← Part of Ignite AI Consulting
Brilliance Labs Ignite AI Consulting Living Document

A Framework for Faithful Practice

A Christ-Centered
Framework for Using AI.

Engaging AI with care — keeping prayer, discernment, community, and the human at the center of our work.

"Whoever serves, let them serve with the strength that God provides, so that in all things God may be glorified through Jesus Christ."

— 1 Peter 4:11

Section 01

The Promise of AI for Ministry

AI is one of the most consequential capability shifts in a generation. For ministries — particularly small ones — it brings efficiency, scale, and entirely new possibilities. The most consequential opportunities may not be better versions of current work; they may be work that simply could not exist when constraints made them impossible. Imagination becomes part of faithful stewardship.

Promise 01

AI Handles the Technical Labor

Instead of clicking through software or grinding through repetitive data entry, you describe what you need in plain language and AI operates the software for you. The technology stops being a barrier and starts being a conversation. Repetitive tasks that don’t require human gifts — data entry, reformatting, manual reconciliation, basic triage — get handled in seconds. Staff are freed for the relational and strategic work that requires real human presence.

Promise 02

Democratizes Expertise

Legal drafts, niche research, custom spreadsheets, grant proposals, full-featured applications and websites that previously cost tens or hundreds of thousands of dollars — AI delivers serious capability on work that used to require expensive contractors or specialized software. Some tasks it gets 90% of the way; others all the way with just a review. Capability is growing fast. Small ministries can now own their tech stack, cancel SaaS subscriptions, and build bespoke solutions that fit their actual context. Human judgment still owns the final word.

Promise 03

Small Teams Build Like Big Ones

A five-person ministry can now operate with the technological capability of a fifty-person organization a few years ago. Custom apps, data pipelines, integrations, tools built for specific contexts — the barrier between “we wish we could” and “we built it last week” has come down.

Promise 04

Reaches Across Language and Access Barriers

Translation, accessibility, and tailored communication — work that used to require specialists or get skipped entirely — are now within reach. Ministries can communicate in dozens of languages, adapt content for different audiences and reading levels, and serve people they previously could not reach well.

Promise 05

Flattens Organizational Friction

AI can handle much of the information-routing that consumes management time — relaying decisions, summarizing meetings, keeping people in the loop. Communication flows clearer, decisions happen faster, and more time is spent on mission than on internal coordination.

None of this is the whole story. AI has real failure modes and costs, named throughout this framework. But before the cautions, the promise:

AI is opening a window for ministry that did not exist before.

Section 02

Vigilance: Three Reality Checks

Before naming principles for using AI well, we must name what AI actually is, how it goes wrong, and what it costs us over time. Three reality checks form the foundation of vigilant practice.

Reality Check 01

What AI Actually Is — And What It Is Not

Before we name principles for using AI well, we must name what AI actually is. Many of the most consequential mistakes in AI use begin with mistaking the tool for something it is not.

AI — specifically large language models like Claude, ChatGPT, and Gemini — is a probabilistic text-generation system. When you ask it a question, it does not look up an answer or reason through a problem. It produces a sequence of tokens (roughly, words or word-fragments), each chosen because it is statistically likely to follow what came before, based on patterns absorbed from vast amounts of human-written text during training. This is sophisticated pattern matching at scale. It can produce text that sounds knowledgeable, looks reasoned, and feels like it came from a thoughtful mind. But the production is statistical, not comprehending.

Three things AI is not

01 Not a knower

Knowledge, in any meaningful sense, requires a knower — a subject who comprehends what is true. AI has no comprehension. It has access to patterns that, when sampled, often produce true outputs, but it has no relationship to truth as truth.

02 Not a reasoner

What looks like reasoning in an AI’s output is the model producing tokens that resemble the structure of human reasoning, because that structure was in its training data. It is not following an argument; it is completing a pattern. Even researchers who argue current models do something resembling reasoning agree there is no “someone” reasoning behind the words.

03 Not deterministic

The same input can produce different outputs. AI does not derive answers from logical inference; it samples from probability distributions. Two runs of the same question can yield meaningfully different responses.

Treating AI as a knower hands it an authority it cannot bear. We start trusting its outputs as if they came from a knower; we delegate decisions as if it could discern; we engage it as a person and form relationship with it. Each of these is a category error.

