For most of modern commercial history, the sales stack was not really a stack at all. It was a notebook, a telephone, a desk diary, a pile of business cards, a few handwritten reminders, and the slightly terrifying confidence of a salesperson who insisted they had everything “under control”. Relationships lived in memory. Opportunities lived in drawers. Forecasting was part judgement, part theatre, and part ritual humiliation in a Monday morning sales meeting.
It is easy to forget how recent the modern sales technology stack really is. For all the noise around artificial intelligence, automation, revenue orchestration and predictive systems, the first wave of sales technology was solving an extremely basic problem: how do we remember who we are speaking to, what we promised them, and what needs to happen next?
That is why the history of the sales tech stack is not simply a history of software. It is a history of management. Each generation of technology has tried to solve a problem created by the previous generation of selling. Contact management tried to solve the limits of individual memory. CRM tried to solve the limits of organisational visibility. Marketing automation tried to solve the widening gap between buyer research and seller engagement. Sales engagement tools tried to solve the problem of repeatable outbound execution. Conversation intelligence tried to solve the gap between what sellers said happened and what actually happened. Revenue intelligence tried to solve the problem of forecasting fiction. Artificial intelligence is now trying to solve almost everything at once.
The promise has always been broadly the same: less administration, better visibility, more consistency, greater productivity, and stronger commercial outcomes. Yet the irony is that the sales stack, originally designed to remove friction, has often created a new kind of friction. Sellers now operate in a world of CRM fields, sequencing platforms, call recorders, enrichment tools, enablement systems, forecasting dashboards, intent data, mutual action plans, AI assistants, and internal reporting rituals. The average seller is not suffering from a shortage of technology. They are often suffering from the operational consequences of too much of it.
That makes this a timely moment to look backwards. Not nostalgically, but critically. Because the evolution of the sales tech stack reveals something important about the evolution of selling itself. The tools have changed dramatically, but the central question has barely moved: how do we help buyers make decisions with greater confidence?
The earliest sales tools were gloriously unglamorous. Contact managers such as ACT!, launched in the late 1980s, and GoldMine, which emerged around the same period, helped sellers organise contacts, notes, activities and follow-ups. These tools were not trying to revolutionise the profession. They were trying to stop salespeople losing important information.
That may sound modest, but it was a significant step. Before contact management software, the best organised salesperson often had an unfair advantage. They remembered birthdays, renewal dates, previous conversations, internal politics, objections, preferences, and commitments. Their systems were personal and idiosyncratic. Some were brilliant. Some were chaotic. Most were invisible to the wider organisation.
The first generation of sales technology began to make that knowledge more durable. A customer relationship no longer had to disappear when a salesperson left the business. A manager no longer had to rely entirely on verbal updates. A company could begin to treat customer information as an asset rather than a private possession held by individual sellers.
The next major shift came with sales force automation and CRM. Siebel Systems, founded in the 1990s, became one of the defining companies of the era. It represented the belief that enterprise sales could be systematised, inspected and controlled. For large organisations, this was intoxicating. Sales leaders could finally see activity, pipeline, accounts, contacts and forecasts in one place. At least, that was the promise.
Siebel also represented the old model of enterprise software. It was powerful, expensive, complex and often heavily customised. Implementation could be painful. Administration could be significant. The software was something the business installed, configured, maintained and endured.
Then Salesforce arrived with one of the great positioning moves in enterprise software history. Marc Benioff did not simply launch a CRM company. He launched a campaign against software itself. The famous “No Software” message was not really anti-software, of course. It was anti the experience of buying, installing and maintaining traditional enterprise software. Salesforce turned the frustration of the old model into a brand.
The story of Salesforce staging a mock protest outside a Siebel user conference has become part of Silicon Valley folklore because it captured a genuine market tension. One company represented the established world of enterprise applications. The other represented a new world of cloud delivery, subscriptions, faster deployment and lower upfront friction. The stunt worked because the enemy was familiar. Buyers did not need to be educated on why cumbersome software was frustrating. They already knew.
