Three objectives Amber, none Red. One cross-functional signal worth your attention first.
Your agents right now
Overnight leadership brief assembleddone
Monday meeting written up to decision logdone
Board pack narrative · this quarterdrafting
Reply to Caldera on SOC 2 timingneeds your OK
01 · The before
You are the memory. Every single time.
Context scattered across inboxes and drives. Days lost to work that does not need you. Always one step behind your own business.
It is not that AI cannot help. It is that nothing you own actually knows your world, so you brief it from scratch every time and correct half of what comes back.
02 · The reference company
This is not a feature tour. It is a whole company.
Lumen is a mid-market software company. Every leader, every function, running on a Business AI Operating System we built with the same programme, engine and system we install for our clients.
The company and the people are invented. Everything else, the knowledge base, the skills, the dashboards, the connectors, is the real thing, running. We built it with synthetic data so we can open every door without exposing a real client. Nothing here is cherry-picked.
The Lumen leadership team · a fictional reference company
Synthetic company · real system · the actual programme output
03 · Find your person
Whose job is closest to yours?
Pick the leader who does what you do. See their working week before and after, and the system that changed it. Start with Erin, the founder, for the whole-company view.
Erin Whitfield
Founder & Managing Director · Lumen
Erin runs a 140-person software company with eight leaders reporting to her. She owns the whole scorecard, from ARR growth to the board relationship, and she is the person who has to hold the entire picture in her head.
Before · you do the work
Half a picture, assembled by hand
Monday morning, seven inboxes deep, building a view of the company to carry into the leadership meeting. She walks in knowing about half of what is happening across functions, the rest pieced together at the table.
The cross-functional connections are real, but she has to make them herself, after the fact, when she happens to notice.
The board pack is two weeks of evenings at quarter end, a picture she already carries, extracted from eight sources under pressure.
A general AI assistant helped with the writing but knew nothing about Lumen, so every task started cold.
After · it works for you
The whole company, read in five minutes
She opens one thing, the Founder Strategic view, and reads the brief the system produced overnight. OKR exceptions, the cross-functional picture, and the two or three calls that need her this week.
This week it flags one signal early: the Manchester SDR pipeline and Professional Services capacity straining on the same accounts, spotted across two functions at once.
The Monday meeting writes itself up to the decision log, drafted for her approval. She fixes two things and approves the rest.
The board pack is shaped across the quarter, a first draft ready three weeks out. She still makes every call. The system removes the assembly, not the judgement.
Her dashboard, liveLumen · Founder Strategic view · illustrative data
lumen-aios · founder-strategic.html
Illustrative result
“Every Monday, she reads the whole company in five minutes. Then she decides.”
~2 hrs
a week back from assembly work
2 hrs
to finish the board pack, not two weeks of evenings
<24 hrs
decision log current after every meeting
1 signal
cross-functional risk caught early each quarter
Illustrative figures for a synthetic company. Real before-and-after numbers come from the programme measurement spine, captured from actual participants.
What Erin's system is made of
Knowledge base
Strategy, OKRs, the decision log and every function's context, in one place the AI reads.
The Founder Strategic view: exceptions and the calls that need her, not a wall of numbers.
Tasks
The weekday brief, the Monday write-up, the board-pack assembly, all running on schedule.
Connections
Each function's data, the calendar, email, and the meeting recordings, drawn together.
Marcus Bell
Sales Director · Lumen
Marcus Bell runs sales at Lumen. He owns the revenue engine, a £6.5M new-ARR target, four teams of about 28 people and every key account.
Before · you do the work
The number lives in his head
He starts each morning trying to answer one question he cannot yet answer, are we going to hit the quarter.
The pipeline is never quite current, so a deal goes quiet for eleven days and nobody flags it.
The forecast gets rebuilt by hand from a pipeline he does not fully trust, part arithmetic, part gut.
He was in three of twenty meetings, so coaching is guesswork and he takes the rep's sunny word for the rest.
After · it works for you
One screen, a current picture
He opens the Command Centre, reads the brief written overnight, and starts the day knowing where the quarter stands.
The stalled Verda deal is surfaced at day eleven as the biggest risk, with a follow-up already drafted for his review.
The forecast rolls up into Commit, Best Case and Pipeline, informed by what was said in the room, not by optimism.
