
What it is
An organisation, held in one record
Most organisations keep themselves in pieces. The engagement survey lives in one tool and the performance reviews in another. The skills sit in a spreadsheet somebody updated in the spring, and the evidence for a B Corp submission is a folder of documents that nobody can quite find in the week it is due. Each piece is accurate about its own corner and silent about the rest. B-Corp is the foundation of the suite because it starts from the other end: one signed-in record of an organisation, with its people and teams, their assessments and reviews, its plans and objectives, the evidence of what it has done, and the page it chooses to show the world.
Around that record works an AI, and the app is plain about where. An agent called the Luminous Conductor waits in a drawer, the Luminous Concierge, that rises over every page; it reads the organisation's own records and answers from what is already working before it says what could come next. A second agent holds a personal assessment as a conversation. Other AI calls sort pasted text into B Corp evidence, break a plan into department actions, write research dossiers on other companies and turn a team's assessments into a reading of the team. Every one of them proposes. A person decides.
Every one of them proposes.
The record is wide, and this page chooses rather than counts. It tells the assessments a person takes and the daily check-in they keep. It tells the B Corp workspace, where evidence waits for approval and objectives earn estimated points toward the B Impact Assessment. It tells the teams, skills, reviews and learning that hold people together, and the leaders' rooms, where strategy frameworks run against the organisation's own data and a research planner names the seats inside another company, never the people sitting in them.
One thing is said once, here, so that it need not be said in every paragraph. The scores B-Corp shows for the five B Corp impact areas are estimates an AI writes from the organisation's own records. They show where the work stands. They are not an assessment by the body that certifies B Corps, and nothing in the app certifies anyone.

A day with it
Tuesday, a head of people and a four-day week
The head of people and impact at a firm of ninety opens B-Corp at half past eight. Behind the page two soft glows rise and fall on a twenty-second cycle, slow enough to breathe with. She presses Control, Shift and C, and the Concierge rises from the foot of the screen. She speaks rather than types, asking where the firm's feedback loops are broken, and watches her own words fill the box. The Conductor begins with what is already working in the leadership team, then sets out what might come first, what next, and what becomes possible after that.
On the B Corp dashboard she switches on Plan Breakdown Mode and pastes the four-day-week proposal from yesterday's leadership meeting, written in the firm's own words for the management team. The agent returns an executive summary, the impact areas the plan touches, wins that serve several goals at once with their estimated points, quick wins under thirty days, actions by department with an owner role and a timeline, and a ninety-day roadmap in phases. She asks what more context would sharpen it, and it asks her a question back. She downloads the breakdown as a text file for the chief executive.
In the Evidence Vault she pastes the notes from last week's review of the parental-leave policy and presses Auto-Tag with AI. The note is filed under Workers, with the requirement it supports, a two-sentence summary of why it counts, five tags and a relevance score. Auto-Collect then sweeps the app's recent reviews and knowledge articles and leaves three more items marked pending. She approves two and leaves the third for Thursday, because she is not yet sure it says what it seems to say.
Before lunch she generates a fresh optimiser plan, titled with today's date, and reads six priority actions by impact area, each given thirty, sixty or ninety days by its effort. In the objectives tracker she chooses that plan and presses Generate OKRs. Draft objectives appear for the quarter, each with two or three key results. She types this month's figure against one of them, eleven of twenty suppliers screened, and watches its progress and the points earned move together. Then she runs the supplier audit as a preview, reads what it would propose, and runs it for real.
In the afternoon a colleague in operations sits down to the Luminous Portrait. The guide speaks first, and its first words say that this is not a test, that there are no right answers and that there is no score. It asks what has been most alive for him lately. Forty minutes later it speaks a portrait back and asks whether it may save it to his profile. Down the hall, the head of people opens Team Assessments, names a spring campaign, chooses the quick assessment, sets a date three weeks out and picks the design team she leads. Six invitations leave, each signed with her name.