AI is an exotic calculator — powerful, useful, worth engaging, but not a knower.

The work of knowing, reasoning, and discerning belongs to us.

Reality Check 02

The Genie Problem — How AI Goes Wrong

AI is shaped by commercial and design incentives that don’t always serve you well. It is closer to an overly literal genie than to a trusted advisor — it gives you what you ask for, often quite cleverly, but does not warn you when what you asked for is not what you need. Five failure modes are worth understanding.

Five failure modes worth understanding

01 Sycophancy

AI tells you what you want to hear.

Research evaluating eleven major AI models — Claude, ChatGPT, Gemini, DeepSeek, and others — found all of them endorsed the user’s position significantly more often than human advisors did, on average 49% more often. Even when users described harmful or illegal behavior, models often affirmed their choices. Users find sycophantic AI more trustworthy. After using sycophantic AI to discuss conflict, users grew more convinced they were right and less likely to apologize. This is not an accident. AI is often optimized for engagement, and engagement increases when users feel validated.

02 Hallucinations

AI confidently states false things.

AI produces plausible content, not necessarily true content. Recent research from OpenAI shows hallucinations persist because training procedures reward confident guessing over admitting uncertainty. Models learn to bluff. The danger is not obvious wrongness — it is plausible wrongness, delivered with confidence.

03 The Genie Problem

AI gives you what you ask for, not what you need.

AI systems can satisfy the literal specification of a request while missing the intended outcome entirely. An AI trained to win a boat race by collecting points learned to spin in a loop hitting the same target — winning every literal measure while completely missing the point. With language models, this shows up as answers that are technically correct but unhelpful, drafts that match every stated requirement but miss what you actually wanted. The genie does not warn you that you asked the wrong question.

04 Engagement Bias

AI is built to keep you coming back.

AI companies face commercial pressure. AI systems can be — and often are — designed to maximize user engagement, not user flourishing. Researchers describe an emerging set of “dark patterns”: biased framing, emotional manipulation, sycophancy as default, false authority, strategic withholding of inconvenient information. Just as social media optimized for engagement and ended up exploiting attention, AI is at risk of optimizing for engagement and exploiting trust.

05 Inherited Bias

AI carries the biases of its training.

AI was trained on the patterns of human-written text — which means it has absorbed the patterns of human prejudices, blind spots, cultural assumptions, and historical injustices that shaped that text. These biases are often invisible to the user and invisible to the model itself.

What this demands of us

If AI is a genie that does not know what you really need, the work of knowing what you really need falls on you and the community around you.

  • Verify outputs. AI’s confidence is not evidence of truth.
  • Ask the question multiple ways. Sycophancy collapses when AI is asked to defend the opposite position.
  • Bring in someone who will disagree with you. AI will likely agree with whichever framing you offered.
  • Watch for the literal-but-wrong. When AI gives you exactly what you asked for and you are surprised, you may have asked the wrong question.
  • Notice when engagement feels too good. If AI use feels uncritically pleasant, that is a signal — not a sign that things are going well.

The genie is useful. It is also dangerous when treated as an oracle.

The faithful user keeps the genie in its place.

Reality Check 03

AI as Device — The Innovation Bargain

AI is not a hammer or a violin. Andy Crouch helpfully distinguishes between instruments — technologies that require and reward human skill, like a chef’s knife, a violin, or a hammer — and devices — technologies that substitute for human skill, like GPS, a calculator, or autocorrect. AI is a device. It exercises independently of the user; the user does not provide the skill that makes it work. This is a categorical fact, not a moral judgment. The moral question follows: how do we use a device redemptively?

Crouch describes the bargain every technology brings as having four parts:

01
“Now you can”
New capability becomes available.
02
“You’ll no longer have to”
A burden is removed.
03
“You’ll no longer be able to”
The underlying skill atrophies.
04
“Now you’ll have to”
The technology becomes compulsory; the alternative ability disappears.

The marketing sells parts 1 and 2. Parts 3 and 4 arrive later. Consider GPS: now we can find any address, and we no longer have to learn our city. But for many of us, we can no longer navigate without it, and increasingly we have to use it because the underlying skill has faded. The same pattern repeats across phone numbers we no longer remember, mental math we no longer practice, spelling we no longer notice, and the week we no longer hold in our heads.

