Sales

Lessons From Okta’s CRO

This is Part V in our series on lessons from some of the top 1% leaders in B2B software & AI. In case you missed the previous editions:

Part I: Canva’s CCO on CPG discipline and human-centric marketing.
Part II: Vanta’s CRO on execution, competition, and post-sales loyalty.
Part III: Snowflake’s CRO on scaling from zero to category leader.
Part IV: Vercel’s COO on GTM as a product and picking winning companies.

Most AI agents can chat, but Agent A can actually do marketing work.

Built by Ahrefs, Agent A connects to your marketing data and helps you research competitors, analyze content, uncover traffic opportunities, and turn insights into action using real-world web data.

Ask questions in plain English and get answers backed by Ahrefs’ search analytics data. No complex setup. Just connect your workflow and start getting things done.

Try Agent A at ahrefs.com/agent-a

In fiscal 2023, Okta posted an operating loss of $812 million. Three years later, it closed fiscal 2026 with $766 million in non-GAAP operating income, its first full year of GAAP operating profit, and $252 million in free cash flow in the final quarter alone.

Most turnaround coverage reaches for the same explanation: somebody cut costs. Okta did tighten spending, but that isn’t the interesting part of this story.

Nearly every lever Jon Addison pulled was a go-to-market lever.

Jon was named interim Chief Revenue Officer in February 2023, during a period when Okta was absorbing a security breach and a market that had stopped rewarding growth at any cost. He took the role permanently that November. We sat down with him to take the playbook apart, and four lessons stood out in particular:

Specialization is what unlocks a broad platform.
Partner-led growth is a multi-year cultural rebuild (not just a program)
The discovery call is dead, and modern selling is validation.
Every market moves through a five-phase disruption cycle, and your go-to-market should be built for the phase you are actually in.

Below, we dive into each.

Okta’s swing from an $812 million operating loss to operating profit and the metrics underneath it.

1. Specialization is the key to a broad platform

Okta sells one of the widest portfolios in enterprise security: workforce identity, customer identity through Auth0, privileged access, identity governance, and now a set of products for securing AI agents. That breadth is the company’s biggest asset. It’s also its biggest trap.

The trap is asking a generalist rep to carry all of it. Handed five things to sell, a rep sells the one they understand best, leaves the newer products in the deck, and the platform quietly stops compounding. The newer modules never get the at-bats they need to turn into real revenue.

Jon’s fix was structural. He reorganized the field around specialization, with separate motions for the install base, for new logos, for Auth0 and customer identity, and for the broader Okta platform. Each motion builds the specific muscle it needs instead of asking one person to be an expert at everything.

The numbers moved. On Okta’s Q4 fiscal 2026 earnings call, CEO Todd McKinnon said new products, including Identity Governance, Privileged Access, and the new AI agent products, made up roughly 30% of bookings in the quarter. When those products were part of a deal, the average contract value came in about 40% higher! Okta Identity Governance alone now has more than 2,000 customers, a little over three years after launch.

Shipping more products is easy, and most platforms do it. Salesforce spent a decade learning that industry clouds and multi-product expansion only pay off when the field is organized to sell them, which is why it runs specialist overlays and industry teams rather than asking core account executives to carry every cloud. Rippling, built deliberately as a compound product, staffs distinct motions for each module for the same reason. New products convert to ACV only when the go-to-market org is rebuilt to sell them.

If your reps walk into accounts with three things to sell and consistently sell one, then it’s an org design gap.

Why new products stall under a generalist motion, and what changes when the field is specia3

2. Partner-led growth is a 3 year cultural rebuild, not a program

Last fiscal year, 95% of Okta’s top 100 deals involved a partner.

The caution that Jon shared is that you can’t “program” your way to that figure. Okta got there through three years of consistent direction-setting, incentive alignment, friction removal, and real investment in what Addison calls partner practices rather than partner programs. The distinction matters. A program is a portal and a margin table. A practice is a partner who has become a genuine expert in the customer’s business, while Okta stays the expert in identity.

