Gathering information
We run surveys in SurveyMonkey. We record meetings, and tools like Granola and Plaud turn the talk into written transcripts.
The team's own words become evidence it can return to and check.
Our approach
Gestalt Empowerment Evaluation
Gestalt Empowerment Evaluation helps staff and community members see the whole picture of their work. It matters most when key pieces are hidden, scattered, or hard to read.
The whole is greater than the sum of its parts
Dr. David Fetterman and Dr. Paul St. Roseman developed Gestalt Empowerment Evaluation together. It builds on empowerment evaluation, which puts the tools of evaluation in the hands of the people doing the work. The goal is lasting skill inside your organization, not reliance on outside experts.
The Gestalt part adds a simple idea. A single number rarely tells the full story. Meaning comes from how the pieces relate to each other. So the approach stresses three things: (a) seeing patterns, (b) keeping context whole, and (c) making sense of the work together.
Culture, history, and lived experience shape what counts as evidence. For that reason, cultural responsiveness is part of the design from the start. It is not an add-on.
The four steps
How the work moves
Empowerment evaluation is used in more than eighteen countries. It usually runs as three steps, with monitoring as an ongoing fourth.
-
Mission
The group agrees on its values before it sets any measures.
-
Taking Stock
The group ranks its most important activities. Then it rates each one and talks through the rating.
-
Planning for the Future
The group sets goals, chooses strategies, and decides what evidence will count.
-
Monitoring
The group tracks its progress. The point is to see where support is needed, not who to blame.
The conversation is the heart of it
When a team explains why it rated an activity a 4 or an 8, evidence comes to the surface. The team also questions its "sacred cows." Additionally, the group learns to focus its energy on a few activities. The method calls this a "dose effect." A few things done well beat many things done thinly.
A natural flow, not a fixed script
The steps stay, but they do not have to happen in a fixed order. A team may clarify its mission while it ranks priorities. Planning may send the group back to its values. We call this two steps forward and one step back. The step back is not a failure. It is often where the deepest learning happens.
The tools we use
Online tools that carry the method
We are building online tools that carry Gestalt Empowerment Evaluation from start to finish. Each tool matches a step of the method. Many organizations can work in them at once, and each one keeps its own data separate.
| Step | What the team does | What the tools provide |
|---|---|---|
| Mission | Agrees on values and priorities | A shared space where the team records what it values |
| Taking Stock | Rates its key activities and talks them through | Surveys and ratings shown next to the team's own priorities |
| Planning for the Future | Sets goals and decides who does what | Goals, partner roles, and decisions, ready to send to a task tool like Asana |
| Monitoring | Tracks progress over time | Dashboards, a history of changes, and flags where the evidence is thin |
Taking up AI one step at a time
AI works best when a team takes it up in steps. That way the team builds confidence and skill as it goes. Handing a complex system to a team that is not ready can overwhelm it. We usually follow this path:
- Personal AI assistants. A team asks questions of its own chosen data, and nothing else.
- Shared libraries (example: NotebookLM). A team gathers trusted materials for training and shared understanding.
- Full evaluation tools (example: Rural Senses). These draft logic models, early findings, and reports a team can edit. Rural Senses is part of DCG's own set of tools. We use it to learn and to build an evaluation information system.
What we work with
Everyday tools, chosen for the method
Gestalt Empowerment Evaluation asks a team to see the whole picture. That is hard when the pieces sit in a dozen places. So we choose familiar tools that bring the pieces together, and that your staff can run on their own.
Keeping it safe and organized
Teams keep shared files in Google Drive. We are adding a secure shared database with real logins and clear rules for who sees each record.
When it is done, each organization's data will stay apart from every other organization's data.
Making sense of it
AI helps us read transcripts faster and draft analytic memos, which are short notes on what the data shows. Each memo cites its sources.
A person reviews every memo before it is final, so each claim stays tied to its source.
AI helpers
The coach inside the ARCHES ACDT Coaching Workbook runs on Claude, an AI assistant. NotebookLM holds a team's trusted materials so the team can ask questions of them. Rural Senses drafts logic models and early findings that a team can edit.
The AI offers ideas and asks questions. The team decides.
Sharing and acting on it
We build dashboards in Power BI, a tool for charts and reports. We also build scorecards, like those in the ARCHES ACDT work. Next, we are building a way to send a finished plan to Asana, a task tool.
Findings turn into plain stories, visuals, and next steps the whole team can see.
How we choose tools
- Your staff can run them after we leave. We coach and write how-to guides, then hand the tools over.
- Your data stays yours. It belongs to your organization, and you can ask us for a full copy at any time.
- Each tool serves a step of the method. We do not add a tool just because it is new.
- We start simple. We add tools as the team gains skill and confidence.
Where AI fits
AI helps people see. People decide.
AI is good at spotting patterns across large amounts of information. That work can overwhelm a team that is short on time and staff. However, AI can also repeat old power imbalances. So we set clear limits on its role.
What AI does
- Finds patterns and gaps in the team's own data
- Points out repeated or missing information
- Drafts summaries and visuals that pull the pieces together
- Prompts new questions and reflection
What AI never does
- Defines a community's values
- Decides the merit or worth of the work
- Replaces conversation or group judgment
- Overrides the team's reading of its own context
Every claim stays tied to its source. AI never uses a quote it cannot check. It works as a thinking partner, while the meaning stays in human hands.
Working examples
The approach in practice
The ARCHES ACDT Coaching Workbook
DCG built and runs this online coaching tool for ARCHES. It draws on a library of partnership strategies and walks a team through its own plan, step by step. The finished work exports to PDF or Word.
It answers a common problem. Teams write a plan for a grant, then never connect it to the work. As one ARCHES planning call put it, the plan becomes "a great ornament." The workbook keeps each page tied to the team's own priorities.
The Just Trust
The Just Trust is a $350 million criminal justice reform effort that works across seven states. There, evaluation was part of daily work, not an outside review. Staff from across the organization helped: (a) collect data, (b) analyze it, and (c) make sense of it.
At a retreat, leaders used a funnel image to sort their priorities. The new frame cut scope and helped the group let go of work that no longer fit. Later, staff used an AI assistant that drew only on their own chosen data. In their words:
"It's answering my questions." · "It's making me ask questions I wouldn't have thought to ask."
The team still read the results together and talked them through. The team owned the meaning.
Where this comes from
Three decades of practice
Gestalt Empowerment Evaluation grows out of DCG's work since 1995. Our early work centered on data management, which we see as a cycle of five steps. The steps overlap and repeat. Each round shapes the next one as a program matures. We call that growth data use.
Dr. St. Roseman's research added another piece. It found that evaluation builds the most skill when people take part across the organization. The gains deepen when people are motivated, connected, and trusting. Those findings shape how we coach today.
- Develop
- Calibrate
- Collect
- Analyze
- Inform
See how this could work for you.
Every organization starts in a different place. We will help you find the right first step.