VP of Product warns that “project manager” will no longer be a standalone role — and shares what PMs must do next
Learn how AI expands a product leader’s role, speeds go-to-market work, and replaces traditional PM processes with faster, insight-driven decision-making.
AI Impact: AI has expanded Scott Jones's role from product leader to deeper involvement in specialties.
Role Evolution: The traditional role of project managers will change, leading to more individual empowerment.
AI Tools: AI tools like Claude and Cursor enhance delivery speed and reduce project management complexity.
Creative Intuition: Intuition and creative skills remain crucial for effectively leveraging AI in product delivery.
Future Vision: AI will enable smaller, more efficient teams and redefine project delivery over the next five years.
Scott Jones is a product leader with technical chops. He's spent much of the last twenty years learning to scale from nothing, and has built zero-to-one products in a wide range of industries. He is currently VP of Product at Realeyes, but thanks to AI, his role has expanded into owning go-to-market as well.
We caught up with Scott to understand how he was able to increase his "surface area" without decreasing quality. He said it's all about AI and intuition.
Connecting the dots between the product and the market
I've spent over 16 years building zero-to-one products across a wide range of industries — adtech, martech, IoT, and cloud infrastructure at places like HPE and Lenovo. Each one taught me something different about scaling from nothing.
Unlock for Free
Create a free account to finish this piece and join a community of forward-thinking leaders unlocking tools, playbooks, and insights for thriving in the age of AI.
Have an account? Log In
Today I lead product, marketing, sales, business development, and partnerships at Realeyes, where I built and run VerifEye — a human verification platform that anonymously validates personhood, uniqueness, and demographics in seconds and for cents. The product went from concept to mid-seven-figure ARR in 18 months with a lean team, and AI is a big reason this velocity is possible.
I work alongside our cofounders, and my role is to own the full go-to-market picture: ensuring our product roadmap, customers, partners, and internal teams align and move in the same direction. There are a lot of plates spinning, but that's what I love most — I get to connect the dots between what we're building and the problems we're solving in the market.
How AI expands a product leader's "surface area"
AI has fundamentally changed how I operate as a product leader — and I don't mean theoretically. I'm living it every day. In lean, zero-to-one environments, this kind of AI leverage is especially powerful. It's my force multiplier.
Customer discovery that used to take weeks of desk research — mapping competitive landscapes, understanding a prospect's tech stack, identifying where our solution fits in their workflow — I can now do in a fraction of the time with dramatically better depth. I run competitive intelligence on companies like Onfido or Entain, preparing for enterprise sales conversations with granular context that used to require an analyst team.
I'm prototyping positioning, building partnership narratives, drafting go-to-market strategies, creating promotional content — all of it faster, more iterative, and honestly, better than when I context-switched between ten different tools and people to complete a first pass.
AI didn't just make me faster; it expanded my scope. My job description didn't change, but my job surface area did — dramatically. I went from being a product leader who coordinates specialists to being a product leader who can go deep in those specialties, at a level of quality that holds up in enterprise conversations.
AI didn’t just make me faster; it expanded my scope. My job description didn’t change, but my job surface area did — dramatically.
I now spend less time on grind work — initial research gathering, first drafts, synthesizing information across sources, and building frameworks from scratch. I spend more time on the higher-order stuff: pattern recognition across markets, making strategic bets on partnership direction, having richer conversations with customers because I walk in better prepared, and moving faster through the decision cycle.
I think about it this way: AI hasn't replaced any role I play. It's made me dangerous in all of them simultaneously.
Why AI will change the role of project managers
The "project manager" role as a standalone function will dramatically change in scope over the next five years. And it will be not be a loss; it will be a liberation.
Here's what I mean: The entire discipline of project management solved a coordination problem. When you have specialized humans who each hold a piece of the puzzle — engineering, design, marketing, sales, legal — you need a human in the middle to ensure information flows, timelines hold, dependencies are tracked, and everyone rows in the same direction. This coordination layer became an entire profession, with certifications, methodologies, and software ecosystems built around it.
