Companies building AI app portfolios face tough questions after they figure out the basics: What do we do next? How do we find the right path in a landscape full of wrong turns?
Savvy replies are piling up at Nutanix, the cloud infrastructure software provider, where developers have built three in-house AI apps and are finishing their fourth. For all the obstacles in choosing language models, configuring APIs, connecting to knowledge bases and nailing down data governance, there’s a simple starting point.
“You have to know your use case,” advised Rajesh Bikky, principal product manager at Nutanix, in an interview with The Forecast.
“You have to know what you're trying to solve and why you're trying to solve it.”
Bikky and his team had a straightforward use case in mind: Using AI chatbots to create more win-wins in Nutanix’s enterprise channel-partner programs. Their experience underscores what it takes to stay grounded in real-world needs when expanding an enterprise AI portfolio.
Growing an Enterprise AI Toolkit
Enterprises worldwide are putting AI tools to work to drive business improvements. The 2025 Nutanix-commissioned Enterprise Cloud Index survey noted that nearly 85% of respondents in global enterprise companies had deployment strategies for generative AI (GenAI). More than half of the global IT decision makers and practitioners stated that they were putting their AI strategies into action.
Forrester Research provided an instructive use case with Izola, a GenAI chatbot, to explore the company’s technology insights. Forrester has decades of research reports in its archives, making it a natural fit for an AI chatbot for internal users and external clients.
Nutanix, meanwhile, has extensive experience in developing software to manage hybrid cloud IT infrastructures that power AI applications. AI chatbots are ideal for querying Nutanix’s immense volumes of software documentation and implementation guidance.
Recognizing these connections, an in-house team at Nutanix started building a portfolio of home-built AI chatbots for product support staff, sales pros and system engineers. Their latest effort is PartnerGPT, which adds conversational AI for the channel-partner-marketing programs Nutanix depends on to give its products global reach.
“It has access to all of our partner-related documentation,” Bikky said. Because it’s a GPT, he added, it can draft emails, summarize complex documents and speed up processes that used to be slow and cumbersome.
This straightforward addition can add a lot of value to enterprise partnerships.
Why a Partner Program is an Appropriate AI Use Case
Like most enterprise software vendors, Nutanix leans heavily on channel partnerships to serve its global clientele. Partners such as distributors, value-added resellers (VARs), service providers, and system integrators (SIs) have local knowledge, reach and expertise that help enterprises focus on innovation and growth.
For all their appeal, vendor partner programs and tools aren’t immune to processes that slow things down when partners want speed. Vendors often provide incentive programs that boost partners’ profitability, for instance.
Naturally, partners want fast, easy answers to questions about specific products, their business performance, or available incentives. Uncertainties can arise if partners can’t find the answers they need quickly, for example, if incentive plans change every year and vary by location and product offering. Nutanix partners in Brazil or Spain, for instance, might have subtle differences in their vendor’s program performance or sales-incentive guidelines.
Traditionally, partners would ask their counterparts at Nutanix to clear up these questions. The typical response: “We’ll get back to you.” That might take hours for a question that an AI chatbot can answer in seconds. If a dozen Spanish speakers ask the same question, GenAI gives them the same answer, reducing the risk of human error.
These kinds of everyday pain points bolstered the rationale for adding an AI chatbot in Nutanix’s partner program. The new bot is called PartnerGPT, which works like this: Nutanix’s partner-facing teams take questions from their partners and feed them into the chatbot, which spits out fast, current information.
“It's always based on the latest documentation and data,” Bikky told The Forecast.
He said Nutanix partners can get fast, consistent replies to questions rather than spend time searching for and parsing through files stored somewhere in a database.
Bikky and his team asked their partners what they needed most from supplier relationships.
“All of them said if a supplier makes it easier for us to sell their products, they will automatically get more mind share,” Bikky said.
Removing friction from partner relationships creates a greater incentive to stick with Nutanix and a lesser incentive to work with other companies. PartnerGPT nurtures these incentives.
Lessons Learned from a Fourth AI App
Bikky recalled that Nutanix’s first GPT apps, for product support, took up to a year to develop. The sales-centered GPT took about nine months, and the system engineers’ app took four to five months.
“Partner GPT took something like two to three months,” he recalled. “Our speed in deploying these completely distinct GPT apps is now much faster because the platform is the same.”
API, data and application integrations used on previous GPTs can be plugged into new ones.
“If we have to spin up a new GPT based on some other use case, we know exactly what to do. We just need to find the right sources, train the model and fine-tune it and do some testing on it to see if it's responding accurately.”
PartnerGPT is currently used in-house, but the company intends to make it accessible directly to their partners through the newly launched Nutanix Partner Central platform.
“We asked the team to identify the source of documentation that they would like to use,” Bikky said.
Different data labels would be required for external-only vs. internal-only users.
“That metadata, that tagging, is important,” he said.
Documents analyzed by the GPT also must be heavily curated for accuracy, utility and timeliness, Bikky added.
Getting to Success with Enterprise AI
With a handful of GPT apps in their holster, Nutanix insiders can look for new problems to solve.
“When you start building an AI app, you can do a lot of things,” Bikky said.
Many of those things don’t pan out. For instance, a report out of MIT grabbed headlines for suggesting 95% percent of enterprise GenAI projects didn’t prove financially worthwhile.
Keen to avoid that fate, Nutanix’s GPT developers sought another persistent challenge that GenAI could address. They found an attractive prospect in Nutanix’s years of accumulated question-and-answer exchanges with channel partners.
“We have over a hundred thousand emails, so we can see the kind of questions partners were asking us,” Bikky said.
Channel partners also get large volumes of questions from their clients — Nutanix end users. The GPT developers also asked their partners these questions.
“We put all of that together and the use case was staring right at our face,” Bikky said. “This is the exact one we should go solve.”
Across four GPT applications, Nutanix’s developers prioritized the most beneficial use cases they could address on a tight timeline. That meant creating a practical foundation that would become more useful with each new GPT application.
“Instead of technically building a completely new solution and coming up with something fancy, we've taken what we already know and applied it to the channel partner use case,” Bikky concluded.
“Now we can develop something that has the highest ROI in the shortest amount of time.”
Tom Mangan is a contributing writer. He is a veteran B2B technology writer and editor, specializing in cloud computing and digital transformation. Contact him on his website or LinkedIn.
© 2025 Nutanix, Inc. All rights reserved. For additional information and important legal disclaimers, please go here.