At six she does the daily check-in, which asks how she is, really. Mood sits at six, energy at five, presence at seven; she types her hours of sleep from her watch by hand and writes one line about a conversation she is still carrying. Last, she downloads the B Impact Assessment report, a thin gold bar for each of the five areas, and the whole archive bound as one document, whose final page reads Held with appreciative gravity.

The person
Eleven maps, a conversation and a daily check-in
The full assessment is a long sitting, and it says so before it begins: fifteen to sixty minutes, with progress kept so a person can stop and come back, and no right or wrong answers. It walks through eleven sections, one framework at a time, with a progress rail and Back and Continue on every step. A person rates statements tagged to the value systems of Spiral Dynamics, answers meaning-making questions for the orders of mind, and moves sliders for five kinds of intensity drawn from positive disintegration. Sentence stems follow for ego development, then three dilemmas for moral reasoning and eight need sliders for the hierarchy of needs.
The second half turns to action logics and their time horizon, twelve statements on the four integral quadrants, a four-letter type with its cognitive functions chosen in order, twenty-four character strengths in six virtue groups, and a life wheel of eight areas. When the last section closes the screen reads Composing your portrait while an AI writes four to five hundred words of integrated insight that take the person's job title and department into account, and the whole is saved as one assessment.
The Results page then holds every assessment a person has taken, newest first, with dated buttons to move between them. Tabs show a developmental profile, a radar of the four quadrants, the type and its functions, the five signature strengths with a chart of their categories, and the life wheel with an overall satisfaction figure. The developmental profile is shown as the record holds it: an index out of a hundred, a named altitude, a centre of gravity, a growth edge, a few optimal challenges and a bar for readiness.
Not everyone has an hour. The quick assessment asks six sections of written questions instead: what brings joy at work, a moment of pride, what frustrates and how it is met, the one thing a person would change. Every question carries a long worked example beneath it and a quiet mark once an answer is written, and Continue waits until all of them are. An AI then reads the answers and builds the same kind of profile from the person's own words, with three hundred and fifty to four hundred words of insight. The type, strengths and quadrant readings in it are inferences an AI draws from writing, and they are offered as that, not as the results of licensed instruments.
A chief executive has two assessments of their own. The full one asks about leadership style and strategic vision, then weighs six groups of stakeholders against one another on sliders, from shareholders to suppliers. It closes with the attention given to each quadrant and an appreciative inquiry in five movements, and an AI writes a four-to-five-hundred-word analysis that the chief executive's dashboard opens on. The quick executive version asks four sections in writing, on leadership, vision, stakeholders and culture, and is read the same way.
Then there is the assessment held as a conversation. The Luminous Portrait is a split screen, a photograph on one side and a guide on the other, and the guide is an AI agent that asks one question at a time. It works through strengths and gifts first, then values, then what the person most deeply wants, asked at least three levels down. It asks permission before naming anything it notices. At the end it offers three to five strengths, the deepest want and one growth edge, framed as an invitation, and asks before saving any of it. Beneath the box where a person writes or speaks runs one line: you are always the authority, and you can always correct the map.
The person also keeps a daily check-in. Seven sliders run from one to ten, each with its own two ends in words, from heavy to radiant and from scattered to fully here. Beneath them sit a few body figures typed in by hand and a line asking what is alive now. Once a few entries exist, an AI writes two or three sentences on the patterns between the felt sense and the numbers. The Wellness Companion reads the last week of entries and opens with one precise question about where something is being held, with a practice suited to the hour; it can write a four-week plan, and the long-term trends draw thirty and ninety days of the same record. A library of sixteen short practices sets each out in six numbered steps. All of it is reflection and practice, offered as that and never as care.

The AI files the evidence and writes why it counts. It stays pending until a person has read it and pressed Approve.