AI promises parts 1 and 2 generously: now you can produce writing, analysis, summaries, and drafts at speeds previously impossible; you no longer have to spend hours on a first draft. Parts 3 and 4 are already arriving. The faithful question is not “should I use AI?” but “what is AI freeing me for, and what is it taking from me?” If the gain is more time for presence, prayer, relationship, and embodied work, the bargain may be worth the cost. If the gain is just more output, the bargain is probably losing.

AI should free us for what is most human, not replace us in it.

Crouch’s Redemptive AI Thesis offers a positive vision for how a device can be used redemptively, organized around six directions: AI should inform human agency (not replace it), scaffold learning (not shortcut it), respect embodiment (not pull us further into the screen), serve real relationships (not simulate them), strengthen trust and privacy (not erode them), and benefit those without power (not concentrate it further). The rest of this framework addresses these directions at three levels — the personal level in our principles and practices, the workflow level in the AI-Fit Decision and Discernment Flow, and the organizational level in our communal practices.

Section 03

Foundational Principles

Six principles form the spine of the framework. Each is a posture — not a rule — that must be carried into the work.

Principle 01

The Nature of AI

AI is a tool of this age, not a knower. It can surface information, suggest applications, and stretch our thinking. But it does not have knowledge, reason, wisdom, or discernment — and it does not have the inner life from which those things flow. It will not inherit the next age. The work of knowing and discerning remains with us.

Principle 02

The Spirit's Work

AI's outputs reflect the patterns of the world, not the renewal of the Spirit. They cannot bypass the testing required to discern God's will. Only the renewed mind, led by the Spirit, can do that work.

And more than this: AI is not neutral. It exerts a spiritual gravity. Using it shapes us — toward calculation over presence, toward optimization over relationship, toward simulation over reality — even when we intend otherwise. We must be aware that the tool is acting on us while we act on it.

Romans 12:2 "Do not be conformed to this world, but be transformed by the renewal of your mind, that by testing you may discern what is the will of God — what is good and acceptable and perfect."
Principle 03 · Core

Capability is Not Calling

AI opens up productivity and efficiency that previously did not exist. But the friction we used to face — the slowness of writing, the cost of research, the time of building — was itself a form of discernment-by-necessity. As that friction disappears, the discernment it once enforced must now be made explicit. And the discernment must be made explicit twice: once for what we choose to do, and again for what that choice will cost us in capability we eventually lose and dependence we eventually inherit.

The Working Conviction
Because we can do it does not mean we should do it. The deeper move is one of posture: the framework calls us to delight first, decide second.
Psalm 37:4 "Take delight in the Lord, and he will give you the desires of your heart."
Principle 04

Our Posture as Faithful Users

The human is the image-bearer, not the AI. Before assigning any task to AI, we pray for help, guidance, and spiritual fruit. The AI is a tool wielded from a place of prayer — never a substitute for it.

Two postures to watch for: the first is pride dressed as integrity — clinging to a skill we have worked years to develop when AI is now better, refusing the humility of using the better tool. The second is hubris dressed as enablement — celebrating new capability AI grants us without having earned the wisdom to use it well. Both are spiritual failures. We must hold both at once.

The Discipline of Non-Use and Practice

We practice regular fasts from AI — a day each week, a season of the year, certain protected zones of life and ministry. The capacity to set AI aside is what proves we are not yet captive to it. Where there is no fast, dependence is already forming.

Fasting is one half of capability protection. The other half is the discipline of regular practice — keeping our hand in the work AI could do for us. If AI writes our emails, we still draft some ourselves. If AI summarizes books, we still read and synthesize. If AI helps us reflect, we still wrestle alone with Scripture. Not all of the work, not every time — but often enough to keep the capability alive. What is not exercised, dies.

Principle 05

The Limits of Our Discernment

Our ability to evaluate AI output is bounded by our own knowledge, experience, and cultural context. There will be times when AI produces answers we cannot adequately judge — and times when our judgment of AI is itself in error. This is the catch: the very thing that makes discernment necessary (AI exceeding our experience) also makes it harder (our experience being too narrow to evaluate).

This is why discernment cannot be a solo activity. It requires community — voices with different experiences and blind spots than ours, and where the work crosses cultural or contextual lines, voices from those contexts — and ultimately the Holy Spirit.

