The More Technology We Build, the More Human Guidance Matters

TL;DR

AI makes it easier to build software, but harder to manage a growing mix of apps, agents, and data. Nicolas Desmarais’s vision is to pair AI with connected commerce and trusted advisors, helping businesses choose technology and put it to work.

AI is making software easier to build and business technology harder to manage

As AI changes how software is built and used, teams and individuals can create custom apps and agents around their own needs, opening up new possibilities for how they get work done.

At Thrive 2026, AppDirect Chairman and CEO Nicolas Desmarais described a future where AI helps people do more, while trusted relationships bring context, judgment, and trust to the decisions along the way.

That future is emerging as software itself expands rapidly, bringing the challenge behind AppDirect’s founding back into focus: how can businesses find and manage the right technology as their choices multiply?

AI is accelerating the software explosion

AppDirect’s foundational idea began with a gap in the online economy. Marketplaces were already changing how goods moved from sellers to buyers, yet SaaS had no comparable marketplace. Businesses still faced a fragmented process for procuring software, while providers needed a platform to create marketplaces of their own or distribute their products through other marketplaces. Desmarais described the founding thesis as a bet that “the cost of compute would continue to free fall,” leading to what he called a “Cambrian explosion in the amount of business applications and data”—and, in turn, ever-growing IT complexity.

That prediction has held up, and AI is accelerating it. Desmarais described the growth from roughly 2,000 SaaS applications around the company’s founding to nearly 100,000 today. AI development tools are lowering the barriers to building software even further. As he noted in his keynote, some AI-enabled developers are now publishing more than 300,000 applications a day.

More choice has a cost. Desmarais cited figures that put the friction in practical terms: selecting a technology solution can take more than six months. Businesses with more than 500 employees waste an average of more than $9.5 million a year on unused software licenses. Managing subscriptions across separate billing portals can also leave buyers facing late fees, while employees spend time searching for information across disconnected systems.

AI adds a new layer to the challenge. 82% of organizations have unknown AI agents running in their infrastructure, and they struggle to understand what those agents cost and to control what information they can access or share. The issue isn’t only how to adopt more AI, but how to make growing collections of apps and agents visible, connected, secure, and useful.

The lesson is straightforward: adding tools without improving how they’re discovered, governed, and managed can add complexity instead of reducing it.

More technology makes trusted guidance more valuable

“Where some companies view human interaction as a bug to be fixed, we celebrate relationships as a core feature.”

—Nicolas Desmarais, Chairman and CEO, AppDirect

As the range of available technology grows, choosing the right solution becomes more consequential. In Desmarais’s keynote, six of the eight factors influencing a buyer’s preference for purchasing through a marketplace involved human interaction. They included assisted sales, implementation, training, security assurance, centralized support, and a single service-level agreement.

That human element matters especially when a decision affects a company’s costs, security, operations, and employees. An advisor can understand the customer’s circumstances, help compare options, and support the work of putting a solution into practice. AI can make information and capabilities easier to access; it doesn’t make context or trust less important.

Desmarais calls this relationship-driven commerce: technology commerce designed around the combination of AI and people. AI can help businesses search, compare, and complete tasks. Trusted advisors can help customers make choices that fit, bring services together, and support them throughout the customer relationship.

That model also shapes the role of AppDirect’s platform. It brings together a marketplace for discovering and monetizing digital services with a workspace for managing billing, identity, and data. The aim is to help businesses and their partners move from finding technology to buying, managing, and using it through a more connected experience.

The next interface is conversational, but the work still needs people

“Development costs have now dropped enough that you can build software for a market of one—yourself or a colleague.”

For years, employees have logged into separate cloud applications to create records, ask questions, and receive notifications. Desmarais sees those familiar interactions moving toward a conversational interface, where an AI agent can call on the relevant software capabilities as needed.

Instead of navigating a browser, an application, and a series of screens to complete a task, an employee could tell an agent what they want to accomplish. The agent could then use connected applications and data to help carry it out. In this model, business applications don’t simply disappear. Their capabilities become skills an enterprise AI can use, within the permissions and governance a company sets.

Desmarais describes this as a shift from the consumerization of IT to the humanization of IT: technology that adapts to people and recedes from the work, rather than forcing people to work around technology.