This is one of the recurring themes in the history of the sales stack. The companies that win categories rarely just sell features. They name the frustration of the previous era.
Salesforce did not just offer CRM in the cloud. It offered liberation from the old software model. HubSpot would later do something similar with inbound marketing, positioning itself against interruption-based marketing and aligning itself with the way buyers increasingly researched online. Outreach and Salesloft would do it again with sales engagement, giving structure to a world where outbound activity had become both essential and increasingly messy. Gong would do it with conversation intelligence, turning the sales call from a private performance into a data source.
The early CRM story is important because it reveals the first great trade-off in the sales stack. CRM created visibility, but it also created administration. It gave management a clearer view of the business, but it did not always give sellers a better way to sell. In many organisations, CRM became less a system of value and more a system of inspection. Sellers updated it because they had to, not because it helped them have better conversations.
That tension has never fully gone away.
By the mid-2000s, another shift was underway. Buyers were changing faster than sales organisations.
The internet had eroded the seller’s information advantage. Prospects could research suppliers, compare alternatives, read reviews, consume content, ask peers, and form opinions long before speaking to a salesperson. In response, companies began investing heavily in content, lead nurturing, email marketing, scoring models and campaign automation.
This was the world in which brands such as Eloqua, Marketo, Pardot and HubSpot became important. They were not all identical, but they were all responding to the same underlying reality: buying was becoming more digital, more self-directed, and more fragmented.
HubSpot’s founding story is particularly useful because it was built around a simple observation. Buyers did not want to be interrupted in the same way anymore. They wanted useful information. They wanted to educate themselves. They wanted to move at their own pace. HubSpot turned that observation into the language of inbound marketing, which became one of the most influential go-to-market ideas of the 2000s and 2010s.
At the enterprise end of the market, acquisitions showed how strategically important marketing automation had become. Oracle bought Eloqua. Salesforce bought ExactTarget, which included Pardot. Adobe bought Marketo. The large platforms understood that CRM alone was no longer enough. If customer relationships increasingly began before a seller entered the conversation, then the systems managing those early interactions became commercially critical.
This period also changed the internal politics of revenue teams. Sales and marketing had always had tension, but technology made the handoff more visible. Marketing could now generate and score leads at scale. Sales could complain about quality at scale. Leadership could inspect conversion rates, campaign performance, source attribution and pipeline contribution. The funnel became a dashboard.
This was progress, but not without consequences. Lead scoring models often created a comforting sense of precision. A prospect who downloaded a whitepaper, attended a webinar and visited a pricing page could look mathematically more ready to buy than they really were. Marketing automation helped organisations create and manage demand, but it also created a new managerial temptation: confusing digital activity with genuine buying intent.
The more the stack grew, the more sales organisations had to learn a difficult lesson. Data is only useful when it improves judgement. Otherwise, it becomes decoration.
This is where the history of sales technology starts to become more complicated. The first wave of tools helped sellers and managers remember. The CRM wave helped organisations see. The marketing automation wave helped companies attract and nurture. But the next wave was more aggressive. It was about speed, scale and repeatability.
The 2010s were the decade of the sales engagement platform.
By then, outbound selling had become a strange combination of art, science and volume. Email was cheap. LinkedIn had become central to professional identity. Data providers made prospect lists easier to build. Sales development teams were becoming more common. Venture-backed software companies needed predictable pipeline creation. The market was ready for tools that could turn outreach into a repeatable operating system.
This is where companies such as Outreach and Salesloft became central characters in the sales tech story.
Salesloft, founded in Atlanta, grew around the idea that sales activity needed cadence, structure and reproducibility. One of the more interesting parts of the Salesloft story is that it had to make a difficult product shift. The company moved away from earlier prospecting products and doubled down on what became its Cadence product. That kind of decision is easy to admire in hindsight and brutal to make in the moment. It means telling customers, employees and investors that the thing creating revenue today may not be the thing that wins tomorrow.