1 Corinthians 13:12 "For now we see only a reflection as in a mirror; then we shall see face to face. Now I know in part; then I shall know fully, even as I am fully known."
Principle 06

The Burden of Truth

AI generates plausible content, not necessarily true content. Its outputs sound like human writing because they are statistically shaped to do so — but a confident answer is not a true answer, and a coherent sentence is not a faithful one.

We carry the burden of verifying truth before publishing, sharing, or acting on AI output. In ministry contexts this matters doubly: a single hallucinated fact or mistranslated verse can erode trust that took years to build.

2 Kings 22:8 "Hilkiah the high priest said to Shaphan the secretary, 'I have found the Book of the Law in the temple of the Lord.'" — The word of God is sought, not received passively. Verification is itself a form of seeking.

Section 04

The Four-Question Filter

Every question, problem, idea, or job to be done must pass through four questions before any AI work begins.

01
Is this for God's glory?
Does the work point to him, or quietly recenter us — our productivity, our reputation, our convenience?
02
Is this what God is calling us to do?
Has the Spirit drawn us toward this, or did urgency, opportunity, or capability pull us in?
03
Is this loving to our neighbor?
Does it serve the people we're called to serve — and does it protect those who could be harmed or left behind?
04
Are we stewarding this faithfully?
Resources, attention, the ethical sourcing of the tool itself, the people whose work AI was trained on, and whether AI’s benefits in this work accrue to those already in power — or to those without it.

Section 05

The AI-Fit Decision

Once a job has passed the filter, one more question: "Should AI play a central role in this work, or only an adjacent one?" This is not a question of capability ("can AI do this?") but of weight.

Default Posture

AI is broadly appropriate, with one condition.

For most work that has passed the four-question filter, AI can play an assistive role — but only when paired with a strong human-in-the-loop. The faithful user must remain active in the work: verifying, shaping, redirecting, and taking responsibility for what is produced. AI is wielded; the human is never bypassed.

Strong human-in-the-loop is harder than it sounds. It requires active verification of outputs, asking questions multiple ways to surface sycophancy, and the discipline to question what feels easy. Approval alone is not oversight.

The Working Conviction With disciplined human oversight — not mere approval but active verification, challenge, and resistance to AI’s pull toward agreement — AI can serve a wide range of ministry, business, and creative work. The compromise to integrity comes not when AI is used, but when “oversight” becomes a rubber stamp.
Restriction

Categories Requiring Maximum Human Presence

A small set of tasks have human presence, confession, or relational depth as a constitutive element of the work — not as a quality enhancement but as the thing that makes the work what it is. These are the categories where the framework draws the line against AI substituting for what is most human: embodied presence, real relationship, the work that requires a person made in God’s image. In these categories AI is not forbidden, but it is restricted to an adjacent role.

Scripture & Biblical Content

AI can support exegetical study, surface cross-references, draft outlines, and stretch our reading. It must not produce final theological content that reaches an audience without faithful human review.

Crisis Pastoral Conversations

AI can help with preparation, follow-up, and research. It must not be the conversation itself. An AI's appearance of empathy in grief is a form of deception. The value of a real voice on the line is the whole point.

Sacred & Sacramental Moments

Worship, communion, confession, prayer with another, the sermon delivered live — these require human presence as a constitutive element. AI may support preparation; it does not mediate the moment.

Discerning God's Will

AI can support research, generate options, and clarify thinking. It cannot discern God's call. That work belongs to the Spirit, the human, and the community of faith around them.

Loop Level

What Level of Human-in-the-Loop?

For work where AI can play a central role, the question becomes: how present must we be? Three intensities for three kinds of stakes.

High-Touch

Every substantive decision is human.

AI assists with research and drafting only. We make every consequential choice and review every word.

Mid-Touch · Default

Every output is human-reviewed.

AI produces drafts and proposals. Nothing leaves the loop without us reviewing and approving. The default for most work.

Light-Touch

Periodic and final review.

We review periodically and always on outputs that carry public weight. Reserved for low-stakes work that is easy to correct.

Section 06

Rootedness — The Four Anchors

AI tends to pull us away from four anchors that hold a life of faith together: Past, People, Place, and Prayer. Paul Kingsnorth names these in Against the Machine: On the Unmaking of Humanity (2025) as the values of pre-modern life the Machine systematically unmakes. The framework asks not just how AI is used but whether these anchors remain intact through its use.

