Discipline is the product

When I started Tactical in 2017, the standard agency pitch was "we use marketing automation." The implication was that the tooling itself was the differentiator. Most prospects nodded along, because most of them were not yet using marketing automation, and the gap between "uses Act‑On" and "does not use Act‑On" was visible.

That gap is gone now. Every agency uses marketing automation. Every agency uses AI. The pitch decks are roughly interchangeable: a slide on platforms supported, a slide on AI capabilities, a slide on industries served. The reader has no way to choose between them on capabilities, because the capabilities are the same.

What separates the agencies that last from the ones that quietly disappear after a year is not the capability list. It is the discipline applied to the capability. That has been true since we started, and it is more true now than it was then, because AI has compressed the cost of acquiring the capability while leaving the cost of acquiring the discipline exactly where it was.

What I mean by discipline

Discipline in marketing operations is not aspirational. It is a small set of habits that determine whether the work survives contact with reality.

Documenting workflows before you ship them. Naming segments so a person joining the team in nine months can read the name and know what is in the segment. Writing field descriptions in the CRM so the field has a stated purpose. Maintaining a single source of truth for the contact record. Defining what counts as a marketing‑qualified lead and not changing the definition every quarter to make the dashboard look better.

None of this is glamorous. None of it shows up in a pitch deck. It is also the difference between a marketing operation that compounds - every quarter the system is more useful than it was the quarter before - and one that runs on a treadmill, rebuilding the same broken segmentation every twelve months because nobody documented the last one.

Why AI made this harder, not easier

The expected story was that AI would automate away the boring discipline work. Workflow documentation generates itself. Field descriptions generate themselves. Naming conventions enforce themselves. In practice, it has not played out that way.

What AI actually did was lower the floor on production volume. Teams now produce three times as much work, in roughly the same number of hours, with roughly the same number of people. The discipline practices were already under‑resourced at the old volume. At the new volume, they are deeply under‑resourced. The result is not that the discipline got automated; it is that the discipline got skipped entirely on the new tranche of work.

I am cleaning this up, in slightly different shapes, in most of our active engagements. The team three years ago had two hundred contact records with inconsistent fields. The team today has eight thousand contact records with inconsistent fields, generated automatically by a workflow nobody wrote down. The original problem is the same. The volume is harder.

The firms that will still be here in five years

I do not know exactly what the agency market looks like in 2031, but I have a guess about the shape. The firms that pitched themselves on AI capabilities will mostly have repositioned, repeatedly, every time the capability set changed. Some of them will still be standing. Many of them will not.

The firms that pitched themselves on operating discipline will be roughly where they are now, with a slightly larger book of business. Their pitch will not have changed because the underlying argument has not changed. The clients who hire them are looking for the same thing they have always been looking for: a partner who treats their CRM like a long‑lived system that has to keep working, not a content machine to point at a quarterly target.

I am biased toward this view because it describes what we do. I am also reasonably confident in it because I have watched two prior cycles play out the same way. The agencies that pitched "we do social media" in 2010 and pitched "we do mobile" in 2013 and pitched "we do programmatic" in 2016 mostly did not survive the next cycle's reframing. The ones that pitched "we operate your marketing engine well" in 2010 mostly still exist.

What this looks like for a buyer

If you are evaluating agencies right now and the pitch you are getting is heavy on AI capabilities, ask three questions before you sign anything.

How will you document what you build inside our instance? If the answer is a vague reference to "internal documentation," push for specifics. If the answer is "we will produce a runbook your team owns at the end of the engagement," that is a different posture.

What is your default position on adding new fields, workflows, and integrations to a stack we already own? An agency that wants to add things on day one is going to leave you with a more complicated instance than you started with. An agency that wants to remove things first is doing the right work.

What does your QA process look like before something goes to production? "We test it" is not an answer. The actual answer involves named reviewers, written checklists, and a deliberate slower step between build and ship. Teams that cannot describe this in operational terms do not have it.

These questions are boring. They will also tell you, in five minutes, which agencies have discipline and which are decorating.

Where this leaves us

Tactical Marketing's positioning has been the same since 2017. The current version of it - AI is your accelerant, not your driver - is just the 2026 phrasing of an argument we have been making the whole time. The accelerant is whatever the current capability layer happens to be: marketing automation in 2018, predictive lead scoring in 2020, generative AI in 2024. The driver has always been operating discipline, applied by people who have been doing this for a while, against systems that have to work on Monday morning.

If that is the kind of partner you are looking for, we are reachable. The conversation is short, low‑pressure, and usually starts with what you already own.