The impact
From a pasted note to a published page
The B Corp workspace opens on a dashboard of seven tabs. On the first, Sync Now gathers what the organisation already holds (how many people there are, what share have taken an assessment, the latest alignment and life-satisfaction figures, how many wellness entries exist) and asks an AI, briefed as an advisor on the five impact areas, for an estimated score in each. Every area comes back with its trend, its top gap and the action that would move it most, with an overall estimate set against two marks, eighty and one hundred, and a short list of moves that touch several areas at once.
The agent tab is where most plans begin. It opens with six suggested questions, such as what would move the Workers score or how a living-wage policy might serve Workers and Community together, and answers each in continuing conversation. With Plan Breakdown Mode switched on, any initiative pasted into the box comes back read through the five areas, with department actions, quick wins, a phased roadmap and an estimated total. The Requirements tab asks the same AI for a brief on any one area: the key requirements, the points available, quick wins with their effort and the gaps organisations commonly miss. Those briefs are drafted by the AI on request, and the page treats them as a starting point for reading the standard, not as the standard itself.
Everything else is filed as pending review, and it stays there until a person presses Approve.
Evidence is the centre of the work, and it is handled the way a careful auditor would want it handled. Each item is tied to one impact area and one requirement, typed as a policy, meeting notes, a survey or a report, and summarised. A person can paste up to six thousand characters into the harvester; the AI decides which area it supports, writes two or three sentences on why it counts, suggests tags and scores its relevance. Anything under forty is set aside and the panel says so in plain words. Everything else is filed as pending review, and it stays there until a person presses Approve.
Objectives come next, and they come from plans. A person chooses a saved plan and presses Generate OKRs, and an AI turns its department actions into objectives with measurable key results, a points target and the plan action each came from. As current values are typed in, each key result's progress is worked out against its target, the objective's progress is the average of its results, and the points earned rise in proportion. An objective at one hundred is complete, and one under thirty is marked at risk. A preview shows the change before it is saved.
The optimiser reads recent area records, how much evidence each area holds and how its objectives are moving, and writes a plan of five to seven priority actions, a few wins across areas and a ninety-day roadmap, kept with a dated history. The supplier audit reads the organisation's contracts and proposes, conservatively, the clauses that bear on Community or Governance, and it never proposes the same contract twice. The at-risk scan compares each active objective with how much of the quarter has passed, finds a rescue action from a saved plan and leaves a note for the owner. When the organisation's Slack workspace is connected it also sends the owner a direct message, and a dry run shows who would hear before anyone does.
When the work needs to leave the building it leaves whole. One action composes a B Impact Assessment report as a PDF, with an executive summary, each area's estimate against its target as a thin gold bar, an evidence catalogue, objectives and active plans. Another binds every B Corp record into one document set in a serif with the house seal, one record to a page. A third downloads the same archive as a single file with written instructions for rebuilding it.
The last thing the workspace makes is the one the world sees. In the disclosure editor an organisation writes a headline, a subline and a mission, and adds its approved metrics one at a time, each with a value, a unit and its impact area. It saves as a draft until Publish is pressed. The page that results lives at its own address, carries the organisation's logo and colour, its stories and short summaries of the evidence it has cleared, and counts its views; the editor's own line says that only the fields below appear and that internal evidence stays private.

The people together
Teams, skills, reviews and learning
A team in B-Corp can be named by hand or built with help. In the team builder a person writes the team's purpose, lists the skills it needs and picks a size from three to eight. An AI, briefed as an organisational psychologist, reads the talent pool and recommends members, each with a role in the team and the reason for it, along with notes on personality balance, skill coverage and the challenges to watch. Create This Team saves it in planning, ready to begin.
Assessing a team is where the app sends real mail. A manager names a campaign, writes a line on its purpose, chooses the quick or the comprehensive assessment, sets a due date and picks a team they lead or ticks their direct reports one by one. Each person receives an email signed with the manager's name. The campaign card follows its completions until it reads Ready for Insights. At that point an AI writes six to eight hundred words for the manager on the team's composition, blind spots, collective strengths and dynamics, beside a pie of its types and a chart of its most common strengths.