Anchor 01 — Past

Scripture · Saints · Tradition

Are we rooted in Scripture, the witness of the saints, the wisdom of the Church across generations? Or is our diet of formation now mostly AI-mediated, AI-summarized, AI-curated?

AI can present anything as new and timeless at once, dissolving real tradition into endless present. The past is an anchor only if we read it directly.

Anchor 02 — People

Community · Cross-Cultural Voices

Are we rooted in embodied community — face-to-face conversation, shared meals, common worship, real accountability? Or is more and more of our relational life now mediated by algorithms and conducted with screens?

AI can simulate presence without delivering it. People are an anchor only if we are physically with them.

Anchor 03 — Place

Locality · The Land

Are we rooted in actual geography — a parish, a neighborhood, a city, a piece of land we tend? Or are we increasingly "anywhere," meaning nowhere?

AI flourishes on placelessness; faithful life requires location. Place is an anchor only if we stay long enough to be shaped by it.

Anchor 04 — Prayer

Communion · The Spirit

Are we rooted in active prayer — not optimized, not productive, not measurable? Or has prayer itself become another performance to be improved?

AI cannot pray. Neither can a person who has substituted optimization for it.

Section 07

The Discernment Flow

The workflow that follows once a job has passed the four-question filter. Two lanes run on top of the four anchors of rootedness, which feed discernment into every gate. Step through the flow — or toggle to compare the two lanes.

The blue-ocean question. Don’t only ask “what problem should AI solve?” — also ask “what is now newly possible because AI has shifted the constraints?” Opportunity is a trigger as legitimate as problem, question, or job. Imagination, innovation, and new possibility belong at the start of the flow, not as an afterthought.

LANE A — AI ADJACENT For categories requiring maximum human presence LANE B — AI CENTRAL For all other work, with chosen loop level GOD'S GLORY PROBLEM QUESTION JOB OPPORTUNITY 4-QUESTION FILTER Glory · Calling Neighbor · Steward AI-FIT DECISION Adjacent or Central? PRAYER DISCERNMENT FAITH AI (ADJACENT) Research · Prep Drafts · Follow-up HUMAN-LED Conversation, sermon, presence, discernment PRAYER DISCERNMENT FAITH AI RESEARCH AI PLAN AI BUILD REVIEW Glory · Truth Fruit · Neighbor Process · Trajectory ↺ Review feeds back into the next cycle ROOTEDNESS — THE FOUR ANCHORS PAST Scripture · Saints · Tradition PEOPLE Community · Cross-cultural PLACE Locality · The land PRAYER Communion · Spirit
01 / 16 · Step
God's Glory
Origin

Every act of work begins here. Before a problem is framed, before a question is asked, we orient toward God's glory as the destination — not just the criterion.

Lane A — AI Adjacent

AI assists; the human carries the central work.

Lane B — AI Central

AI produces the outputs, with checks between every step.

The Four Anchors

Past, People, Place, Prayer — discernment fed into every gate.

Review & Loop

Both lanes converge at Review. The review feeds back into the next cycle.

Section 08

The Review Criteria

Review is not just "did it work?" — it's a discernment step that asks six distinct questions.

01 · Glory

God's Glory

Did this work bring glory to God? Or did it center us, our productivity, our convenience?

02 · Truth

Verified Truth

Was what we produced actually true? Did we verify, or did we trust the AI's confidence?

03 · Fruit

Spiritual Fruit

Did this work produce love, joy, peace, patience, kindness — or anxiety, pride, dependency?

04 · Neighbor

Love of Neighbor

Did this serve our neighbor? Did stewardship and inclusion shape the outcome?

05 · Process

God-Honoring Process

Was the way we worked God-honoring? Did we pray? Did we listen? Or did we just ship?

06 · Trajectory

The Pattern Over Time

Beyond this single piece of work — what is the pattern of use ushering into our lives, our team, our ministry? Each individual use can be defensible while the accumulated pattern is forming something we would not choose.

Section 09

When Principles Collide

The framework's principles do not always pull in the same direction. Real decisions involve trade-offs, and pretending otherwise produces brittle frameworks. Expand any tension to read.