The reason most operators underinvest here is that they misread the math. In enterprise cybersecurity, a company making a generational, multi-year platform bet doesn’t decide on a single vendor call.

“We need seven partners to surround that customer who also trusts us.”

—Jon Addison, Chief Revenue Officer, Okta

Seven trusted voices around the buyer before a company on a multibillion-dollar trajectory commits to an identity platform for the next decade. That math reframes everything downstream. F1 sponsorship, account-based marketing, global system integrator investment, brand spend, and ecosystem events: those stop looking like separate budget lines and start looking like one line, a surface area for trust to accumulate around the buyer.

Cyber has seen this before. CrowdStrike and Palo Alto Networks both built the bulk of their enterprise distribution through channel partners and global system integrators, not direct reps alone, because a security leader making a platform commitment wants independent corroboration before signing. Most companies treat partners as a distribution channel bolted onto the side of the sales org. Okta treats them as the growth engine and builds the org around that conviction, which is why it took three years and why Jon says they’re still early.

For the full playbook on building a partner ecosystem, see GTMnow episode: Partner Ecosystem That Sells For You with Brian Weinberger.

Enterprise buyers gather corroboration before they commit. Partner-led growth is the math of that trust.

3. The discovery call is now a validation call

Buyers don’t show up to a first call ready to be educated. They show up with strong, already-formed opinions about your product, your competitors, and where you fit. Jon’s reframe of what that does to the seller’s job is the sharpest idea in the conversation:

“The first call is no longer discovery. It’s validation.”

—Jon Addison, Chief Revenue Officer, Okta

Showing up with a generic discovery deck is now the fastest way to lose to the rep who shows up with a hypothesis.

Sam Senior, CEO of TestBox, has been making the same case on GTMnow: 70–80% of the B2B purchase decision now happens before the first call, the vendor shortlist has shrunk from three or four names to one or two, and the rep’s job on that first call is to validate or correct what the buyer already believes. More in GTM 186: How Buyers Now Decide Before Talking to Sales with Sam Senior.

Okta’s answer is a new internal sales methodology called APEX, built on the Command of the Message framework and recalibrated for the AI era. In practice, that means value frameworks specific to each new product, a differentiated point of view for each stakeholder in the room, and AI tools, conversational intelligence, quoting agents, and pre-sales assistants embedded directly in the rep’s flow of work.

Okta’s AI rollout inside the field started bottom-up, with reps experimenting with whatever tools they could find. That was useful and genuinely messy. The shift this year was top-down: industrialize the experimentation through one sanctioned methodology, set governance, and place deliberate bets on the tools that scale. Most go-to-market orgs are somewhere mid-arc on this right now, and the ones that move from rep-by-rep tinkering to a single methodology fastest are the ones where the productivity is massively compounding.

This redefines rep skills. What gets automated is the repetitive, knowledge-based administrative layer of selling. What gets preserved, and amplified, is the part AI can’t do: empathy, contextual awareness, and relationship depth.

The first call has changed shape, and Okta industrialized the shift through one sanctioned methodology.

4. Build your go-to-market for the phase of the cycle you are actually in

Jon has lived through two prior platform shifts, the internet and the cloud, and he reads AI through the same lens. The framework he uses is the most portable thing in the entire conversation.

Every major platform shift (like internet and cloud, previously) moves through five phases. It starts with disruption, when the new technology arrives and breaks the old assumptions. Then comes complexity, as companies experiment and stacks fragment into point solutions. Consolidation follows when buyers tire of the patchwork and demand fewer, deeper platforms. Governance arrives next, as regulators catch up and security and compliance harden. Growth closes the cycle when the new normal compounds and the durable winners emerge.

“I think there’s a significant amount of disruption still to happen.”

