Nearly Two Decades at the Bleeding Edge Taught Me That.
I read Rob Bernshteyn’s piece on the Software Apocalypse and found myself nodding throughout.
Not because it was new information. Because it finally named something I’ve been watching play out across every major wave of B2B technology for nearly two decades.
Rob built Coupa around a deceptively simple idea: that software companies shouldn’t be paid for software. They should be paid for the value that software delivers. He called it Value as a Service. And the fact that it felt radical at the time tells you everything about how broken the industry’s default model had become.
That broken model is what the AI market is now exposing at speed. But it isn’t the first time this has happened. And the companies that get through it won’t do so by accident.
Rob’s piece covers two parallel tracks for software CEOs: a bold strategic call on AI transformation or cash preservation, and ensuring every existing customer can measure the value they’re receiving right now. It’s the second track that connects directly to what I’ve spent twenty years building.
The Software Apocolypse is real. It’s also a pattern.
Rob’s argument was direct. Traditional enterprise software was built on complexity, proprietary integrations, and switching costs. Buyers tolerated it because they had no genuine alternative. They absorbed the friction and paid the renewals because leaving felt harder than staying.
AI has changed that calculation. Buyers now have real alternatives to software they’ve been paying for for years. The ‘trust us, the value is coming’ conversation no longer holds. Renewals are being scrutinised in ways they never were before. Companies that built their commercial model on lock-in rather than proven value are finding that lock-in doesn’t hold when the lock can be picked.
What struck me about Rob’s piece wasn’t just the diagnosis. It was that he’d already built this into Coupa’s operating model. The rest of the market is now being forced to catch up.
And here’s the thing about that: this cycle is not new. The same sequence has played out across every enterprise technology wave I’ve been part of. Product excitement arrives. The market scales. Then the market asks ‘where is the actual value?’ and the companies that can prove it win. The ones that can’t lose customers, lose renewals, and eventually lose the argument.
It happened with Business Intelligence. With Big Data. With Marketing Technology. With HCM. The technology changes. The reckoning doesn’t.
What I’ve been doing at the edge of this for nearly twenty years
Throughout my career, I’ve been the search partner that companies at the bleeding edge of AI and data technology have brought in to build their GTM teams. Not as an employee inside those businesses, but as the specialist partner trusted to find and place the people who would make the technology commercially viable.
And across every engagement, the same two-part challenge has been present.
First: how do you make a buyer feel a problem they can’t yet see, put a hard number on what it’s costing them, and build a business case so watertight that it moves from discretionary spend to strategic imperative?
Second: once you’ve sold that commitment, how do you actually deliver on it?
Selling the value is one half of the job. Proving it is the other. The companies that only got the first half right didn’t survive the second question. The ones that got both right built something durable.
Autonomy: selling the answer to a question buyers hadn’t asked yet
Autonomy was one of the first companies I worked alongside at the serious edge of what AI could do. Using Bayesian inference to find meaning and patterns in unstructured data at a time when most enterprise software was still operating on rigid, keyword-based logic. Documents, emails, voice, video. It didn’t need you to tell it what to look for. It found the patterns itself.
The technology was extraordinary. The commercial challenge was that the problem it solved didn’t yet exist in the buyer’s mind. Most enterprises had no framework for understanding what unstructured data was costing them because they’d never had a way to measure it. The pain was real. The cost was significant. But until someone showed them both, they couldn’t act.
The GTM leaders and teams we built there had to do something very specific. Go into a business, surface a problem it didn’t know it had, attach a credible cost to it, and make the case for addressing it. Turn technology capability into a business outcome the buyer could justify at board level.
That is a very particular skill. It’s not standard enterprise sales. And the search for people who genuinely have it is where I’ve spent most of my career.
Splunk: when the data starts telling you things nobody planned for
Splunk was one of the first platforms to turn machine data into real-time operational intelligence. And the use cases that emerged were largely unplanned. Which is part of what made the GTM challenge so interesting.
One example I come back to regularly: airports running security video feeds through Splunk. The platform could detect when a package was left on a floor, track whether anyone returned for it, and raise an automatic alert if it hadn’t moved past a defined threshold. Suspected unattended item. Potential security threat. The kind of judgement call that previously required a human watching a monitor for hours, now handled automatically and at scale.
Then there were the patterns that nobody had specifically gone looking for. Hot weather and BBQ food sales: an obvious correlation. But Splunk’s customers started finding the other signals that travelled with it. What else shifts in consumer behaviour when temperatures rise? What moves before the BBQ sales data does? What can a retailer act on twelve hours before their competitors see the same trend?
That’s not analytics. That’s operational intelligence. But to a buyer encountering it for the first time, it looked like a demo. The GTM job was turning that demo into a board-level business case.
The teams we helped build across sales, presales, customer success, and marketing at Splunk were built around that translation. People who could identify the measurable operational value in what the platform could do, cost the problem it was solving, build the business case, and then stay accountable to delivering the outcome that case promised.
Celonis: finding the problems companies didn’t know they had
Celonis built process mining: the ability to map exactly how work actually flows through an organisation. Not how the org chart suggests it does, or how the process documentation says it should. The real version. Every variation, every exception, every workaround.
The gap between the official version and the real version is where significant cost lives. A business might believe it operates a clean, efficient purchase-to-pay process. Celonis would show it had 47 different execution paths where there should be three. Approval bottlenecks that had become routine. Exception handling that had quietly become standard practice. Manual workarounds built over years because nobody had fixed the underlying system.
None of that was visible before. Most companies had no idea of the scale of what they were losing to operational inefficiency until someone quantified it.
Again: extraordinary technology. But a buyer who can’t see the problem won’t fund the solution. The GTM leaders placed at Celonis were specifically the people who could make invisible cost visible, turn it into a credible business case, and then drive the operational change that delivered the stated return.