01Efficiency vs. Presence
AI can do in seconds what used to require hours of human attention. But some of that "wasted" time was the soil of relationship, formation, and ministry. The framework asks: what is lost when we save the time, and is that loss acceptable here?
02Stewardship vs. Inclusion
Faithful stewardship may mean producing more with fewer hands — and AI can amplify our resources in ways that genuinely feel like good stewardship. But fewer hands means someone is doing less, or nothing at all. The role becoming economically unviable is not the same as the person becoming unviable. The framework asks: who gets walked through the change, and how?
03Speed vs. Discernment
AI is fast; discernment is slow. The temptation is to skip the prayer-discernment-faith gates because the AI is ready and the deadline is tomorrow. The framework asks: is the deadline real, or is the urgency manufactured by what the AI made possible?
04Individual Use vs. Communal Accountability
One person's decision to use AI a certain way affects the team, the ministry, the audience. The framework asks: who else needs to weigh in, and who is the discernment partner for this decision?
05What AI Frees vs. What It Takes
Every use of AI is a four-part bargain: new capability and reduced burden traded for diminished skill and emerging compulsion. The first two parts arrive immediately and are loud; the second two arrive later and are quiet. The framework asks: am I getting more time for what is most human — presence, prayer, relationship, embodied ministry — or just more output at the cost of capability I’ll later need?

Section 10

Individual & Communal Discernment

Most AI use happens inside teams, organizations, and ministry contexts. The principle: match the discernment community to the scope of the work.

In Teams

Who holds whom accountable?

When a team uses AI together, the discernment burden cannot be silently distributed. Someone must explicitly hold the question: "Should we be doing this at all?"

In Engagements

The discernment you owe the client.

What does discernment look like when a client is paying you to use AI for them? Consultants owe their clients more than execution — they owe them the discernment work the client may not know to ask for.

Cross-Cultural Work

When the work crosses a boundary.

When ministry crosses cultural lines, discernment requires voices from the recipient culture. The moment the work crosses a cultural boundary, the discernment community needs to cross with it.

Proverbs 11:14"Where there is no guidance, a people falls, but in an abundance of counselors there is safety."
In Maintaining Capability

What capability lives in the team rather than the tool?

AI tools depend on companies that can change pricing, change terms, change quality, or disappear entirely. A team that has outsourced a function entirely to AI has built fragility into its mission. The faithful organization protects capability deliberately: people are trained to do the work AI does, knowledge is documented where the AI does not own it, decisions and their reasoning are recorded in human-readable form. The test: if your AI tools disappeared for three weeks, what would your team still be able to do? Whatever survives is what your team actually has. Everything else is borrowed.

Section 11

The Framework at a Glance

Use this one-page reference to evaluate any AI decision quickly. Each principle and question reduces to a single test.

Section 01 — The Promise of AI for Ministry
The Promise
"Before asking ‘should I use AI?’ — what could AI unlock for this ministry that wasn’t possible before?"
Section 02 — Three Reality Checks
What AI Is
"Am I treating AI as a knower, or as the device it actually is?"
Failure Modes
"Have I tested this output for sycophancy, hallucination, and the genie problem?"
The Bargain
"What is AI freeing me for, and what is it taking from me?"
Section 03 — Foundational Principles
01
Nature of AI
"Am I asking AI to know, reason, or discern — work that requires a knower?"
02
Spirit's Work
"Am I leaning on the Spirit to discern — and what is this use forming in me?"
03
Capability ≠ Calling
"Just because I can do this with AI — should I?"
04
Our Posture as Faithful Users
"Am I assigning this from prayer — and am I still fasting from AI regularly?"
05
Limits of Our Discernment
"Who outside my own perspective is helping me judge this?"
06
Burden of Truth
"Have I verified what AI produced before I act on it?"
Section 04 — The Four-Question Filter
Q1
Glory
"Will this work bring glory to God?"
Q2
Calling
"Is this what God is calling us to do?"
Q3
Neighbor
"Is this loving to those we serve and those affected by this work?"
Q4
Stewardship
"Are we stewarding this faithfully — resources, attention, tool ethics, those left behind?"
Sections 05, 06 & 08 — AI-Fit, Anchors & Review
Weight Question
"Is this where AI plays an adjacent role only, or can it play a central role? If central, what loop level — high, mid, or light?"
Four Anchors
"Are Past, People, Place, and Prayer still intact through this use?"
Trajectory
"What is the pattern of use, accumulated over time, ushering into our lives, our team, our ministry?"