But this transition takes more than a conversational interface. Organizations need to build their ability to work with AI over time. Desmarais outlined a maturity path that begins with using AI to answer discrete questions, then moves through supervised co-creation, delegated assignments, ongoing workflows, and ultimately the orchestration of multiple agents toward larger outcomes.

AppDirect’s own adoption shows how broadly employees are beginning to work with agents. Desmarais said the company has created more than 3,000 agents using AppDirect’s Devs.ai—more than two per employee, with every department building them.

In his Devs.ai walkthrough, Desmarais showed how an agent can move beyond answering questions: it can help create a focused application using App Builder. Once built, the app can be published to a private app store with controls over access and permissions, then made available to agents as a tool through the model context protocol (MCP) (an open standard connecting AI apps with tools and sources). He also demonstrated how agents can be set up for recurring workflows and described coordinating multiple agents toward larger outcomes. The person who built the app still defines the goal and the guardrails. The platform helps turn that direction into a working, governed capability.

AppDirect teams have also built more than 2,000 agentic applications across nearly every department using the Devs.ai Agent Builder, while the company subscribes to more than 300 external applications. Desmarais said there was very little overlap between the internal and external applications—evidence, in his view, that custom applications can create net-new value rather than replace software built for broader markets.

AI makes it feasible to build software for needs that once seemed too narrow to justify a product. Desmarais described the shift this way: “Development costs have now dropped enough that you can build software for a market of one—yourself or a colleague.”

The opportunity is to make these focused capabilities easy to connect, govern, and manage—not just easy to create.

Growth depends on more routes to market, powered by trusted partners

As technology choices multiply, providers need to reach buyers wherever they prefer to discover and purchase services. But adding channels can also add operational complexity. Desmarais described AppDirect’s approach as a single commerce platform that connects providers’ own marketplaces with third-party, hyperscaler, procurement, referral, and reseller channels. Providers can onboard products, manage distribution, track revenue, reconcile transactions, and handle payouts across those routes.

Tackle and PartnerStack extend that reach in complementary ways. Tackle helps providers list and sell through hyperscaler marketplaces, coordinate co-selling, and connect those transactions with their broader go-to-market activity. PartnerStack helps providers recruit, manage, and pay referral partners. Together, they give providers more ways to reach customers without treating every channel as a separate operation.

Trusted partners make those routes valuable to buyers. AppDirect’s network spans more than 150,000 partners across over 170 countries, including affiliates, advisors, and resellers. Their roles differ: affiliates can introduce potential customers, while advisors and resellers help businesses evaluate services, shape solutions, and support customers through purchase and ongoing use. That human guidance matters in a market where buyers face more options and need help choosing what fits.

AppDirect’s catalog gives partners room to grow with those customers. Desmarais said the platform offers more than 50,000 SKUs across eight categories, from connectivity and energy to SaaS, infrastructure, AI, and security. More than a third of partners now sell across multiple categories. Partners can add categories and platform capabilities as their businesses evolve, from quoting and ordering to customer management, training, and support.

The opportunity, then, isn’t simply to add more channels. It’s to connect providers with the partners who help customers find, choose, and manage technology—and to give both sides the infrastructure to work together as customer needs grow.

AI is expanding what people can build and who gets to build it

As technology changes, it’s natural to focus on what might disappear. Desmarais offered a different perspective: past technologies have changed how people work without eliminating the need for people or for new kinds of work. He compared AI’s potential for cognitive work to the way heavy equipment expanded what people could accomplish in physical work.

“AI will do for mental work what the excavator did for physical work.”

AI can help people take on more, increase productivity, and make new kinds of businesses possible. But more businesses and more technology will also mean more decisions about what to use and how to make it work well. That’s where connected platforms and trusted advisors can make a difference.

Desmarais’s vision is an “everything store” for business technology, paired with an agentic operating system and a network of trusted advisors. It’s a vision built not on choosing between people and AI, but on combining AI’s scale with human judgment and relationships.

“Powered by AI plus people, not AI minus people.”

As software becomes easier to create and agents take on more work, businesses will need both powerful tools and people who can help put those tools to work. The more technology we build, the more important it becomes to make it understandable, connected, and human.

Access more insights and resources

• Listen to Nicolas Desmarais' full keynote address

• Go to our Thrive26 resources page to access all the main stage keynotes and spotlights, breakout session insights, and many related resources.