Outreach, founded in Seattle, had a different but equally compelling story. Manny Medina and the founding team built the company after their previous start-up struggled, essentially creating a sales engine that became more valuable than the original business. Outreach became one of the defining companies in the category, competing not only with Salesloft but also with earlier and adjacent players such as Yesware, ToutApp, and later Apollo, Groove and others.
The competition between Outreach and Salesloft was fascinating because it was not merely a feature battle. It was a fight to define the category. Was this sales automation? Sales acceleration? Sales engagement? Sales execution? Revenue workflow? The naming mattered because category design shapes how buyers think, how analysts write, how investors value companies, and how sales leaders allocate budget.
Outreach and Salesloft both understood that the future was not just sending more emails. The category expanded from sequences and cadences into dialling, task management, analytics, coaching, forecasting, buyer signals and broader revenue workflows. Each company wanted to avoid becoming a narrow tool that sat beside CRM. The strategic prize was much larger: become the daily workspace for the seller.
That ambition explains some of the competitive moves. Outreach acquired Sales Hacker, a media and community business for sales professionals, which was a clever recognition that owning attention and education could strengthen a software category. It later acquired Canopy to expand into revenue intelligence and forecasting. Salesloft, meanwhile, continued expanding its platform, achieved a multi-billion-dollar valuation with Vista Equity Partners, acquired Drift to extend into buyer engagement and conversational AI, and later merged with Clari to create a broader revenue AI platform.
The story of Salesloft and Outreach is also a story about the venture-backed sales culture of the 2010s. Growth was the dominant religion. Sales development teams expanded rapidly. Companies invested in more data, more sequences, more touches, more automation and more pipeline generation. For a while, the logic seemed obvious. If a good salesperson could create a certain amount of activity manually, a good salesperson with a sales engagement platform could create far more.
The problem was that buyers were on the receiving end of the same logic.
As more teams adopted similar tools, the customer experience deteriorated. Inboxes filled with templated personalisation. Sequences became longer. Follow-up became automated. The marginal advantage of having a tool declined as everyone else bought one. What once felt sophisticated started to feel noisy.
This is another recurring pattern in the sales stack. Technology creates an advantage, then adoption erodes that advantage, then the market needs a new layer of differentiation.
Sales engagement did improve sales execution. It gave structure to activity. It helped managers understand what was happening. It made follow-up less dependent on memory. It created consistency across teams. But it also contributed to the industrialisation of mediocre outreach. The best teams used it to support relevance, timing and discipline. The worst teams used it to send more average messages to more people more efficiently.
That is not really a technology problem. It is a leadership problem made more visible by technology.
The next major evolution was the rise of conversation intelligence and revenue intelligence.
For decades, sales calls were largely invisible. A manager might join occasionally. A seller might summarise afterwards. A CRM note might capture a sanitised version of events. But the actual conversation, the place where trust was built or lost, where customer language emerged, where objections surfaced, where urgency was created or weakened, was mostly inaccessible.
Gong changed that perception. So did Chorus, ExecVision and other players in the category. By recording, transcribing and analysing calls, conversation intelligence turned customer interactions into searchable data. Suddenly organisations could inspect talk ratios, competitor mentions, pricing discussions, next steps, objections, questions, themes, risks and coaching opportunities.
This was a profound shift. Sales leadership had always relied heavily on interpretation. Conversation intelligence reduced some of that dependency. It allowed managers to see what actually happened rather than relying entirely on what a seller remembered, understood or chose to report.
It also created one of the more interesting competitive dynamics in the stack. Gong became one of the most valuable private software companies in the revenue technology world, raising significant capital at a huge valuation. Chorus, one of its major competitors, was acquired by ZoomInfo. Clari acquired Wingman and later Groove. Outreach expanded into revenue intelligence. Salesloft expanded into conversation intelligence and then buyer engagement through Drift. The market started converging.
This convergence matters. It tells us that the sales stack has been moving from point solutions toward platforms. CRM wanted to own the system of record. Sales engagement wanted to own the system of action. Conversation intelligence wanted to own the system of truth. Revenue intelligence wanted to own the system of prediction. AI now wants to sit across all of them.