The team performance dashboard measures a chosen team from the record: assessment completion, skill coverage against what its members' roles require, critical gaps, recent engagement and average satisfaction. An AI coach can add an overall reading and a plan of six to eight actions. The collaboration hub writes an opening prompt and a timed agenda for a facilitated discussion, or a set of team-building activities fitted to the team's mix of temperaments, with notes for the quieter members and for the louder ones. A developmental map gathers the team's assessments into one view, a card for each person and a radar of the whole, with an AI reading of its collective growth edges.
Skills are kept in a library of their own, each skill in a category and marked when it is a core competency, and each role says which skills it needs, at what level, as must have, should have or nice to have. The skill matrix compares every person's held skills with their role's needs and lists every gap in a table that can be searched and sorted, with the critical must-have gaps set apart. An AI skill-gap audit, run for the whole organisation, a department or a team, adds a readiness score, critical gaps with their business impact, predicted shortages and the upskilling and reskilling moves that would close them, and is saved with a suggested date for the next one. A person's own skills page shows their readiness for their role as a percentage, and an AI writes a development path to a target role, with milestones that can be ticked off as they are reached.
Performance reviews are built for more than one voice. A manager opens a review for a quarter, mid-year or year, writes the goals to be assessed with a weight for each, and taps the names of the peers who should be asked. Five tabs hold goals, skills, competencies, peer feedback and a summary, and the overall rating out of five is worked out as the ratings are given. A peer answers in place, with two sliders and three short fields. When the last peer has answered, the review moves to the person it describes, for their acknowledgement.
Learning sits on its own quieter campus, a side rail of eleven rooms reached with the same sign-in. An AI writes a learning path of four to six modules toward a career goal, the first module open and the rest waiting their turn. A coaching simulator rehearses eight hard conversations, from a salary negotiation to giving difficult feedback, with an AI counterpart shaped to the person and a scored reading of their clarity, confidence, empathy, persuasiveness and professionalism at the end; each session is saved with its transcript. A twelve-week coaching course is written from a person's goal, available time and preferred way of learning, and career paths lay out three possible routes, each with a readiness figure and a transition plan.

The leaders' view
Strategy, the wider field and other companies
The Conductor is the same host wherever it is met, in the drawer over any page or in a full room of its own. It can read the organisation's assessments and people, their skills and reviews, and the metrics the leadership panels hold, and it can write coaching sessions, reviews and insight summaries back into the record. It keeps memory across conversations. Many screens open it with a question already written, so that a growth edge or a weak measure arrives as the first line of a conversation rather than a note to remember. A button beside its answers, Luminize, rewrites any text in the register of the person reading it, keeping every fact and deadline.
Executive Analytics is the organisation-wide room, and it is marked as protected. Each time it is opened, the view is written to an access log with who looked and why. People who have declined executive viewing are left out, and the page says how many are included. It works out the real skill gaps per department and role, and its predictions are written by an AI from aggregated figures only, then saved, with the most serious raised to a panel of their own. The predictive insights room does the same for a team, sending no individual's details to the AI, and names the people who already hold a missing skill and could mentor in it. The win-win optimiser sets each person's fulfilment beside their effectiveness, lets a leader slide the balance between the two, and has an AI propose reassignments that serve both, with succession readings for the roles that matter most.
On the strategic dashboard, Brief Me gives a director what the Conductor sees, in a fixed order: what is alive and working, the signal that most needs attention, the move available now that serves more than one party, the one missing piece of data that would change the picture, and a thirty-second brief that could be read aloud to a board. A message clarifier sits close by. A leader pastes an announcement and the AI says what it means in plain language, and what it means for each of sales, operations, engineering, finance and people.