Coupa and Value as a Service: the standard everyone is now being held to
Of everything I’ve been part of over the course of my career, Coupa is where I saw the full model executed at its absolute best. And Rob Bernshteyn’s leadership of that company is the reason I want to name it specifically.
Value as a Service wasn’t a marketing concept at Coupa. It was a commercial operating model. Every customer conversation was anchored to what the technology would actually deliver in measurable business terms. Not features. Not roadmap promises. Not theoretical ROI built in a spreadsheet that nobody ever revisited. Actual, committed, tracked, proven value.
Sales makes the promise. Customer success proves it.
The sales team’s job was to identify the business pain accurately, cost it credibly, and build the business case that made investment a rational decision rather than a technology bet. The customer success team’s job was then to take ownership of delivering on that case. Not to manage the relationship. Not to chase NPS scores. To be genuinely accountable for the outcomes that had been promised.
That distinction, between relationship management and outcome accountability, is one of the most important in B2B GTM. Most companies talk about customer success as a retention function. Coupa ran it as a value delivery function. The difference in commercial outcome was significant and sustained.
The teams we built across Coupa’s sales, presales, customer success, marketing, and RevOps functions were built around that philosophy at every level. Not just the leaders at the top of each function. The full teams beneath them. The account executives who could build a watertight business case. The presales engineers who could prove the technical case in the language of the buyer’s business. The CSMs who understood they were accountable to an ROI commitment, not just a satisfaction survey. The marketers who positioned around business outcomes rather than product features. The RevOps professionals who kept the data honest across the whole system.
That is what a proper GTM engine looks like when it’s built around value proof rather than value promise. And it’s why Coupa performed consistently on net revenue retention in a way that most of its competitors couldn’t replicate.
The same motion across SuccessFactors, GitHub, Seismic, and beyond
The same fundamental GTM challenge ran through every other major engagement across my career.
SAP SuccessFactors. HR inefficiency doesn’t appear as a line item on a P&L. The cost of poor people data, manual processes, and talent blind spots is real and significant, but almost entirely invisible until someone builds the case. The GTM job was making that cost impossible to ignore, and then building customer success teams that could prove the people outcomes that justified the investment.
GitHub. Enterprise engineering teams shifting to modern development practice. The business case wasn’t ‘your developers will be happier.’ It was quantified speed to market, measured security risk reduction, and the hard cost of technical debt left to compound. Sold on those outcomes. Proven in delivery.
Seismic. Sales reps spending hours searching for the right content, using outdated materials in front of buyers, taking direct hits to win rates as a result. The pain was specific and quantifiable once it was surfaced. The ROI was trackable. And the CSMs who drove adoption were the ones who made the numbers real.
Different technologies, different buyers, different markets. The two-part motion was identical every time: find the pain, cost it, build the case that makes doing nothing the most expensive option, then be accountable to delivery.
The part of this that most search firms consistently undersell
I want to be direct about something, because it gets missed in most conversations about GTM hiring.
Building a great GTM team is not about placing a leader. A great leader with the wrong team beneath them fails. Slowly, expensively, and usually in a way that damages both the business and the leader’s reputation.
The work is building the complete engine. Sales people who can construct a business case, not just present a product. Presales engineers who speak the buyer’s language and prove the technical case in business terms. Customer success managers who are accountable to outcomes, not just satisfaction scores. Marketers who position around the problem being solved rather than the features being sold. RevOps professionals who keep the data honest and the commercial model coherent.
That is the team. At every level, across every function. The companies I’ve worked with throughout my career that got this right, the ones where the full GTM engine was built with the same rigour at every level, were consistently the ones that delivered sustainable growth rather than lumpy, unpredictable revenue.
Why AI companies are now being held to this standard whether they’re ready or not
The Software Apocalypse Rob described isn’t something that’s coming. It’s here, and it’s moving faster than any previous market correction I’ve seen in nearly two decades.
Every enterprise software, SaaS, and AI company is currently sitting on capability that exceeds its buyers’ current understanding. The models are powerful. The platforms are impressive. But the gap between what the technology can do and what customers are actually realising from it is large, visible, and closing fast from the wrong direction.
Buyers know this. They have more leverage than at any previous point in the enterprise software cycle. More genuine alternatives. Less patience for roadmap promises and theoretical ROI. They want the business case made credibly before they sign, and they want someone accountable to delivering it afterwards. That accountability, the Coupa model, is now the market baseline rather than a differentiator.
The AI companies that scale won’t just be the ones with the best models or the most impressive demonstrations. They’ll be the ones that hire GTM leaders who can build and run the full machine: sales teams that sell on measurable outcomes, presales that prove the technical case in business terms, customer success that owns value delivery rather than relationship management, marketing that builds around the business problem rather than the product capability, and RevOps that keeps the whole system honest.
Finding those people, at every level, across every function, is a very specific capability. It is not what a generalist recruiter does. It’s not what a contingency agency staffing a headcount does. It’s a search discipline built on deep knowledge of what genuinely good looks like inside each of these functions, across the specific context of complex B2B technology.
What this means if you’re building a GTM team for an AI company right now
Nearly two decades of building GTM teams at the bleeding edge of B2B technology, across Autonomy, Splunk, Celonis, SAP SuccessFactors, Coupa, GitHub, and Seismic, is the foundation of what Strong Search exists to do.
Not to fill headcount. Not to send CVs. To find the people who can run the full GTM motion: identify the business pain, cost it accurately, sell the business case, and then prove it in delivery. At every level of the function, not just at the top.
Rob Bernshteyn built Value as a Service into Coupa’s operating model long before the market was ready to demand it. The market has now caught up. If your GTM team isn’t built around the same principle, the reckoning arrives faster than you might expect.