Every meeting writes itself up against the playbook and MEDDPICC, so he can coach from evidence he can see.
Marcus's dashboard, liveLumen · Sales Command Centre · illustrative data
lumen-aios · pipeline.html
Illustrative result
“You carry the number in your head, and your head is tired by Thursday.”
~5 hrs
a week back for coaching and customers
1-2 hrs
weekly assembly, was 6 to 8
day 11
stalled deal caught, not quarter end
£6.5M
annual new-ARR target he owns
Illustrative figures for a synthetic company. Real before-and-after numbers come from the programme measurement spine, captured from actual participants.
What Marcus's system is made of
Knowledge base
The sales playbook, MEDDPICC, competitive battlecards, deal-desk rules, product information and Marcus's own voice samples.
Skills
Keep the pipeline true, research accounts, draft quotes, roll up the forecast, write up meetings, coach reps.
Dashboard
Pipeline by stage, forecast in three categories, key-account progress, and at-risk or stalled deals as the exception view.
Tasks
A weekday morning brief, an email pass drafting into review, a weekly forecast roll-up, knowledge-base lint.
Connections
The CRM, the knowledge base, recorded meeting transcripts and the pending-review folder where every outbound draft waits.
Nadia Hassan
CTO · Lumen
Nadia Hassan is CTO at Lumen, owning an engineering organisation of about 44 people across delivery squads, platform reliability, security and data, plus six product-embedded vendor relationships.
Before · you do the work
The connective tissue is her memory
Her morning starts with overnight Slack threads and ninety minutes assembling a picture she does not quite have yet.
A vendor incident has to be reconstructed from a Slack thread, and a leadership question takes forty minutes of contract archaeology.
Delivery status is triangulated from squad-lead updates against a release plan last touched four days ago.
Six vendor relationships live in her head, and she cannot name the payments renewal date without checking.
After · it works for you
Fifteen minutes of reading
She opens the engineering brief built overnight and knows what she needs in two minutes, not ninety.
The overnight incident is summarised with cause, resolution and SLA impact, and vendor-health answers Erin's question before she asks.
A delivery view reads the filed squad updates, so she sees what shipped, what is in progress and what is blocked without every standup.
The vendor register surfaces the payments renewal three weeks early, with contract date, terms and the squad's flagged API dependency.
Nadia's dashboard, liveLumen · Engineering view · illustrative data
lumen-aios · engineering.html
Illustrative result
“She knew the platform uptime. The system knew it the moment the incident closed.”
~90 min
a day back for architecture and hiring
30-45 min
daily assembly, was 2 to 3 hrs
99.94%
uptime against a 99.9% target
6 vendors
dependencies tracked with renewals
Illustrative figures for a synthetic company. Real before-and-after numbers come from the programme measurement spine, captured from actual participants.
What Nadia's system is made of
Knowledge base
The platform architecture overview, squad charter and release plan, vendor-dependency register, incident log and security controls summary.
Skills
Assemble the engineering brief, surface delivery across squads, maintain vendor health, prepare the board section, structure post-mortems.
Dashboard
Delivery flow by squad, release status, uptime and incidents, and vendor-dependency health as the distinctive exception panel.
Tasks
A weekday morning brief, a weekly vendor-health check, knowledge-base lint, and post-mortem prompts when an incident closes unfiled.
Connections
The knowledge base, monitoring and uptime data, vendor SLA and contract records, the incident log, and recorded vendor and post-mortem calls.
Olu Adeyemi
Head of Product · Lumen
Olu Adeyemi is Head of Product at Lumen, owning the roadmap, prioritisation and the definition of what gets built. He runs about nine people and is the thread through the Assets 2.0 launch.
Before · you do the work
Synthesis done from memory at night
He loves the discovery calls, but everything that follows, the write-ups and the synthesis, piles up unfinished.
He cannot say whether five customers or fifteen raised a theme, because the notes sit in three formats in three places.
A spec decision waits forty minutes while he hunts for the customer quote that settles it, then argues from memory instead.
The monthly themes summary takes a full day of his best thinking, mostly spent collating rather than reasoning.
After · it works for you
A running synthesis, always current
Each recorded call writes itself up against the roadmap and the themes to date, so the picture assembles itself.