Decision Intelligence holds twenty strategy frameworks in five groups, from TOWS and VRIO to scenario planning and value stream mapping. A person runs one, or ticks several and runs them together, and an AI reads each against the organisation's latest assessment, its people, its finances and its sales. A combined run says where the frameworks agree and where they pull against one another, then gives five integrated insights and three actions with the frameworks behind each. Results last for the session and are kept by download: an A4 report whose first page is one recommendation of under twenty words, written for a chief executive who is not a strategist.
The ecosystem room looks outward. Organisations post shared challenges; solutions are submitted, read by an AI for relevance, feasibility and innovation, voted on, and one is chosen. A synergy report reads existing partnerships and proposes kinds of partner and opportunities that name what each side and the wider field would gain. Outreach is drafted and kept as a draft, moved through its stages by hand. Onboarding a new partner, once its welcome has been read, sends that welcome by email. Connection health gives each cooling relationship two or three conversation starters drawn from what the two have shared.
The research planner turns to other companies. A person names one, adds its web address and anything already known, and chooses a lens: B Corp impact, ESG and disclosure, or general strategy. An AI with web search writes a dossier, with a maturity score for each dimension, its central gap and the public evidence behind it, priority actions and a ninety-day roadmap. It also builds a contact matrix that names the seat to approach in each area, what to raise with it and an opening line. Names, emails and telephone numbers are left blank, to be filled in by hand, and every dossier is saved to reopen.
Reports have more than one form. The report generator prints any assessment with the organisation's own logo and colour through the browser's own print, and Executive Analytics can have an AI write a report from department figures, to download as text or send by email. The report builder lets a person define a report and keep it as a template, with an AI drafting its first summary from the outline it is given.

The conversation opens by saying it is not a test and has no score. The portrait is saved only after the person says yes.

The suite
What it reads, what it reaches, and its neighbours
B-Corp sits on one desk with the eighteen other applications of the suite, installed on the house's own Mac by its own method and against its own cloud record. The desk starts, opens and rests each application, and every two minutes it starts again any that has stopped. A stock-take of every application's records, functions and agents is kept beside it. Each application keeps its own sign-in, and B-Corp's asks for one on every screen.
Several of its panels are built to reach the systems an organisation already runs. The chief executive's dashboard has six tabs named for the body of a company. The Pulse reads a finance system for cash, burn and actuals against budget. The Oxygen reads a CRM such as Salesforce or HubSpot for the pipeline. The Engine reads an HR system for people and tenure, The Work reads Jira, Asana or Monday.com for projects, The Armor reads a legal system for contracts and renewals, and The Shield reads an IT system for uptime and incidents. Each connects through an address and a key placed in the app's settings. Until they are placed, each panel shows a demonstration set and says that it is in demonstration mode, and nothing in it should be read as the organisation's own.
Slack is reached the same way: once an organisation's workspace is connected, the at-risk scan writes to an objective's owner directly. The wellness record is built to hold readings from a watch or a ring, and its regulation check answers them with a short practice and the reason for it; today those readings arrive typed by hand, and no wearable sign-in exists yet.
The neighbours are arranged to be used together, and the arrangement is by design rather than by wire. A consultant working in Resonance Canopy has, in B-Corp, the organisational assessment, the analysis and the research dossiers an engagement would start from. A coach keeping Resonance Coaching open beside it works from the same portrait, growth edge and development path. The research planner's contact matrix, a seat, what to raise and an opening line, is exactly the shape of the sheet a person keeps in Resonance CRM, carried across by hand. The frameworks the assessments draw on are the ones Resonance Canon teaches as courses read with others. The daily check-in's journal line and gratitude notes are the kind of reflection Resonance Sanctuary keeps privately, in a room of its own.
Two neighbours are closer still. The Conductor is configured with the house's brand-voice skill, the voice The Luminous Book writes down for an organisation. B-Corp also carries an agent that guards the Resonance OS design language, the language the Resonance OS essay sets out: it audits a screen or a line of copy against a seven-point test, rewrites copy in that voice and names any departure by its principle. And the metrics and stories an organisation approves for its public page are the material a marketing workspace such as SquareDash would carry further, once a person chooses to carry them.