The Clari and Salesloft merger is particularly interesting because it reflects this convergence very clearly. Sales engagement on its own was no longer the whole game. Forecasting on its own was no longer the whole game. Buyer engagement, seller workflow, conversation data, pipeline inspection, coaching signals, forecasting and AI agents are all being pulled into the same strategic conversation. The phrase “revenue orchestration” may sound like software marketing, because it is, but it is trying to describe a real shift. Companies no longer want disconnected tools that each optimise a narrow slice of the revenue process. They want connected systems that can see, recommend and act.
The same pattern is visible elsewhere. ZoomInfo started with data and intelligence, then acquired Chorus to move closer to conversations and revenue workflows. HubSpot expanded from inbound marketing into CRM, sales, service, operations and AI. Salesforce moved from CRM into platform, analytics, collaboration, marketing, service and now agentic AI. Microsoft is embedding Copilot for Sales into the applications sellers already use every day.
This is the gravitational pull of the modern stack. Every successful tool eventually wants to become the platform. Every platform wants to become the place where work happens. Every workflow wants to become intelligent. Every intelligent system wants to become predictive. Every predictive system now wants to become agentic.
The risk, of course, is that the buyer does not care about any of this.
The buyer does not care whether your CRM is perfectly updated. They do not care whether your sequence has fourteen steps. They do not care whether your call recording has been analysed by a model. They do not care whether your forecast category has changed from commit to best case. They care whether the salesperson understands their world, helps them think clearly, builds confidence, navigates complexity and makes the buying process easier.
This is where the sales stack has always been slightly vulnerable. It improves the seller’s operating environment, but it does not automatically improve the buyer’s experience. In some cases, it actively worsens it.
Artificial intelligence represents the most significant shift in the sales stack since the arrival of cloud CRM.
The reason is simple. Previous generations of sales technology mostly stored, organised, distributed, measured or automated information. AI systems are beginning to interpret, generate, recommend and act. That changes the nature of the stack.
Salesforce’s Agentforce, HubSpot’s Breeze, Microsoft Copilot for Sales, Apollo’s AI capabilities, Gong’s AI features, Outreach’s AI agents, Clari’s revenue AI positioning and dozens of newer AI-native tools are all competing around a similar promise: less manual work, more intelligence in the flow of work, better prioritisation, faster execution, and a more productive seller.
The ambition is enormous. AI can summarise calls, draft follow-up emails, update CRM fields, research accounts, identify stakeholders, generate talk tracks, suggest next steps, surface opportunity risk, analyse competitor mentions, recommend coaching, build business cases, and help managers inspect pipeline more efficiently. In theory, much of the administrative burden that has haunted sales teams for decades could be reduced.
That is the optimistic view, and there is truth in it. Sales teams have spent years complaining that sellers do not have enough time to sell. Research from Salesforce has repeatedly shown that sellers spend a surprisingly small proportion of their week actively selling, while more recent research also points to tool overload and stack complexity. If AI can remove genuine administrative drag, it could create meaningful productivity gains.
But there is a more sober interpretation too.
Every major generation of sales technology has promised to give time back to sellers. Yet the modern seller still spends too much time on administration, internal coordination, tool management and reporting. That should make us cautious. Efficiency gains do not automatically become better customer conversations. Time saved is not the same as time reinvested wisely.
The numbers tell a slightly uncomfortable story. The sales technology market has expanded dramatically, from roughly 600 sales technology vendors in 2018 to more than 2,000 by 2024. At the same time, Salesforce reports that sellers now use an average of eight tools to close opportunities, while 42% feel overwhelmed by too many tools and overwhelmed sellers are 45% less likely to attain quota. Yet this expansion has not produced a corresponding improvement in commercial performance. Ebsta and Pavilion’s 2025 GTM benchmark analysis, covering 655,000 opportunities and $48bn of pipeline, has been widely summarised as showing B2B win rates falling to 19%, down from 29% the previous year. The conclusion is not that technology has caused lower win rates. That would be too simplistic. The more interesting point is that more technology has not automatically created better selling. In many organisations, it has simply created more activity, more data, more administration and more visibility into the same underlying problem: sellers are still struggling to help buyers make confident decisions.
