How Blue Star Innovation Partners brought its whole firm along on AI

Company

Blue Star Innovation Partners, a growth equity firm that invests in B2B software and payments companies (Frisco, Texas)

Engagement

A firm-wide AI kickoff for expanded internal use cases, then a hands-on pilot for the people who wanted to go further

Scope

A short voice intake with each person, a 2-hour in-person kickoff for the whole firm, three 2-hour sessions with operational pods working on real firm projects

Partner

OneTwo Growth Studio

Timeline

June to August 2026

Key outcome

Daily Claude users at the firm grew about 3x after the sessions

Blue Star Innovation Partners (“Blue Star”) is a growth equity firm in Frisco, Texas, founded and run by business founders of software and payments businesses. It invests in B2B software, AI and tech-enabled services, and fintech companies, usually with $5 million to $30 million in annual recurring revenue. In 2026 it launched BSIP AI Labs, an in-house team that helps the firm leverage AI to operate and invest with sharper insight, and enables its portfolio companies to build AI into how they work and what they sell.

Cameron Barragan is a Partner and the firm’s COO, where he co-leads the firm’s and the portfolio’s AI transformation strategy. He joined Blue Star in 2023 after operating roles at Storable, Sprinklr and Bazaarvoice. He sits on several portfolio company boards and has stepped in as an interim leader at portfolio companies.

Roylee Sanchez is Blue Star’s SVP of Value Creation, and he works with the portfolio companies on AI. In a 2026 value creation benchmark from York IE and Dartmouth, he described where his field is going: “The next-gen value creation model is less about sitting in every meeting and more about building systems that consistently surface the right work.”

In 2025, Blue Star made AI its top priority. Cameron and the team saw the potential for the technology running on two fronts.

The first was the firm itself: AI could help with sourcing, deal execution and the back office, so those teams could move faster and do better work. That meant changing the firm’s people, process and technology, and its culture.

The second was within their portfolio: Blue Star works with management teams to implement AI to speed up software development, improve go-to-market, customer success and support, and build automation into their customer-facing products.

Both fronts depend on people at Blue Star understanding AI well enough to use it in their own work, and to judge how the companies they back use it. OneTwo Growth Studio’s work sat on that first front within the firm itself, focused on getting the team thinking and operating with AI on a new level.

When the work started, a small group of senior leaders and operations people were well ahead on AI, Cameron and Roylee among them. They took it seriously and taught themselves fast, each on their own.

That had a cost. “Because it had been a choose-your-own-adventure for everybody across tech and investing, not just at our firm, we’d all developed our own points of view and our own levels of expertise,” Cameron told me. Each of them had picked up language from different thought leaders and from their own experiments. “So we might actually be saying the exact same thing, but using different terms and talking past each other.” This ended up causing misalignment that drove delays, and made it hard to agree on internal use cases and standards.

Behind that group was a much larger one. Most people at the firm were eager to learn how AI could help in their jobs, but didn’t know where to start. The early adopters helped when they could. But they didn’t yet have the “conscious competence,” as Cameron put it, to teach what they knew to everyone at once.

Then there were the skeptics, who would only take AI seriously once the use cases were clear and proven. They would need to see the first two groups make it work before they bought in.

The job was to get all three groups aligned and moving in one direction.

Bringing a whole firm along on AI takes two different jobs. The first is raising the floor: everyone gets the same basic understanding and the same language and frameworks for it. The second is raising the ceiling: the people ready to go further get room to go deep on real work, and they show everyone else what’s possible.

We built the engagement in two parts, one for each job. A kickoff for the whole firm raised the floor, so everyone started from the same frame on the same day. A smaller pilot raised the ceiling, so the people ready to go further could go deep on real work.

Each part spoke to one of the three groups. The kickoff gave the early adopters a shared language, reducing their risk of talking past each other. It also gave the eager majority a place to start. And the pilot projects were scoped to become the proven use cases the skeptics were waiting for.

Before the kickoff, people across the firm did a short interview with a custom voice agent about their work and how they were using AI, and their answers shaped the agenda. Then in late June the whole firm spent two hours together in Dallas. We covered how these tools have moved from answering questions to doing real work, and what a model, its context and its tools each do. Then people mapped one of their own workflows in small groups, with their hands on the keyboard.

The pilot was a smaller group of eight from across the firm. Over three two-hour sessions from late July to mid-August, they worked in pods by function, each on one real project for the firm. The back office pod worked on employee onboarding and offboarding. The deal team pod started building skills for its own deal work, beginning with call prep. The operations pod started on a shared layer for the firm’s portfolio company data.

The sessions went from a deeper understanding of how models and agents work, to designing work for AI instead of only doing today’s work faster, to agents that hand work to other agents. Cameron said the work got the firm to “crystallize some of the lowest hanging fruit and/or highest impact opportunities that we could tackle in each of those areas,” with a project in each one acting “as a forcing function to develop some baseline expertise. And importantly, develop some of that common language across a broader set of people across the business.”

Pushing a whole firm to get better at AI can be a challenge, so I asked Cameron and Roylee how they’d framed it for people.

They started with one idea, which Cameron put plainly. Nobody at the firm would be replaced by AI, but they could be “replaced by somebody who’s really good at AI.” So everyone needed to “develop a baseline of competency,” because even a basic understanding would help in every part of the firm’s work.

Then they led by example. Blue Star’s leaders didn’t order the firm to adopt AI. They put it to work in their own jobs first, and gave other leaders the resources and room to do the same. The bet was that watching a partner or team lead rebuild part of their own day around these tools would move people further than any directive, because it gets them looking at their own workflows and asking where AI could take on part of the load.

