Stop Buying AI Seats. Start Rebuilding How Work Gets Done
Buying everyone an AI seat is not AI transformation.
It can be a useful starting point. It can create individual productivity gains. But from what I am seeing in the enterprise, it is rarely the thing that changes how a business operates.
The companies that will win with AI are not simply the ones that give every employee access to another tool. They will be the ones that redesign work itself.
And that matters to me because I want enterprises to win this next era.
AI Should Be Built Into Work
Most employees are never going to become great prompters. They should not have to.
We should not expect every person across a large organization to understand models, prompts, agents, context windows, or which tool is best for a given task. That is not a scalable transformation strategy. It is asking people to add one more skill and one more application to an already crowded workday.
The real opportunity is to make AI part of the workflow itself.
Build AI directly into how work gets done
Remove tedious tasks instead of asking people to AI-assist them
Consolidate fragmented SaaS tools into AI-native systems
Give employees outcomes without requiring them to understand the technology underneath
Allow revenue and output to grow without headcount needing to grow at the same rate
The best AI experiences will often feel almost invisible.
A sales team should not have to manually summarize calls, update customer records, research an account, draft a follow-up, and coordinate internal next steps across multiple systems. An operations team should not have to reconcile spreadsheets, chase approvals, or turn disconnected reports into decisions.
Those workflows should be redesigned.
The system should capture information once, route it where it needs to go, automate repeatable actions, surface exceptions that require human judgment, and help people make better decisions faster.
That is transformation.
What I’m Seeing in Enterprise
Across enterprise organizations, I am seeing a growing divide.
On one side are companies running AI pilots, buying point solutions, and encouraging employees to experiment. There is value in that. It helps teams learn and creates momentum.
But the other side is more interesting.
These are companies asking deeper questions about the way work moves through their organization. They are looking at where employees lose time, where customers feel friction, where teams duplicate effort, and where data gets trapped between systems.
They are not trying to bolt AI onto every existing process.
They are asking whether the process should exist in its current form at all.
That is a much more difficult conversation. It requires leaders to look beyond isolated efficiency gains and confront the accumulated complexity inside the organization. But it is also where the biggest opportunity lives.
Enterprises have enormous advantages if they are willing to act on them:
Deep domain knowledge built over years or decades
Established customer relationships and distribution
Valuable proprietary data
Experienced teams who understand the real edge cases
The resources to build and scale durable platforms
The goal is not to discard those advantages in pursuit of the newest AI tool. The goal is to turn them into better products, better operating systems, and better experiences for employees and customers.
The Build Decision Has Changed
This is why I am increasingly bullish on building custom software again.
For years, the default enterprise answer was buy.
That made sense when building software was slow, costly, and difficult to maintain. It was usually easier to accept a vendor’s opinionated workflow than to design and build something tailored to your organization.
The result was often a growing collection of SaaS tools, integrations, handoffs, duplicate data, and workarounds.
AI has changed the economics.
The cost and barrier to building custom software are falling. Teams can prototype faster, test ideas earlier, and build systems that fit the actual needs of the business rather than forcing the business into a generic template.
That does not mean enterprises should build everything.
They should still buy commodity capabilities where buying makes sense. But the workflows that create differentiation should be evaluated differently today.
Buy when | Build when |
The capability is standardized and not strategically unique | The workflow is core to your customer experience or operating advantage |
Speed to initial deployment is the primary need | Existing SaaS tools create friction, manual work, or inflexible processes |
The organization can use the product largely as designed | The business has proprietary data, logic, or processes that should shape the product |
Customization would create more complexity than value | The solution can consolidate tools, automate decisions, and create long-term leverage |
The question is no longer just, “Can we buy software that does most of this?”
It is increasingly, “What could we build now that was not practical to build two years ago?”
The Enterprise Opportunity
Large enterprises will not transform overnight. Legacy technology, governance, security, data architecture, and organizational complexity are real constraints.
But those challenges should not become excuses to wait.
In fact, enterprises have a chance to use AI to simplify. They can retire disconnected systems, reduce unnecessary process, and redesign the workflows that slow down their people and customers every day.
Smaller, faster companies may have an advantage in moving quickly. They can rip out entire workflows and rebuild them AI-native. But enterprises have assets that startups cannot easily replicate: scale, trust, distribution, customer access, and institutional knowledge.
I want enterprises to use those strengths.
I want them to build systems that make their people more effective without demanding that every employee become an AI expert. I want them to reduce the operational drag that has built up over years of software purchases and process layers. And I want them to create businesses where growth is not automatically tied to adding more people to manage more complexity.
The enterprises that win will not necessarily be the ones with the most AI licenses.
They will be the ones that make work meaningfully better.
Make AI Invisible
The future is not a company full of people spending their days prompting tools.
The future is a company where the tedious work disappears, information moves cleanly, systems understand context, and people have more time to do the work that needs judgment, creativity, and relationships.
Do not make AI another task.
Build the company so using AI becomes invisible.