The “project manager” role as a standalone function will disappear within five years — and this is not a loss; it’s a liberation.
AI dissolves the coordination problem. It does not replace the project manager with an AI project manager — that's the shallow prediction. It dissolves the problem by collapsing the specialization that created the need for coordination.
I am living proof of this. I run product, marketing, sales, business development, partnerships, competitive intelligence, legal research, and technical prototyping. Two years ago, a team with a project manager kept everyone aligned. Today, I can operate at a level that previously required multiple roles, with Claude as a copilot. The work did not get simpler; it got faster. The bottleneck shifted from "How do we coordinate across specialists" to "How does one person with AI leverage make better decisions faster?"
How AI tools transform project delivery tech stack
My delivery stack has evolved significantly over the past year. AI has moved from a nice-to-have to my operating system.
- Claude — my primary AI copilot: This is the backbone. I use Anthropic's Claude for tasks that would normally require three or four different team roles. I use it for customer discovery, competitive intelligence, partnership strategy, content creation, follow-up communications, pitch deck narratives, sales enablement, market sizing, and thinking out loud with a collaborator who has infinite patience and broad domain knowledge.
- Claude Code — bringing AI into the terminal: I recently installed Claude Code, Anthropic's command-line tool for agentic coding. Instead of having AI in a chat window while you work in a separate development environment, the AI operates directly in your terminal — it reads your files, executes commands, and iterates on code in your actual working directory. I'm early with it, but it significantly impacts delivery. When scoping an API integration for a customer pilot or evaluating how our SDK fits into a partner's environment, I can now explore it technically in real time, rather than writing a spec and handing it to engineering.
- Cursor — AI-powered development environment: This is a newer addition, but it already shifts how I think about prototyping and technical work. Cursor brings AI directly into the coding environment. Instead of context-switching between a conversation with Claude and an IDE, I can iterate on technical concepts, prototype integrations, and work through implementation details with AI assistance embedded in the workflow. For someone who sits at the intersection of product and engineering — scoping APIs, thinking through SDK integration paths, evaluating technical feasibility for customer pilots — an AI-native dev environment closes the gap between strategic thinking and hands-on building.
- AI-driven selling and outbound: This is a game changer for pipeline generation. I use AI-powered tools, including Claude, to automate intelligent outreach sequences — outreach, personalized follow-ups, drip campaigns — all driven by targeting logic.
- HubSpot — CRM and sales operations: HubSpot is our system of record for pipeline management, deal tracking, and customer communications. I track everything from qualified opportunities through technical evaluation to close, log call notes and next steps, and manage follow-up sequences. It's not the flashiest tool in the stack, but it's the connective tissue that keeps everything organized as I run multiple enterprise deals.
AI became the layer that connects everything. Claude is the strategic brain, the social selling tools are the outbound engine, HubSpot is the operational backbone, Cursor is closing the gap on technical prototyping, and the marketplace platforms are the distribution layer. Each piece is stronger because AI runs through all of them.
How Claude enabled every aspect of a sales workflow
Here's an example of what Claude can do. I used it when dealing with an enterprise sales cycle with a major gig economy platform evaluating verification solutions for its contractor workforce.
The complexity here is multi-dimensional: A prospect has an existing verification stack, specific technical requirements for continuous re-verification, data privacy concerns, and multiple stakeholders across product, engineering, trust and safety, and procurement. Normally, preparing for and executing on a deal like this would require a small team — an analyst doing competitive research, a solutions engineer preparing technical positioning, someone drafting follow-up communications, and a strategist developing the account plan.
I did all of it with Claude as my copilot. Here's how:
- Before the initial meeting, I conducted deep competitive intelligence on their incumbent provider — one of the largest identity verification platforms in the market. Claude researched their entire pricing model for me: cost per session, if they charge for failed verifications, how they bundle liveness and document checks, contract structures, and add-on pricing. Within 20 minutes, I had a competitive teardown that would have taken an analyst days — and I formatted it for a Slack message to share with my engineering leads so they had the intelligence before the call.
