How Gong and Ironclad drive AI feature adoption with certification, targeted learning, and separate tracking, and what customer education teams can learn.
AI feature adoption is how quickly and widely customers start using new AI capabilities in a product. As AI speeds up release cycles, customer education teams keep pace with three moves: certify the roles that drive usage across an account, intervene with targeted learning where adoption lags, and track AI features separately from overall usage.
- The problem: Product teams ship faster than customers can learn. Unused features don’t create value, or renewals.
- The proof: Gong customers with certified Program Managers show 21% higher adoption; one Ironclad accelerator program influenced $1.2M in ARR.
- The shift: Customer education moves from publishing content to intervening where the data says customers are stuck.
AI has changed how fast software ships. Product teams release new capabilities every few weeks, sometimes every few days. That’s exciting, until you look at adoption. A feature customers don’t use doesn’t create value for them, and it doesn’t help you at renewal.
Most customer education teams were built to publish: courses, docs, webinars, release notes. That model can’t keep up with AI release cycles. Teams keeping pace have shifted from publishing content to driving behavior, and they measure success by AI feature adoption, not course completions.
We asked customer education leaders at two fast-growing AI software companies, Gong and Ironclad, how they do it. Their answers boil down to three moves: certify the roles that multiply usage, intervene where adoption lags, and track AI features separately.
How does certification increase product adoption?
Gong, the revenue AI company, reports more than $500 million in annual recurring revenue [VALIDATE]. It uses one of the oldest learning practices to drive adoption of a very new product: certification. Its customer and partner certifications include proctored exams, built-in badges, and shareable credentials.
The key is choosing which roles to certify. Tim Halter, Gong’s instructional design manager, shared in a webinar that Gong’s Program Manager certification drove a 46% increase in team-wide activity. Certifying the admin role lifts usage across the whole account, not just for the person who earned the credential. Gong Chief Customer Officer Simon Frey has also shared on LinkedIn that customers with certified Program Managers show 21% higher adoption than those without.
Certification is the gold standard, but shorter milestones keep people moving. Gong awards badges early in the learning journey and uses Accredible so customers can share credentials on LinkedIn, giving them something to be proud of and putting the Gong brand in front of their networks.
“We have badges just for getting through the ‘Getting Started’ portion of training,” Halter explains, “and we use the nurture campaigns in Thought Industries to follow up with an email that says, ‘Hey, congratulations, you completed this, you’re not done, but here’s your badge. Continue your learning, come back soon.'”
For more on where certification fits, see how to get the most from certification in the customer success lifecycle.
When do you need targeted learning interventions?
Certification builds broad capability. But when a specific feature goes unused, you need to address it directly.
How does Ironclad run its Adoption Accelerator Program?
Ironclad, an AI contracting company that reports more than $200 million in annual recurring revenue, built an Adoption Accelerator Program to make sure customers get value from what they bought. Bill Kelleher, senior manager of learning experience design and documentation, explained how it works: his team tracks adoption milestones across the customer journey, finds accounts lagging on key features, and offers them a short, focused learning series.
“Let’s say, for example, we know that having really clean data is going to make a real difference in a customer’s ability to derive value from Ironclad, but there’s a good subset of customers who haven’t touched their data manager,” Kelleher says. “We go in and identify those customers, and we can offer them a series of three or four sessions, as well as pre-work and homework, and Q&A with Ironclad experts, and really go through it in depth.”
Ironclad then tracks attendance and compares it with product adoption and annual recurring revenue (ARR) for those customers. The results from a single accelerator program:
| Measure | Result |
|---|---|
| Participants | 91 |
| ARR influenced | $1.2 million |
| Participants who attended two or more sessions | 63% |
| Increase in customers using the platform | 200% [VALIDATE wording] |
| Increase in meaningful feature use among data managers, such as workflows with presets | 58% |
“The numbers that we’ve seen are pretty impressive,” says Kelleher. “As a training organization, it’s great for us to be able to put that into perspective for our stakeholders, because it’s always so hard to really show a correlation between users taking training and what the influence is on the bottom line.” For a deeper look at this approach, read “Accelerate Product Adoption with Targeted Learning.”
How does Gong use learning data to warn customer success teams?
Gong intervenes too, but through a different channel. Instead of contacting customers directly, its Education Services team gives customer success managers (CSMs) the data. Mark Banuelos, senior manager of education services, worked with Operations to add education consumption to a Tableau dashboard that CSMs already use.
