For most of the history of software, the user has been fairly easy to picture.
A person sits down in front of a screen.
They log in.
They navigate somewhere.
They enter information, retrieve information, make a decision, initiate an action, or move something through a workflow.
We have spent decades improving that relationship.
We moved from command lines to graphical interfaces. From desktop applications to browsers. From browsers to mobile devices. We invested enormous amounts of effort in usability, navigation, onboarding, personalization, workflow design, and reducing clicks.
Entire disciplines grew around understanding what a person was trying to accomplish and helping them accomplish it more easily.
Product management grew up in that world.
So did SaaS.
But I am beginning to wonder whether one of our most basic assumptions is changing:
What if the primary user of software increasingly isn't a person at all?
01
The interface may stop being the center of the product
Imagine that I manage employees in multiple states.
Today, I might use one system for HR records, another for compliance information, another for policy management, another for payroll, and yet another for legal or regulatory research.
If I need to make sure my employee handbook is compliant everywhere we operate, I, or someone working for me, may have to move between all of those systems.
Now imagine instead that I tell an AI agent:
Make sure our employee handbook is compliant everywhere we have employees.
What happens next?
The agent doesn't necessarily need a beautifully organized dashboard.
It doesn't need thoughtfully designed navigation.
It doesn't need to click the same buttons a human would.
It needs capabilities.
It needs to know where our employees are located. It needs authoritative regulatory information. It needs access to our existing handbook. It needs permission to recommend or make changes. It may need an audit trail explaining what changed and why.
In other words, it may need the product very badly.
It just may not need the interface.
That distinction seems increasingly important.
02
What are we actually selling?
For a long time, the application itself has been the visible embodiment of a software company's value.
Customers buy seats.
Users log in.
Engagement gets measured.
Product teams watch adoption metrics.
We celebrate daily active users and worry when customers don't return often enough.
But imagine a future in which a customer's AI agent uses your capabilities hundreds of times a month while the customer rarely opens your application.
Is that a disengaged customer?
Or an extremely engaged one?
What does a "seat" mean in that world?
What does adoption mean?
What does good UX mean?
Perhaps the most important experience no longer happens entirely between a human and a screen.
Perhaps it happens between systems.
And that raises a question I think every SaaS company should at least be considering:
If nobody ever opened your application again, what would still be valuable enough for an AI to come back for?
The answers might include:
I suspect some software companies will discover that those things were always the real product.
The interface was simply how humans reached them.
03
This does not mean the interface disappears
I don't think this leads to a world without human interfaces.
Humans still need to understand what happened.
We need to inspect, approve, override, question, configure, and sometimes simply explore.
There are decisions we may not want an agent making autonomously, even if it technically can.
And there are experiences where interaction itself creates value.
But the role of the interface may change.
Instead of being the place where all the work happens, it may increasingly become a place for supervision, exceptions, explanation, configuration, and judgment.
That would be a substantial shift in how many products are conceived.
04
And then there is product management
This is where the question becomes personal for me.
I've spent much of my career helping companies build software for people doing work.
That has meant understanding users, workflows, jobs to be done, friction, adoption, and all of the other familiar components of product management.
I don't think those skills suddenly stop mattering.
But I do wonder whether the object we're managing is changing.
If agents increasingly mediate between people and software, product managers may need to think much more deeply about capabilities rather than screens.
About permissions rather than navigation.
About the quality and provenance of information.
About what an agent should be allowed to do.
About how a product communicates uncertainty.
About what needs a human checkpoint.
About what becomes an API, what becomes infrastructure, and what remains an experience deliberately designed for a person.
There is a technical dimension to all of this.
There is also a deeply human one.
05
I'm encountering this as a builder, too
My son and I recently started Jackson Solution Works and launched our first product, ScopeAccord.
Building something ourselves has made these questions feel less academic.
When you are starting with a blank page, you have the luxury, and the responsibility, of asking whether assumptions inherited from existing software still make sense.
Do we need this screen?
Should this workflow exist because it serves a person, or simply because applications have traditionally been designed this way?
Is the valuable thing the experience we're putting on the screen, or the capability behind it?
If an agent eventually wants to use this capability for the customer, have we designed the product so that it can?
I don't have settled answers to those questions yet.
In fact, I would be suspicious if I did.
The technology is moving too quickly, the economics are still emerging, and human behavior has a stubborn habit of making clean predictions look ridiculous.
But I think these are questions worth asking now.
06
Bearing Forward
That is why I'm starting this series.
Bearing Forward will be a place to think publicly about AI, product management, SaaS, work, careers, trust, human judgment, and whatever else becomes impossible to separate from those questions.
I am not approaching it as a futurist with a map.
I don't have one.
I am approaching it as someone who has spent a long time building products, who is now building something of her own, and who finds herself both uneasy about the scale of the change ahead and genuinely exhilarated by the intellectual problems it creates.
The terrain is changing.
Some of the old paths may disappear.
New ones will emerge.
We may not be able to see very far ahead at any particular moment.
But we can still pay attention.
We can test our assumptions.
We can make choices.
We can keep building.
And we can keep asking perhaps the most important question underneath all the others:
What do we carry forward?
For now, I'll start here:
