TechFides — May 2026
A COO at a mid-market financial services firm told me last month that her organization had spent $340,000 on an "AI governance platform" 18 months ago. She walked me through the dashboard. It was a beautiful product. Risk registers, policy templates, model inventories, audit logs.
Then she paused. "We still don't have AI in production."
I have heard a version of this conversation three times in the last six weeks. Different industries. Different vendors. Same outcome.
This is the part of enterprise AI that vendors do not talk about clearly, because their incentive structure does not let them. AI initiatives do not fail because the technology is hard. They fail because the change is hard. And no software product, no matter how well-designed, governs change.
This is the gap that AI Transformation Management fills, and it is the reason TechFides built it as a resource-based engagement rather than a SaaS product.
The 70/30 rule that nobody briefs the board on
Every major AI initiative in the enterprise has two components: the technology that needs to be installed, and the change that needs to happen around the technology so it produces value.
The technology component is about 30 percent of the total effort. Models, data pipelines, infrastructure, integrations — all of it can be planned with reasonable predictability by a competent engineering team.
The change component is the other 70 percent. Stakeholder alignment across business units. Workforce training and capability transfer. Regulatory adaptation in coordination with legal and compliance. Communications strategy for parliament, board, donors, citizens, or shareholders. Process redesign for workflows the AI will touch. Phased sequencing that respects the political and administrative calendar.
This 70/30 split is not a TechFides observation. It comes out of Six Sigma and PMP research on enterprise transformation over the last two decades. The number has been stable across industries: roughly 70 percent of transformation failure traces to change-management deficits, not technical execution.
The implication is uncomfortable for vendors selling AI products. If 70 percent of the work is change, then 70 percent of your investment should be in people who manage change. The platform is the tool. The transformation is the work.
What software does well — and what it does not do at all
I want to be fair to the AI governance software category. The category did not invent itself for no reason. There are real things it does well.
Inventory and visibility. A good platform can tell you which models are running, who is using them, what data they touch, and what their risk classifications are. This is genuinely useful.
Documentation at scale. Policies, runbooks, training materials — the platform stores them, versions them, surfaces them. Better than a SharePoint folder.
Audit trail. When the regulator asks who approved which decision, the platform can answer. This is increasingly non-optional.
Reporting templates. Board-ready dashboards, regulatory submissions, internal KPIs. The platform does this faster than humans would.
These are real capabilities. The category exists because the capabilities are real.
Here is what no AI governance platform on the market does, will do, or can do.
It does not make a skeptical VP sign off on a use case. That is a relationship.
It does not get an internal audit team to sign a memo declaring the deployment compliant. That is a coordinated effort across three departments, four meetings, and a senior advocate who understands the audit team's actual concerns.
It does not negotiate the workforce-impact language in the union contract. That is six months of bargaining.
It does not write the script your CIO uses in front of the board when the board asks why this initiative is different from the last two that failed. That is judgment shaped by 25 years of running transformation programs.
It does not move a 90-day implementation from "slipping by two weeks" to "delivering on time" by reallocating resources between workstreams in real time. That is program management.
If you map the items on these two lists, the gap is obvious. Software handles the parts of the work that are documentable. Senior resources handle the parts of the work that determine whether the initiative succeeds.
What "Transformation Management" actually is
TechFides AI Transformation Management is a resource-based engagement. You are not buying software. You are buying people.
Each consultant on a Transformation Management engagement carries three things at once.
Domain functional expertise. Years operating inside the relevant sector — financial services, healthcare, public sector, manufacturing, professional services. This is non-negotiable. A senior consultant who has never run a transformation inside a regulated industry cannot manage a regulated-industry AI transformation, no matter how strong their general consulting skills are.
AI fluency at the technical decision-making level. Not engineering depth — that is what the technical team is for. But enough fluency to evaluate model selection, deployment architecture, data engineering choices, and governance posture in real time during steering committee meetings. Without this, the senior resource becomes a translator, not a decision-maker.
Certified program management discipline. PMP, PgMP, ITIL, Six Sigma. These are not certifications for the sake of them. They are the operating language that lets a senior resource step into an enterprise program management office, run the cadence, manage the dependencies, and hand off a defensible plan to your internal team at the end of the engagement.
The combination of these three is rare. It is what makes Transformation Management more expensive per hour than software, and more effective per dollar than software.
The four engagement shapes — and which one is which
Transformation Management ships in four configured engagement tiers, each shaped to a specific organizational need.
Strategic Advisory ($50K to $95K, 6-month cycle). For executive teams who need ongoing access to senior judgment without a full embedded engagement. Quarterly strategy sessions, ad-hoc decision support, board-ready deliverables. Best for organizations where the AI strategy is forming and the leadership wants a sparring partner.
Transformation Management ($150K to $350K, embedded). For organizations executing an AI initiative right now. Senior domain experts embedded with the client team. The client leads. The TechFides resources support, structure, document, and shepherd. Best for organizations that have decided what they want to do and need someone who has done it before to keep the program on plan.
Capability Building ($175K to $350K, 18-month cycle). For organizations setting up an internal AI Center of Excellence. Talent development, process standardization, IP retention. Best for organizations that intend to bring AI strategy in-house permanently and need someone to build the function before stepping out.
Rescue & Turnaround ($85K to $200K, 90-day sprint). For organizations whose AI initiative is stalled, over budget, or politically toxic. Root-cause analysis, remediation plan, momentum restoration. Best for organizations whose previous engagement or internal program has lost the plot and needs to be put back on the rails before the budget cycle closes.
The four tiers exist because organizations need different things at different points. The COO with the $340,000 governance platform and no AI in production needed Rescue & Turnaround. The organization that has already done a successful pilot needs Transformation Management to scale it. The organization that has not started yet needs Strategic Advisory to map the path.
How this fits with AEGIS and AI Readiness 360
Transformation Management is one of three TechFides offerings under the same operating logic: diagnose, execute, govern.
Diagnose is AI Readiness 360 — a 60-question network-level diagnostic. Two-week delivery. Defines the actual problem before anyone commits a budget.
Execute is AI Transformation Management — embedded senior resources who run the program from kickoff to handoff.
Govern is AEGIS — the operating system that institutionalizes everything after the engagement ends, so the governance and operating model survives every administration change.
Most engagements move through all three over a 12-to-24-month horizon. Some organizations enter at AI Readiness 360 and stop after the diagnostic. Some enter at AEGIS because they have already executed and need governance retroactively installed. The sequence is not mandatory. The relationships are.
The single decision that changes the math
If you are reading this and your organization is in the middle of an AI initiative that is not delivering, the question to ask in your next leadership meeting is small but specific.
"Are we under-resourced on the change side, or under-resourced on the technology side?"
If the answer is the technology side, you need more engineers or a different platform. Most platforms can solve this.
If the answer is the change side — and based on the 70/30 rule, it usually is — you do not have a platform problem. You have a senior-resources problem. You need fewer dashboards and more people who have run this exact program three times before.
That is the gap AI Transformation Management closes. And that is the gap no software can.
See the AI Transformation Management engagement model with the four engagement tiers laid out in detail.
For governments running the same transformation logic at national scale, see the TechFides Government practice.
Like this? Get the next one Wednesday.
One email per week. No marketing filler. Unsubscribe anytime.