- For the meeting itself, I used Claude to develop narrative threads to guide my discovery and highlight my capabilities.
- After the meeting, I drafted a follow-up email to four stakeholders. Instead of a generic thank-you message, I used Claude to refine the messaging for three specific conversation threads. Claude helped me calibrate the technical depth — suggesting I simplify a line about true positive versus false positive rates for a non-technical buyer. It was like having an attentive and thoughtful colleague reviewing your work.
Working with Claude is like having an attentive and thoughtful colleague reviewing your work.
It's worth mentioning, though, that I treat AI outputs as a first-pass framework — something to refine and validate, not blindly trust.
Concurrently, I applied the same AI-driven approach to a data aggregation partnership where legal complexity arose. Their use case for 1:N face recognition encountered Illinois Biometric Privacy Law. Claude helped me research the case law, understand how liability triggers during collection versus storage, map out legal precedent for dismissed similar claims, and design three different compliance approaches — ranging from complete state-level exclusion to enhanced consent flows. I walked into that call with a legal strategy that would normally require outside counsel billing hours.
Where PMs should focus their agentic experimentation
As far as agents, I deliberately focus my experimentation on areas that accelerate my existing delivery bottlenecks, rather than building elaborate orchestration for its own sake. My bottlenecks are:
- Technical validation speed when scoping customer integrations
- Handoff friction between product strategy and engineering execution
- My personal bandwidth to prototype ideas before committing team resources.
Claude Code and Cursor directly address all three. So, I'm not running multi-agent orchestration pipelines or building autonomous workflow chains. Some of that is coming, and I pay close attention. But right now, the highest-ROI agentic investment for a lean team at our stage makes each human on the team dramatically more capable — not replacing humans with agent swarms.
It's early and messy, exactly what experimentation should feel like.
How AI enables lightweight systems for small teams
I've moved almost entirely away from traditional project management methods. This wasn't a gradual philosophical shift; it was a practical one, driven by the reality of building a product at startup speed with a lean team.
In my earlier career at companies like HPE, Lenovo, and Digital Turbine, I used the traditional stack. Jira for backlog management and sprint planning. PRDs and BRDs served as formal artifacts. Sprint ceremonies — standups, retros, planning sessions. Teams debated roadmap documents for weeks before building anything. And at a certain scale with large cross-functional teams, the structure serves a purpose. It creates a shared language and predictable cadence.
But when I moved into founder-led, zero-to-one environments, most of that overhead slowed us down rather than enabling us. When you're building a product with a small team, traditional PM frameworks have an upside-down ceremony-to-output ratio.
So what replaced it?
When you’re building a product with a small team, traditional PM frameworks have an upside-down ceremony-to-output ratio.
Claude replaced the document factory. Traditional PM generates enormous amounts of documentation — PRDs, competitive analyses, market requirements, stakeholder updates, go-to-market briefs. I still produce all that work, but now it happens in real-time conversation with AI rather than in templated documents that take days to draft, circulate, and revise. When I need competitive intelligence before a sales call, I don't open a Confluence page and start a research project — I converse with Claude and get actionable intelligence within an hour.
Slack became the operating system, not the sidebar. In traditional PM, Slack is where teams discuss work that lives elsewhere — in Jira, in Confluence, in Google Docs. In my current workflow, Slack drives decisions. I'll run a competitive research session with Claude, format the output for Slack, and drop it directly into a channel for my engineering leads. Strategy, intelligence, and decisions all flow through the same channel without the abstraction layer of formal documents and ticketing systems.
How AI accelerates team alignment and execution management
My core delivery rituals haven't fundamentally changed. Defining scope, aligning teams, validating work, and managing execution — those are still the job. But the speed at which these tasks can be done has changed because AI now acts as a force multiplier for every team member.
Defining scope: The ritual remains the same. Do your homework. Understand the problem, the buyer, the competitive landscape, and the technical restraints. And define what we're solving. The homework simply happens at 10x speed.