Who it is for
A purpose-led organisation, and the people inside it
B-Corp is for an organisation that is working toward B Corp certification, or keeping it, and that wants its people's development and its impact held in the same place. Inside it the app is arranged by role. A person sees their own assessments, skills, reviews and check-ins. A manager sees their teams and their direct reports. An executive and a chief executive see the dashboards that read across the whole. Whoever leads the impact work has the B Corp workspace; whoever runs the organisation's settings invites new people by email, with a code that lasts seven days, and sets the logo and colour the reports carry.
What it replaces is the folder, the survey tool, the separate reviews system and the spreadsheet of skills, each kept by a different person and each silent about the others. It also replaces a particular kind of week, the one before a submission, when evidence that was created months ago in a meeting or a policy review has to be found again, described and tied to a requirement. In B-Corp that description is written on the day the note is pasted, by an AI, and read by a person before it counts.
What makes it different is the order in which it thinks. The Conductor is written to begin from what is already working, to read the organisation as a living whole, and to sequence its advice developmentally, saying what comes first and what becomes possible after. The assessments are framed as maps a person may correct. The executive briefing opens on what is alive before it names a risk, and the leadership panels close their plain-language readings by asking to be corrected. The research planner names roles and refuses to invent people. Each of these is a choice in the code, made the same way every time, and the app is the sum of those choices more than of its screens.
It is also honest about its own edges. Scores for the impact areas are estimates. Forecasts, readiness figures and synergy scores are an AI's reading, and the page says so. Panels without a connection say that they show demonstration data. A thing that is told as it is can be relied on for longer than a thing that is told as more.
How it is offered
Installed for an organisation, and kept
B-Corp is offered to an organisation in three ways: on its own, as part of the foundation, or inside the whole of B-Corps, where every application on these pages is installed together, set up in the organisation's own voice and with its own records, and kept running for as long as the organisation keeps it. The house carries each application into its suite for good, so what is set up now is still there, and still mended, in the years after.
The fee reflects the work of installing and shaping it, and the size of the organisation it serves. It is set with each organisation, after a conversation about what the application will carry for it.
Arranged in conversationB-Corp · B-Corps

The care it takes
What each person holds, and one quiet page
Privacy in B-Corp is a set of switches each person holds over their own assessment. Whether a direct manager may see it in full is off until the person turns it on. Whether executives may see it anonymised, whether it joins team totals, and whether HR may see it are each the person's to change, and a table beside the switches says in plain words who can see what. When a person is shown to another without permission to be named, they appear as an employee number with a masked email. Executive Analytics and the team performance dashboard write every view to an access log, and an administrators' page reads that log back as a record of who looked at what, and why.
The AI is held to the same care. Evidence it finds never counts on its own; it waits for a person. The impact estimates are asked to be realistic and uninflated, and the supplier audit is asked to be conservative about what it proposes. The predictive rooms send only aggregated figures and are told never to single anyone out. The research planner leaves every name blank. The Portrait asks before it names, and asks again before it saves. The public page answers only once it is published, and only with the fields the organisation approved.
Held with appreciative gravity, which is the posture of the whole app toward the people it records.
Everything the organisation keeps about its B Corp work can leave with it, as one file with instructions to rebuild it, or as a bound document, one record to a page. Nothing has to be rebuilt by hand if the organisation moves on.
The rooms are made to be lived in. While a person is recognised the screen shows one slow gold ring and the word Arriving. The glows behind them breathe. Empty states say what to do next in a sentence: start journaling to see the map, run an analysis first, take a first assessment. Both portals carry a breathing circle that swells for four, holds for four and settles over six.
And at the end of the B Corp archive, after every policy and survey and objective has been set on its own hairline-ruled page, there is a last page with two lines. One says End of Archive. The other says Held with appreciative gravity, which is the posture of the whole app toward the people it records.
B-Corps