That approach fit how Cameron thinks about adoption, in terms of the diffusion of innovation curve. People take up new tools at different speeds, and that’s normal. Some want to try things right away, and others would rather see a use case that already works before they put in the time.

So the job was to find the people willing to try, starting with the firm’s own leaders, and help them go first. “Those people are going to pioneer the use cases that folks that are later in that diffusion of innovation cycle will benefit from and learn from over time,” Cameron said. “So it’s about finding those people and then empowering them with the resources to accelerate their learning.” The aim was for them to “develop those skills for themselves so others can watch and see where the utility becomes practical. And that really worked well for us.”

The engagement also showed them where everyone sat. Between the early adopters and the skeptics, Cameron said, “there’s a whole continuum of people who want to be taught and want to learn and want to be brought along in that journey. And it helped us better assess and understand who those people were so that we could actively engage them in the process.”

Some of the changes reached the whole firm. Others showed up at the top end, in what people could now build and what they could ask for.

The first is use. In early October, Cameron reported that daily Claude users at Blue Star had grown about 3x since the sessions. Roylee called the work “a catalyst to give everybody a baseline on AI.”

People also get up to speed faster. “I think we’re faster to get context, to be honest,” Roylee said. He sees it when someone starts on a new industry or maps a competitive landscape: they’re better at asking an LLM for what they need, and at using it to learn a new field. As he put it, “Our operational IQ on a per company basis increased.”

The biggest change is at the top end, in what people ask for. The deal team has built skills into its daily work. And because people now know what the tools can do, they can say what they need from them.

Cameron said the sessions got “the people that were involved up to a level of baseline competency.” “That allowed them to experiment and get smarter about what it can do, to then crystallize what AI can help them with now and start to build a mental roadmap for what they want AI to help them with in the future,” he said.

The clearest example is the firm’s shared data project, which grew out of the operations pod’s work. It started as one clean layer across all of the portfolio company data. Now it’s a plan in phases: next comes data on companies the firm might invest in, then a library of standard analyses that can run across all of it. The firm’s AI Labs team can sit down with people and understand what they do, because everyone starts from a baseline. Because people are “sufficiently trained up on what AI can do,” Cameron said, “they can come to us with more thoughtful requirements.”

Before, the few people who knew AI well had to pull requirements out of everyone else and translate them into something buildable. Now people push their requirements to them, already framed around what the tools can do.

Two of the firm’s ongoing projects came out of the engagement, and it shaped the work of a third.

BSIP Synapse. The operations pod’s pilot project grew into BSIP Synapse, the first major proprietary tool used across the firm. The project started as one clean layer that lived across all of the portfolio company data, making key data, metrics, and reports on their portfolio accessible within the firm.

Synapse has evolved significantly, and is now the firm’s in-house system for its deal and value creation teams. It brings portfolio company and target company data together with live deal and market intelligence from internal and external sources. It adds the firm’s own context too: its decision frameworks, past portfolio decisions, industry knowledge, memos and calls. Agents within Blue Star connect to Synapse through an MCP, giving the team the ability to build repeatable skills on top.

On the deal side, Synapse enables Blue Star to target and engage more companies in the funnel, working faster, aligned more accurately, and with better insight. On the value creation side, it means less time diagnosing a company’s problems and more time helping fix them.

The back-office operating system. The back office pod’s pilot project became an internal app for employee onboarding and offboarding. Until now, onboarding and offboarding were tracked by hand across a project board and spreadsheets, with software licenses in another spreadsheet. The new system puts all of it in one place. Every hire or departure gets a standard checklist, a Slack bot chases each task until it’s done, and software seats follow people in and out. It’s built, in final testing, and close to switching over. The people who run the back office designed it around their own work, starting in the pilot and continuing to iterate well after the workshop.

The AI Labs team. Blue Star stood up AI Labs in-house as part of its 2026 AI push, and the engagement shaped what the team works on and how people bring it requirements. AI Labs now drives internal projects and tooling to make the firm run better, on top of its work with the portfolio companies. Its goal is to evolve how Blue Star operates with AI. It partners with Frontier Labs AI advisors to stay ahead, and orchestrates forward-deployed AI transformation teams to drive meaningful progress across the portfolio.

Plenty of leaders are trying to do what Blue Star did: bring a whole team along on AI, not only the few people already ahead. Cameron and Roylee each came back to lessons that other leaders in that spot can use.

Roylee’s advice is to get everyone one early win. With AI, he said, what matters is getting “one win on the board.” After that, the rest tends to follow. For a leader, that means planning how each person gets their first win, and staying close until they get it.

Cameron’s advice is to invest in the baseline, because understanding builds on itself. “I think it’s one of those things, once you learn a little bit, you just start to intuitively understand how all the components fit together, and you can’t know it until you’ve experienced it a little bit.” At Blue Star, the baseline built in the firm’s own work is meant to feed the harder questions it asks about its portfolio, like how defensible a company’s AI is and how a company uses AI across operations, software development and its products.

Blue Star started with a few people ahead on AI and a firm that wasn’t sure where to begin as it related to upskilling the whole group. A few months later, the floor has risen for everyone, the people ready to go further are building the firm’s next tools, and the skeptics have the proven use cases they were waiting for. The next step is to go deeper, not wider: Cameron wants “hardened use cases” in the firm’s own data work and in AI for software development across the portfolio, “done extremely well” before anything new.