This may be the most important leadership question of the AI era. If AI gives sellers five hours back each week, what will organisations do with that time? Will they simply ask for more activity, more sequences, more pipeline coverage and more reporting? Or will they reinvest that time into account strategy, deeper research, stronger discovery, more thoughtful business cases, better stakeholder mapping and more effective coaching?
The answer will determine whether AI improves selling or simply increases the speed at which poor selling happens.
There is also a strategic question about the shape of the stack itself. For years, companies kept adding tools. CRM, engagement, enrichment, enablement, conversation intelligence, forecasting, proposal software, call recording, intent data, learning platforms, content management, chatbots, scheduling tools and more. The result was often less a stack and more a software cupboard.
AI may reverse that pattern. If AI can sit across systems, retrieve information, trigger workflows, generate content and update records, then the stack may start to consolidate. Some point solutions will become features. Some categories will disappear into larger platforms. Some AI-native challengers will replace older tools that were built for a different era. Salesforce, HubSpot, Microsoft, Clari, Outreach, Gong, ZoomInfo and Apollo are all trying, in different ways, to ensure they are not the layer that gets abstracted away.
That is why the current moment feels so strategically charged. The battle is no longer only about which tool sellers use. It is about which system becomes the intelligent interface for revenue work.
In the CRM era, the question was: where does the data live?
In the sales engagement era, the question was: where does the activity happen?
In the revenue intelligence era, the question was: where does the judgement come from?
In the AI era, the question may become: where does the work actually get done?
This is where we should be careful not to become either blindly enthusiastic or reflexively cynical. AI will change the sales stack. It already is. It will remove tasks that sellers should never have been doing manually. It will expose weak process. It will make coaching more targeted. It will make poor CRM hygiene harder to hide. It will help managers see risks earlier. It will help sellers prepare faster. It may also flood buyers with even more synthetic relevance, more automated follow-up and more messages that sound polished but feel empty.
The technology will not decide which of those futures wins. Leadership will.
The best sales organisations will not use AI simply to automate activity. They will use it to raise the quality of commercial thinking. They will use it to improve preparation, sharpen qualification, strengthen business cases, identify missing stakeholders, coach managers and help sellers understand the buying journey more clearly. The worst organisations will use it to produce more noise.
This is why the history of the sales stack matters. It reminds us that every tool carries an assumption about what selling is.
A contact manager assumes selling is partly about remembering relationships.
CRM assumes selling is partly about managing information and process.
Marketing automation assumes selling is influenced by education and timing before direct engagement.
Sales engagement assumes selling requires disciplined, repeatable execution.
Conversation intelligence assumes selling can be improved by observing real behaviour.
Revenue intelligence assumes selling can be made more predictable by connecting signals.
AI assumes selling can be augmented by machine reasoning, automation and action.
None of these assumptions is wrong. But none is sufficient either.
Selling remains stubbornly human because buying remains stubbornly human. Organisations do not make decisions just because a workflow has fired, a field has updated or a sequence has completed. They make decisions when enough people believe the cost of change is justified, the risk is manageable, the outcome is valuable, and the path forward is clear.
The sales tech stack has evolved from memory aid to management system, from workflow engine to intelligence layer, from automation platform to AI assistant. That evolution is remarkable. But the ultimate test has not changed.
Does it help the seller help the buyer?
If the answer is yes, the technology matters.
If the answer is no, it is just another expensive layer between two people trying to have a useful conversation.
In September 2024, Tupperware filed for bankruptcy protection in the United States. For many people,…
Every sales leader wants a world-class sales team. The challenge is that most organisations have…
Leadership has always had an image problem. When most people picture a successful leader, whether…
Quick SummaryMEDDIC and MEDDPICC solve the same problem: better qualification. MEDDIC fits smaller and simpler…
There’s a slightly strange expectation in professional life that your relationship with your manager shouldn’t…
There was a time when qualification lived almost entirely inside the mind of the salesperson.…