We are at one of those moments right now.
Before the internet changed the economics of software delivery, enterprise technology was a services business. The large consulting firms and systems integrators built solutions that were specific to each client. Every implementation was a project. Every project required a team of specialists who understood the client's business deeply enough to translate it into working software.
The results were often impressive. The solutions were genuinely built around how the organisation worked: its specific processes, its industry constraints, its reporting requirements, its way of making decisions. When it worked well, it was transformative.
But it was expensive. It was slow. And the knowledge lived in the consultants, not in the software. When the engagement ended, the knowledge walked out the door with the team. Organisations found themselves dependent on external expertise to maintain and evolve systems that were supposed to serve them. The tail was wagging the dog.
In 1999, Salesforce launched with a simple and radical proposition. Software delivered over the internet. No installation. No implementation project. No expensive consultants. Subscribe and go.
It worked. Not immediately — the first decade of SaaS was met with scepticism from enterprise buyers who had grown up in a world of custom implementations and
on-premise installations. But the economics were undeniable. One codebase serving thousands of customers meant the cost of building and maintaining the software
was spread across an enormous customer base. The product got better faster than any custom solution could. And the barrier to entry dropped from a multi-million
dollar implementation project to a monthly subscription.
By the 2010s, SaaS had won. Almost every enterprise software category had been disrupted: CRM, HR, finance, procurement, workplace management, facilities,
project management. The feature checklist became the primary purchasing criterion. Products competed on functionality. Buyers evaluated on price per seat. The consulting-led era was over.
But the SaaS model contained a quiet assumption that most buyers did not examine closely enough.
The SaaS model works because it is built on a single premise: that most organisations have broadly similar needs in any given software category. One product,
built well, can serve most of them adequately.
For many categories, this premise held. CRM is CRM. Payroll is payroll. The core process is similar enough across organisations that a standardised product,
with some configuration, does the job.
But for categories where the specific way an organisation operates is the product, where the solution needs to reflect the organisation's unique logic rather than
a generic template, the premise began to crack.
Large enterprises started noticing the gap. The product did 80% of what they needed. The remaining 20% required workarounds. The workarounds became
permanent. Shadow IT emerged: spreadsheets and manual processes running alongside the expensive SaaS product, patching the gaps the product was
never designed to fill. Integration costs began to exceed licensing costs. The promised efficiency gains arrived partially, not fully.
More subtly, organisations found themselves adapting to the software rather than the software adapting to them. Processes were redesigned to fit the product's assumptions. Reporting was restructured around what the product could generate. The organisation, gradually and almost invisibly, had been shaped by the tools
it bought rather than the tools being shaped by the organisation it served.
For most organisations, this was an acceptable trade-off. Good enough is good enough when the alternative is an expensive custom implementation.
For large enterprises with genuinely complex operations, specific compliance requirements, industry constraints that no generic product anticipated, and operational
logic that is core to how they compete, good enough stopped being good enough.
The AI conversation in enterprise software has so far been dominated by a relatively unambitious idea: adding AI features to existing SaaS products. Copilots.
Assistants. Automated reports. Smart search. These are useful. They are not transformative.
The genuinely transformative implication of AI has received less attention. It is this: for the first time in the history of enterprise software, it is possible to build
solutions that are simultaneously scalable and specific.
The SaaS era forced a choice between the two. You could have a scalable standard product or an expensive bespoke implementation. The economics of the
pre-AI world made it impossible to have both.
AI breaks this constraint. When intelligence can be trained on an organisation's specific data, shaped by its specific constraints, and updated as those constraints evolve, the solution becomes something genuinely new. It is not a standard product configured to approximate your needs. It is a system that learns your organisation's actual logic and applies it autonomously.
The implications are significant. Domain expertise, the deep understanding of how a specific type of organisation operates that previously lived in expensive consultants, can now be encoded into intelligence that persists, learns, and improves over time. The knowledge does not walk out the door when the
engagement ends. It gets smarter with every data point it collects.
This is not a return to the old consulting model. The era of multi-year implementations, armies of billable consultants, and systems that require permanent external
support is not making a comeback. The economics have changed too fundamentally.
What is emerging is something different. Something that did not exist before agentic AI made it possible.
Domain expertise encoded into intelligent systems. Solutions built around the specific way an organisation works rather than solutions the organisation works
around. Outcomes that improve automatically rather than implementations that degrade over time.
The organisations that recognise this shift early will not simply buy better software. They will gain something more valuable: systems that genuinely understand
how they operate, surface what those systems are learning, and act on that intelligence without waiting to be asked.
The SaaS era gave us powerful, affordable, scalable tools. The era now emerging promises something the SaaS model was structurally unable to deliver: software that knows your organisation rather than software your organisation learns to live with.
I have spent years working inside complex organisations across different industries and geographies. The pattern I keep seeing is this: the categories where standard products have struggled most are the ones where the specific way an organisation operates is inseparable from the problem being solved.
Workspace management is one of those categories. How a bank uses its floors is not how an FMCG company uses its floors. How a large enterprise with hybrid
attendance manages its space is not how a smaller organisation does. The logic is specific. The constraints are real. The standard product gets you most of the way
there.
The next era of this category, and many others like it, will be built around intelligence that understands the specific organisation it serves rather than intelligence built
for the average one.
That shift is already underway. The organisations that act on it early will not just manage their operations better. They will understand themselves in ways their competitors cannot replicate.
Anushka Verghese is Co-Founder and CPO of Smarten Spaces.