Confluence Policy & Strategy Group

In November 2018, just over a year after Hurricanes Irma and Maria devastated Puerto Rico, we held a Workforce Summit that brought together the Governor's Office, every state agency, and dozens of employers from local businesses to Marriott, Microsoft, and World Central Kitchen to create real career pathways for Puerto Rico's students.

More than 1,400 students were matched with employers for work-based learning in the first year. More than 800 were guaranteed a living wage job upon completion. The Clinton Global Initiative recognized the effort with an Ideas Into Action award in 2019.

Last year, one of those students wrote to tell me she's now a manager, mentoring her own interns. That note is pinned above my desk.

The full story and what it means for how we approach every engagement since is on the blog.

When the Island Had to Rebuild Everything at Once Hurricanes Irma and Maria made landfall in Puerto Rico within two weeks of each other. By the time I arrived on the island, the damage was not just physical. Students were returning to schools that in many cases could not reopen, in communities whose economic futures looked nothing like they had twe...

There is a reason fire extinguishers say "break glass in case of emergency." You do not reach for one for an everyday problem. You reach for it when the normal options are exhausted and the situation cannot wait.

Colorado passed an AI law that almost nobody at the table actually wanted. Two further efforts, including a special session, tried to rewrite it and could not build agreement among the stakeholders who would have to live with it. The next attempt ran through a structured process. Every party in the room, consumer advocates, civil rights organizations, technology companies, business associations, ended up supporting the result. It passed the Senate 34-1, the House 57-6, and was signed into law thirteen days after introduction.

This is exactly what Confluence was built to do. Not just for the moment when everything is on fire, but for the room that has time, no crisis forcing anyone's hand, and a problem too complex for the people already at the table to untangle on their own.

The Colorado Educator Safety Task Force is the second kind of room. Nothing forced that work onto anyone's desk on a particular day. It met monthly for nearly a year. What changed the outcome was not urgency. It was a neutral facilitator with no stake in the answer, and a process built to bring in more than the usual insiders. More than 1,100 educators were surveyed. More than half reported being physically injured by a student, a finding that contradicted every existing data set. People said the true thing because the process made it safe to.

The Working Group on Transforming Criminal and Juvenile Justice is a third kind. A twenty-year-old commission had sunset with no agreement on what should follow. Prosecutors, defense attorneys, law enforcement, and reform advocates, people who had spent decades with fundamentally different visions for the system, reached consensus on the mission, structure, and membership for two new permanent bodies. Historic adversaries, at the same table, agreeing on what comes next. That is what the process is built to produce.

And then there is Puerto Rico. Later this week I'll share the story of a school system rebuilt after two hurricanes, where everything had to happen at once and nothing was stable. That is a room with urgency and opportunity at the same time, lives upended, and a once-in-a-generation chance to rebuild something better than what existed before.

Emergencies. Complex rooms with time to get it right. Rooms where both are true at once. Different conditions, same need: a process designed on purpose, not assembled on the fly, by people who run this work and nothing else.

If you are looking at a room like any of these, whatever else is true about it, I'd like to talk about what that could look like.

Managing budgets is tough work - on the expense and revenue side - and always involves difficult decisions. Authentically engaging those directly affected makes the decisions not only improved but-in, it more importantly improves the decisions themselves. Honored to help lead these conversations.

There is a version of regulatory failure that does not come from bad intent or political gridlock. It comes from speed.

Cannabis was legalized in Colorado before most agencies and local jurisdictions had frameworks to handle tax structure, banking access, or municipal revenue dependence. Now psilocybin, ibogaine, and other therapeutic substances are following the same arc and the policy infrastructure is still catching up to the clinical evidence.

AI is reshaping how students learn, how employers hire, how cities plan transportation infrastructure, and how natural resource managers model fire risk and water supply. Most of the regulatory frameworks governing those domains were written for a different world.

Workforce development programs are preparing people for jobs that will look substantially different by the time a two-year certificate is complete. The curriculum approval process in most states runs on a different timeline than the labor market.

We work in all of these rooms. The common thread is not the subject matter. It is the condition: people with genuine expertise and genuine responsibility trying to make durable decisions about something still in motion.

The private sector is navigating the same problem. Compliance leaders, industry associations, and executive teams face the same gap between what they know today and what the regulatory environment will require in two years. The instinct is often to wait for clarity. That rarely works either.

What helps is a process specifically designed for distributed knowledge and moving targets, one that surfaces what the room does not yet know it needs to know before forcing a decision. Not because the facilitator is a subject matter expert in all of these domains. No one is. But because the structure of the problem is the same across all of them.

Government does not have a monopoly on this challenge. Every sector navigating rapid regulatory change is dealing with a version of it.

If your organization is sitting in one of these rooms right now, we’d be happy to talk about what that kind of process looks like in practice.

Over 50 percent of educators surveyed told us they had been physically injured by a student.

74 percent said a student had attempted to physically injure them or another adult in their presence. Many said it had happened more than once. Some said it happens every day.

Those numbers were not in any existing data system. They came from educators who had largely stopped expecting to be asked.

In the fall of 2024, Confluence PSG designed and facilitated the Colorado Educator Safety Task Force, a 17-member group that spent eight months examining what Colorado schools had not answered well: how safe are educators, really?

The Roadmap for Action the task force produced is built so that educators, building leaders, district administrators, state agencies, and preparation programs can each act on the findings with the authority they already have, without waiting for anyone else to move first.

The findings also changed what state agencies believed was true about the problem, and that changed how they work.
Read the full post, access the Roadmap, and hear from educators in their own words at the links below.