—Jon Addison, Chief Revenue Officer, Okta

This is the strategic logic under every move Okta is making. The company isn’t playing the disruption phase of the AI wave, the phase everyone else is crowding into. It’s playing consolidation and governance: positioning itself as the neutral, independent platform that’s complete across every identity type, human, machine, and AI agent, and designed to absorb complexity rather than add to it.

Okta’s own research found that 91% of organizations are already using AI agents, while only about 10% have a developed strategy for managing those non-human identities. That gap is the consolidation-and-governance opportunity, and Okta is building a product directly into it. Okta for AI Agents reached general availability on April 30, 2026, and the company’s Cross-App Access standard, its approach to governing how agents connect across applications, has been folded into the Model Context Protocol. The partner rebuild, the platform specialization, the AI governance products, and the APEX methodology all rhyme, because each is a bet on the same thesis: as the AI wave moves from everyone deploying agents to everyone scrambling to govern them, the buyer rewards the platform that already looks like the consolidated answer.

There’s precedent for the pattern. Marketing technology ran through exactly this cycle in the 2010s, exploding into thousands of point tools before buyers forced consolidation onto a handful of suites. Security is mid-consolidation now, which is why Palo Alto Networks has spent the last few years openly pushing customers toward platform adoption over best-of-breed sprawl. The companies that win the consolidation phase are rarely the ones that defined the disruption phase, they’re the ones that read the cycle early and are built for the phase the buyer was about to enter.

Jon’s five-phase disruption cycle and the two phases Okta is deliberately building for.

Tag @GTMnow so we can see your takeaways and help amplify them.

The CRM isn’t going away, but it’s becoming infrastructure. A new a16z piece argues AI agents are doing to CRMs what the newsfeed did to the social graph.

The 2026 CNBC Disruptor 50 is out — Anthropic takes No. 1, 43 of 50 companies call AI essential to their model, and total implied valuation tripled to $2.4 trillion. Proud to see GTMfund portfolio companies like Vanta and Armada on the list.

OpenAI co-founder Andrej Karpathy just joined Anthropic’s pre-training team to build a team using Claude to accelerate pre-training research.

Givefront joins Owner. Three Harvard and Berkeley dropouts recruited after they could name Owner’s restaurant customers by face. Now part of 20+ founders building at scale inside Owner (a GTMfund portfolio comapny).

Sendoso acquired Merch.co, their third acquisition in two years, making them the only end-to-end gifting platform — from ideation to delivery in one system.

Mercury raised $200M at a $5.2B valuation, growing and profitable for four years running. Serving more than 1 in 3 U.S. startups, with an AI-native payroll acquisition and conditional bank approval in hand, Mercury is building the financial OS for the next generation of builders.

Exa raised $250M at a $2.2B valuation to build the search engine for AI. Their API serves 5,000+ companies and 400,000+ developers, including Cursor, Cognition, and HubSpot.

GTM 191: SEO Is Not Enough Anymore: Webflow’s CEO on the Rise of AEO | Linda Tong

Listen through the links in the page above or by searching wherever you get your podcasts “The GTMnow Podcast.”

Armada – raised $230M Series B and launched Galleon Forge One, a facility built to continuously manufacture and deploy modular data centers that bring AI compute anywhere in the world. As the AI race accelerates, infrastructure sovereignty is becoming a strategic priority – Armada is building the industrial base for it. GTMfund is proud to have backed the team.

Polsia – raised $30M at a $250M valuation. They are approaching $10M ARR with just the founder. Polsia is an orchestration of agents that runs the work a company actually needs done, coding, research, cold outreach, paid ads, support, the unglamorous middle of operating a business.

Pivot – raised $40M Series B to replace legacy procurement with an agentic AI operating system. Already in 25+ countries, processing $3B in invoices annually, with DoorDash and Lemonade as customers. Procurement is one of the last enterprise functions waiting to be rebuilt for AI — Pivot is doing it from the system of record up.