The most useful innovation often begins with one recurring problem that someone finally decides to make smaller.
Make the problem smaller
Innovation is usually described at enterprise scale.
We hear about major platforms, transformation programs, new services, large investments, and ambitious capabilities that may take years to deliver. Those efforts matter. They can create real value. They can also make innovation feel distant from the people doing the daily work.
I have come to see innovation differently.
Sometimes innovation is a new enterprise capability. Sometimes it is a small change that removes three hours of repetitive work, reduces rework, improves an intake process, or helps a team make a better decision faster.
Innovation does not have to be big.
It begins with friction.
A recurring task that should not require the same manual effort every week. A handoff that repeatedly breaks. A request that arrives without the information needed to act. A search for knowledge that takes longer than the decision itself. A communication that is recreated from scratch every time. A process step that exists because nobody has stopped to ask whether it still creates value.
Those are not minor inconveniences. They are signals.
Innovation is the disciplined act of noticing friction and improving it.
That improvement may involve automation. It may involve an agent, better data, a reusable template, a redesigned process, or the removal of an unnecessary step. The tool matters less than the problem.
A weak innovation story begins with the tool and searches for a problem. A stronger one begins with the problem and chooses the simplest responsible tool.
That distinction matters even more now because the barrier to experimentation is falling quickly.
Many employees have access to tools that can help summarize information, prepare communications, classify requests, search knowledge, assist with analysis, generate test cases, review code, or support other bounded tasks. More advanced platforms can extend those capabilities under stronger governance and broader sharing models.
That does not mean every task should be handed to an agent. It does not mean controls disappear. It does not mean human accountability becomes optional.
Lower friction does not eliminate governance.
Deterministic innovation
My own operating philosophy is simple: deterministic innovation wins the day.
- Notice the recurring friction.
- Define the boundary and choose the simplest responsible tool.
- Keep a human accountable and measure whether the change helped.
I do not want innovation that depends on novelty, hype, or a tool behaving impressively in a demonstration. I want a bounded problem, explicit controls, a repeatable method, measurable evidence, and an accountable human.
Every useful experiment should be able to answer a few basic questions:
- What exact problem are we solving?
- What is in scope and out of scope?
- What controls prevent unintended action?
- Who remains accountable?
- What evidence will show whether it helped?
- Can the result be repeated and understood?
If those questions cannot be answered, the work may still be interesting, but it is not yet dependable.
In my own service area, we have applied this approach in practical ways.
We use assisted development to accelerate engineering work. We use agents to help create and review unit tests, inspect code, support UAT preparation and evidence review, refine engineering intake and user stories, prepare communications, and triage large bodies of information.
The agents are not replacing judgment. They are removing friction around judgment.
That distinction is essential.
They do not independently own production decisions. They do not replace regulated validation. They do not carry final UAT accountability. They do not receive unbounded authority to mutate systems because someone decided autonomy sounded modern.
They perform bounded assistive work under human direction and review.
The article you are reading is itself an example.
The central idea came from a human problem and a human point of view. Agents helped preserve the discussion, sharpen the thesis, challenge weak framing, check factual claims, apply publication controls, develop the structure, and prepare a deterministic handoff for implementation.
The machinery did not decide what I believe. It helped reduce the friction between the idea and a publishable result.
That is exactly the kind of innovation I am describing.
The value is not that artificial intelligence touched the work. The value is that a repeatable system helped move the work forward while preserving human intent, accountability, and control.
Innovation becomes real when someone makes a recurring problem smaller.
Leadership has an important role. Leaders can make tools and patterns understandable, remove unnecessary barriers, protect limited experimentation time, provide practical coaching, recognize small improvements, and create pathways for useful experiments to become reusable capabilities.
But employees also should not wait for innovation to arrive as a finished enterprise service. The people closest to the work often see the friction first.
The starting point is not, “What major innovation should our organization build?” The better question is often much smaller:
“What repetitive, frustrating, or error-prone problem can I make smaller?”
Start there. Define the boundary. Choose the simplest responsible tool. Keep a human accountable. Measure whether the change helped. Share what you learned.
Do not confuse small with insignificant. A modest improvement that returns time every week, reduces risk, improves quality, or makes knowledge easier to use can create more durable value than a highly visible experiment that nobody can repeat, govern, or sustain.
Deterministic innovation wins the day because repeatable improvement creates more durable value than novelty without control.
Innovation does not have to be big.
It has to improve something real.
Check the work, not just the conclusion.
Public research, authority, lineage, and author testimony are labeled separately. Sources can corroborate, challenge, or bound the argument; they do not replace Tony Malott's judgment.
Take the complete artifact with you.
The deterministic package contains a self-contained offline article, the exact public-route snapshot, canonical public metadata, receipt, source text when available, plain-text context, claim ledger, source records, and a member-hash manifest.
Sources, authority, and lineage
Each record states the role it plays. Research support and governance provenance are not treated as interchangeable.
Candidate article: Innovation Does Not Have to Be Big
Governs the locked thesis, manuscript, tone, claim boundaries, and public publication.
Governs the locked thesis, manuscript, tone, claim boundaries, and public publication.
Open sourceOwner approval to reconcile and publish
Authorizes materially faithful exact-main reconciliation, merge, and normal Production publication.
Authorizes materially faithful exact-main reconciliation, merge, and normal Production publication.
Open sourceWhat is asserted—and how it is bounded
Research, author analysis, and personal testimony remain distinct. Supporting links and caveats stay attached to each claim.
Innovation is the disciplined act of noticing friction and improving it.
Agents perform bounded assistive work under human direction and review.
Public boundary. No employer-specific context, internal licensing or tenant details, invented metrics, private conversations, or autonomous authorship claims are published. The public article is a distinct canonical Work from the protected Innovation Made Visible and Take the Mantle experience.