Thank you to the Office of School Safety at the Colorado Department of Public Safety & the Colorado Department of Education for your continuous commitment to the well-being of educators.

Colorado Educator Safety Task Force Members' Interviews The Colorado Educator Safety Task Force (ESTF) was formed by the st...

Policy take: triggers not principles

On technically complex policy questions, you can get a room to agree on principles faster than you might expect.

Protect consumers from harm caused by AI. Do not create compliance burdens so onerous they chill innovation. Make the rules clear enough that businesses can actually follow them. Hold the right parties accountable when something goes wrong.

In a multi-stakeholder process I recently facilitated on AI governance, there was broad agreement on all of those within the first few sessions.

Then we started writing definitions.

The fights that followed were not about whether consumers deserved protection. Everyone agreed they did. They were about what "materially influences" meant when an AI system is one of several inputs into a decision a human still makes. They were about whether liability for AI-related harm should be broad and general or specifically anchored to violations of existing anti-discrimination law. Every word in each of those definitions determined who the law reached, what it required, and what it cost to comply.

The principle was easy. The trigger was nine months of work.

This is the dynamic that makes technically complex policy so hard to get right, and so easy to get wrong. Legislation that resolves at the level of principles feels like agreement but produces chaos in implementation. The compliance officer trying to build a program around it cannot tell her team what they actually have to do. The regulator trying to enforce it cannot explain what a violation looks like. The consumer trying to invoke it cannot tell whether the harm they experienced is covered.

The real test of any complex policy framework is not whether the principles command support. It is whether the definitions are precise enough to govern behavior, specific enough to enforce, and fair enough that the parties who negotiated them can all live with the result.

That last part is harder than it sounds. Precise enough to govern and fair enough to hold a coalition together are not always the same thing. Sometimes they are. When they are, you have something worth passing.

Most negotiation processes start with someone's first draft. Everyone else spends their time pushing back against it rather than building something together.

The CDMP approach we use at Confluence PSG starts somewhere different: with what every person in the room actually cares about and why. Not their position. Their reasoning.

Two recent processes, one on transit governance reform and one on AI policy, show what that looks like in practice and why the sequence matters more than most people expect. The horse trading that produced results nobody loves and everybody likes came later. The work that made it possible came first.

New post on the Confluence PSG blog:

Nobody Loves It. Everybody Likes It. Most groups that are brought together to solve a hard problem want to start negotiating immediately. That instinct is understandable. The people in the room are experienced. They have positions. They have organizational mandates and limited time. They are ready to get to work. The problem is that th...

New blog post: A governance process that jumps from disagreement straight to proposed solutions almost always fails. People vote on proposals before they share an understanding of the problem. The vote becomes a proxy for the underlying dispute rather than a resolution of it.

The Governance Question Most Rooms Never Ask A 14-member committee convened to answer one of the harder questions in regional governance: what would it actually take to make a large, complex, and struggling regional transit agency work the way it needs to?

On Monday, the Colorado General Assembly passed SB26-150, the Modernizing the Regional Transportation District Act, legislation built directly on the recommendations of the RTD Accountability Committee, a 14-member body charged with one of the harder governance questions in regional transit: not just what to change, but whether the existing structure was even the right thing to fix.

That question is harder than it sounds. The committee did not disagree about whether RTD faced serious challenges. They disagreed about what the challenges meant. Two foundational tensions ran through nearly every conversation. The first: the current structure is not working. The second: democratic legitimacy must be protected. Both were true. The dispute was about what each one required.

Inside those two positions lived a harder nested question. If the structure is not working, is it the size of the board, the way members are recruited and supported, or the people themselves? And if democratic legitimacy is non-negotiable, what does that actually demand? Full election, some appointment, or something in between, and who gets to decide?

A governance process that jumps from that level of disagreement straight to solutions almost always fails. People vote on proposals before they share a diagnosis, and the vote becomes a proxy for the underlying dispute rather than a resolution of it.

What moved this committee was sequence. Before anyone proposed a solution, we spent sessions on a single question: what problem are we actually trying to solve? From there, we generated ideas for consideration rather than proposals, a deliberate step that lets people explore options without committing to them. We brought in evidence: data on elected versus appointed boards across peer agencies, testimony from transit board members in other regions, research on board size and governance effectiveness. Members who came in with strong positions found themselves asking different questions once they had a shared factual foundation.

Then came the motions. On board structure alone, the committee considered six different configurations over two sessions. A motion for a 9-member board saw a vote of 6 to 5 with three members not present in the first session, two votes short of the threshold set in the Committee bylaws. The same motion passed 11 to 2 the following week. The difference was not persuasion in the conventional sense. It was that the room had moved from debating positions to examining a problem together, and when that shift happens, people can change their minds without feeling like they lost.

The legislature examined the committee's process closely. They heard extensive testimony, were lobbied hard from multiple directions, and had their own questions about the 9-member hybrid board the committee ultimately recommended. They passed the bill.

This pattern is not specific to transit. Any governance challenge where the people in the room disagree about what is broken before they disagree about how to fix it requires the same discipline: define the problem before you generate solutions, generate options before you force votes, and build a shared evidentiary foundation before you ask people to commit. The structure of the problem is the same whether the room is a transit agency board, a corporate governance review, a nonprofit in crisis, or a public-private partnership trying to decide who is actually in charge.

More on what that process looked like and what it produced on Thursday.

Bill text and history: https://leg.colorado.gov/bills/SB26-150