Aligning teams: It still means aligning product, engineering, and commercial teams. I still share competitive intelligence in Slack with my engineering leads. I still pressure-test positioning before market launch. I still write the narrative connecting what we're building to why customers should care. But now, I generate that narrative in a single session with Claude before bringing a sharpened recommendation to the team. This changes the quality of the alignment conversation. We spend less time wordsmithing and more time making strategic decisions.
Validating work: This is where AI has perhaps the most interesting impact. Traditionally, validation meant waiting — for user research results, for a pilot to run, or for the market to confirm positioning. I still perform all of those steps. But now I can pre-validate at a previously impractical level. When exploring a data aggregation partnership that encountered biometric privacy law complications, I didn't wait for outside counsel's memo. I researched case law, mapped liability triggers, found relevant legal precedent, and architected three compliance approaches — all before the next call with the partner. The validation cycle compressed from weeks to hours. We still validated with real counsel, but we entered with an informed framework rather than a blank whiteboard.
Managing execution: This still involves coordinating moving pieces, but AI changed my personal bandwidth for managing execution across parallel workstreams.
The philosophical point is this: AI didn't give us new rituals. It removed friction from our existing rituals. When you remove friction from a system, the same inputs produce dramatically more output.
The philosophical point is this: AI didn’t give us new rituals. It removed friction from our existing rituals. When you remove friction from a system, the same inputs produce dramatically more output.
How creative intuition enhances AI-driven project success
Value depends on the human, not the tool. Everyone has access to the same AI. That's why my creative life as a musician and visual artist informs how I use AI while at work.
The best AI work isn't mechanical. It's improvisational. It's knowing when to push Claude in a different direction because something feels off about the output. It's sensing when a competitive analysis needs one more layer of depth because the insight isn't sharp enough yet. It's feeling the rhythm of an enterprise sales conversation and knowing which talking point to pull forward and which one to hold back. None of that comes from a prompt template. It comes from the same place as my musicianship — a willingness to be present, to respond to what's happening rather than what you planned, and to trust your instincts enough to act without overthinking.
I'm also a Kriya Yoga practitioner, and one of the core principles in that practice is channeling what you feel rather than what you think. That philosophy runs through everything I do. When I iterate on positioning with Claude, I don't follow a formula — I feel my way toward the version that resonates. When I'm in a customer conversation, I don't run a script — I read the room and respond to energy. AI gives me the raw material faster than ever. But the human craft — the taste, the timing, the intuition — turns raw material into something that moves people.
AI gives me the raw material faster than ever. But the human craft — the taste, the timing, the intuition — turns raw material into something that moves people.
The delivery leaders who will thrive in this next chapter won't master the most sophisticated AI orchestration platform. They're the ones who bring a creative sensibility to their work — who treat every customer interaction, every strategy session, every product decision as an act of improvisation rather than an exercise in process compliance.
The tools will keep getting better. The humans who know how to riff with them — who can feel the music underneath the data — will build things that matter.
How AI will change project delivery in the next five years
Here's where I see things going over the next five years.
First, the individual contributor becomes radically more powerful. Every person on a team will have the equivalent of a small support staff embedded in their workflows. An engineer won't just write code; they'll generate their own test plans, draft their own documentation, and research their own competitive context. A salesperson won't just sell; they'll run their own market analysis, build their own pitch materials, and research their own prospects' legal and regulatory landscape. When every individual can do the work of three, you don't need as many people in the room, and you definitely don't need someone whose primary job is ensuring the room communicates.
Second, teams get radically smaller while output remains the same or grows. The mid-size company of 2030 looks like a startup from 2020 in terms of headcount, but delivers at the scale that used to require hundreds of people. The companies that figure this out first win. Companies that try to bolt AI onto their existing 50-person department structures will get outrun by five-person teams that are fully AI-native.