“Before even six months, ideally a year before [renewal decision], let’s start having the CSMs actively go into a Tableau dashboard, and they can see the education consumption,” Banuelos told his Operations colleagues. “I don’t care if it’s in Academy, I don’t care if it’s done through live training, but have they consumed anything?”
Education data is paired with adoption data. “If they have low adoption in the platform, and/or they have low education consumption, then we immediately have an automated event happen,” Banuelos explains. “It’s an alert [directed to the CSM] saying, ‘Hey, you need to engage with that customer because maybe they haven’t consumed anything, or they haven’t had any training.'”
That alert gives the CSM months to steer customers toward learning that increases adoption, so the renewal conversation focuses on results. “You look like a hero as a CSM, because now you can say, ‘Hey, we went from a year ago to X number, now you have more adoption, you’re seeing your outcomes, you rolled out a new sales methodology, and it’s working.'”
The lesson: intervene early. “We want to make sure we’re operationalizing it and getting ahead of it, and not just cramming training in as a last-ditch, hey, they renew in two months,” says Banuelos.
Why should you track AI feature adoption separately?
If AI capabilities are central to your strategy, overall usage numbers can hide the problem: an account can look healthy while your newest, most important features go untouched. Ironclad’s team uses a product adoption scorecard that tracks AI capabilities and new features separately from the rest of the product. (Ironclad reported that its AI legal assistant delivered nearly six times year-over-year ARR growth in 2026.)
Useful measures include the share of users who used an AI feature in the last 30 days and the share of accounts with any usage. Breaking adoption into smaller milestones makes it easier to see whether learning changes behavior, helps product, customer success, and professional services teams agree on goals, and gives you clearer evidence to share with executives.
What does this mean for your customer education team?
The pattern across Gong and Ironclad is consistent, and it’s a shift any team can make:
- Start from adoption milestones, not a course catalog. Identify the features that predict customer value, then build learning backward from them.
- Certify the roles that multiply usage. Admins and program owners spread adoption across an account.
- Intervene where the data says customers are stuck, with short, focused learning, before the renewal is at risk.
- Put learning data where other teams work, such as the customer success dashboard.
- Track AI features separately, so new capabilities don’t hide in healthy averages.
What does a governed AI knowledge layer change?
Faster releases create a second problem beyond adoption: accuracy. Every new feature changes what’s true about your product, and your courses, docs, help center, and AI assistants have to change with it. When they don’t, customers get contradictory guidance about the very features you’re trying to drive.
A governed AI knowledge layer keeps those answers in step. It connects the knowledge you already have, catches contradictions before customers see them, and keeps guidance current as the product changes. The Customer Knowledge AI Platform from Thought Industries brings this together: Thought Industries Academy builds expertise through certifications, learning paths, and targeted programs like the ones above, and Thought Industries’ Lens, now in Beta, delivers the same governed answers in your product and in the AI tools customers use. For a broader view of AI in customer education, see why AI matters for customer education, and for adoption strategy, see how Gong turned customer education into a product adoption and growth engine.
As Banuelos puts it: “You can build this amazing platform, and you can get everybody into it, but if you don’t have a very great strategy for education, how are your [customers] going to be using what they just purchased day in and day out?”
Frequently asked questions
Why should you track AI feature adoption separately from overall product usage?
Overall usage can look healthy while new AI features go untouched. Tracking AI features separately, for example the share of users or accounts using a feature in the last 30 days, shows exactly where adoption lags, so you can target learning where it will make a difference.
How does certification increase product adoption?
Certification gives customers a clear goal and proof of skill, and certifying the right roles spreads usage across an account. At Gong, the Program Manager certification drove a 46% increase in team-wide activity, and accounts with certified Program Managers show 21% higher adoption.
What is a product adoption accelerator program?
A product adoption accelerator program monitors product usage against targets, identifies customers who are falling behind on key features, and offers them targeted learning, such as a short series of live sessions with pre-work, homework, and expert Q&A, to close the gap.
Which customer education platform helps improve feature adoption?
Look for a platform that supports certification and badging, learning paths by role, automated nurture emails, live and on-demand learning, and integrations that send learning data to your CRM and customer success tools, so teams can see education alongside product adoption.
How do you create a customer training program that drives product adoption?
Start from adoption milestones, not courses. Identify the features that predict customer value, build learning for the roles that use them, certify key roles, watch adoption data for accounts that lag, and intervene early, well before renewal.
How can you track the impact of customer education on product adoption?
Connect learning data to product usage and account data. Compare feature adoption, renewal, and revenue for customers who complete training with those who don’t, and share education consumption with customer success teams in the dashboards they already use.

