take a couple of steps back and think about your maturity level and your readiness level to execute on these things and start to get yourselves ready. Clean up the data. Get your partner information up to date and nice and clean as you prepare for those two or three use cases.”
— Raegan Wilson, VP of Ecosystem Consulting & Solutions Innovation, Spur Reply
She warned against a common pattern: “We start going down the path with a lot of customers and we get three steps down the path and realize there’s a blocker and it’s something that was self-caused.”
Harbinder Khera framed data as the foundation everything else depends on:
“It really comes down to the game of data. You have two types — structured data and unstructured data. Where the data sits — if the data sits in CRM, if the data sits in a PRM, or a combination of both — that’s where the data resides for partner enablement. Start your journey from there.”
— Harbinder Khera, Founder & CEO, Mindmatrix
Start With What You Have
While larger AI projects work through security approvals and data cleanup, Raegan Wilson challenged channel leaders to act now within their own span of control:
“Go look at what you have access to. Many of your organizations are going to have access to different LLMs or agent models that you can take advantage of today. Look at your current span of control — which is you and whoever works for you. What do you have access to today that you might be able to leverage to optimize any area you’re working in? Start here. Then expand that out.”
— Raegan Wilson, VP of Ecosystem Consulting & Solutions Innovation, Spur Reply
Don’t wait for the enterprise AI project to be approved. Build agents with tools you already have access to — Copilot, ChatGPT, your existing tech stack — and start training the muscle. The reps matter as much as the technology.
Key Takeaways
- Security sign-off is the #1 blocker — start the vendor vetting process now, in parallel with use case planning
- Change management, not technology, is why AI fails — per MIT research, people resistance is the primary failure mode
- Create an AI Administrator role — someone needs to own data quality, prompt engineering, response monitoring, and adoption tracking
- Clean your data first — AI on bad data produces bad outcomes. Invest in data readiness before use case deployment
- Start with existing vendors — your current tech stack has likely already passed security review. Ask what AI you can turn on now
- Act within your span of control — build personal and team-level agents while enterprise projects work through approvals
- Pick 2-3 use cases, not 10 — map them to business outcomes and measure results before expanding
This article is based on a Channel Focus Community webinar featuring Harbinder Khera (Mindmatrix), Raegan Wilson (Spur Reply), Laura Cooling Braasch (Ansira), and Anthony D’Angelo (Unifier), moderated by Carlos Blanco. Watch the full webinar.