Third — and this is the prediction people may not like — the planning ritual dies. Quarterly roadmap reviews, sprint planning ceremonies, elaborate Gantt charts, and multi-week prioritization exercises. All of it was designed for a world where changing direction was expensive because you re-coordinated a large group of specialists. When your team is lean and AI-augmented, the cost of changing direction drops to nearly zero. You don't plan for a quarter because you can re-evaluate and pivot in a week. Delivery becomes continuous and adaptive rather than planned and phased.
Fourth, judgment becomes the only scarce resource. AI will handle research, analysis, drafting, prototyping, testing, documentation — everything fundamentally about processing information and producing artifacts. It won't decide whether to pursue a partnership that's technically compelling but carries reputational risk. It won't read the room when a customer's VP of engineering says something signaling internal politics. It won't know when to push hard on a strategic bet versus when to let an idea breathe. The humans who thrive are those with taste, judgment, and the ability to operate in ambiguity — not those who are really good at maintaining a Jira board.
The humans who will thrive are those with taste, judgment, and the ability to operate in ambiguity — not those who are really good at maintaining a Jira board.
And fifth — this is the one I feel most strongly about — the word "delivery" itself evolves. Right now, it means "we built the thing and shipped it." In five years, delivery means "we identified the opportunity, validated it, built it, shipped it, measured it, and iterated" — and one person or a tiny team did all of that in the time it previously took to write the requirements document. The entire cycle compresses so dramatically that the distinction between strategy and execution, planning and delivery, and thinking and building — it just evaporates.
This model works best in lean, zero-to-one environments. At scale, coordination doesn’t disappear — it shifts.
Why it's a good thing that AI made the delivery leader's job obsolete
I've learned three very important things the hard way during this time of change:
- Stop waiting for permission to experiment. There is no AI readiness program, no certification, no approved vendor list that will tell you it's time. Leaders pulling ahead now started using AI in their actual work — not in a sandbox, not in an innovation lab, not in a "pilot program." In their real pipelines, with real customers, on real deadlines. I didn't get executive approval to use Claude for competitive intelligence before enterprise sales calls. I just did it; my preparation quality spoke for itself. Results grant permission. If you wait for your organization to hand you an AI strategy on a silver platter, you're already behind someone who started six months ago with only curiosity and a chat window.
- Invest in your judgment, not your process. Many delivery leaders respond to this moment by doubling down on methodology, trying to figure out "the AI-native project management framework" or "the agentic delivery playbook." That's the wrong instinct. The frameworks will keep changing. The tools will keep evolving. The value of a human who can walk into a room — or a Slack channel — and make the right call with incomplete information won't change. Know your customers deeply. Understand the business, not just the backlog. Develop the pattern recognition to see around corners. AI will make every mechanical aspect of delivery cheaper and faster. It won't replace the leader who knows which mountain to climb. Double down on being that person.
- This one changed my career: redefine your job. For years, I thought my job was tactical product management work: managing backlogs, writing PRDs, running sprint ceremonies, and aligning roadmaps. Then, AI made most of that work trivial — or at least dramatically faster. Once I let go of the rituals and leaned into the judgment, everything accelerated. I went from a product manager to a business builder who uses product skills. That shift — from process custodian to strategic operator — is available to every delivery leader now. AI doesn't shrink your role. It strips away the overhead and reveals what your role always was.
If you wait for your organization to hand you an AI strategy on a silver platter, you’re already behind someone who started six months ago with only curiosity and a chat window.
This moment rewards those comfortable being uncomfortable. I'm installing developer tools I've never used. I'm researching case law without formal training. I'm running social selling campaigns, prototyping SDK integrations, and preparing for enterprise sales conversations with Fortune-level companies — all in the same week. None of that was in my job description five years ago. All of it is possible because I decided the boundaries of my competence are suggestions, not walls — and AI lets me climb over them.
My advice is simple: Start now, trust your judgment, and let go of the version of your job AI just made obsolete. What's on the other side is better.
Follow along
You can follow Scott Jones on LinkedIn and Instagram as he continues testing what's possible and expanding the role of PMs.
More expert interviews to come on The Digital Project Manager!