When Erin asks how many customers raised the SLA reporting gap, he answers exactly, seven in three weeks, five of them commercial FM.
Spec reviews run from cited customer evidence rather than reconstructed reasoning, so decisions are better and faster.
The Assets 2.0 launch checklist updates as actions complete, so he reviews the exceptions instead of chasing the whole list.
Olu's dashboard, liveLumen · Product view · illustrative data
lumen-aios · product.html
Illustrative result
“Forty conversations in the synthesis, not the fifteen he managed to write up himself.”
1 hr
monthly synthesis, was a full day
All
customer calls in view, was 40%
7 calls
SLA theme surfaced with evidence
2-3 hrs
weekly launch overhead removed
Illustrative figures for a synthetic company. Real before-and-after numbers come from the programme measurement spine, captured from actual participants.
What Olu's system is made of
Knowledge base
The roadmap context, Assets 2.0 brief, discovery process and templates, the beta programme spec and the running discovery synthesis.
Skills
Maintain the roadmap and launch tracker, synthesise discovery calls, draft specs and beta summaries, surface launch exceptions.
Dashboard
Roadmap status, the Assets 2.0 launch checklist, discovery themes with supporting call counts, and beta cohort status.
Tasks
A weekday morning brief, a weekly product-review brief assembling themes, knowledge-base lint, discovery write-ups run on the call.
Connections
The product knowledge base, roadmap register, recorded discovery-call transcripts, the decision log and the pending-review folder.
Priya Nandra
Marketing Director · Lumen
Priya Nandra is Marketing Director at Lumen. She owns the whole marketing function, brand, demand generation, content, product marketing and events, and runs a team of eleven to twelve.
Before · you do the work
Four tools, none of them agree
The campaign picture lives in scattered systems that never reconcile, so most of the week goes on assembly rather than judgement.
Checking campaign status means opening the email platform, the tracker, a spreadsheet and the project tool, and finding the discrepancy just before the ten o'clock.
Two ad variants perform very differently but nothing ties them back to pipeline, so the renewal decision gets made on half the picture.
An off-brand AI tool produced fluent text that sounded nothing like Lumen, so every draft needed a full rewrite and the team dropped it within a month.
After · it works for you
One view, drafts that sound like Lumen
The system removes the assembly and drafts in the real brand voice, so Priya reviews and decides instead of rebuilding.
The morning brief shows what has gone, what is queued and what is at risk, with demand-gen numbers ready and the AI-referral share broken out.
Ad and LinkedIn variants come back resolved with the winner shown on click-through and downstream pipeline, and a fresh variant already drafted.
A GEO panel shows whether Lumen is cited when buyers ask ChatGPT or Perplexity about field-service software, and recommends citable content.
Priya's dashboard, liveLumen · Campaigns view · illustrative data
lumen-aios · campaigns.html
Illustrative result
“It drafts in your brand voice because it read your actual voice corpus, not a generic marketing template.”
3-4 hrs
a week back from assembly
1-2 hrs
weekly assembly, was 4 to 6
1 day
A/B test resolved, not left to renewal
4 to 1
launch tools cut to one view
Illustrative figures for a synthetic company. Real before-and-after numbers come from the programme measurement spine, captured from actual participants.
What Priya's system is made of
Knowledge base
Brand and positioning, messaging framework, campaign calendar, content strategy, demand-gen channels, launch playbook, SEO and GEO trackers.
Skills
Unified campaign calendar, brand-voice drafting, A/B tested creative, GEO monitoring, demand-gen reporting, content repurposing and competitor watch.
Dashboard
Campaign calendar, demand-gen funnel with AI-referral line, content pipeline, launch tracker, website, SEO and GEO visibility, A/B performance.
Tasks
Weekday morning brief, email pass drafting to pending review, weekly campaign-performance summary, and a periodic website, SEO and GEO scan.
Connections
Email platform, ad and LinkedIn channels, website and SEO data, AI search engines, the Sales knowledge base and meeting transcripts.
Tom Achebe
VP Customer · Lumen
Tom Achebe is VP Customer at Lumen, accountable for retention, the company's priority metric. He runs three teams, Customer Success, Support and Professional Services, about twenty-eight people in total.
Before · you do the work
Account health lives in people's heads
Tom reconstructs the customer picture from Slack, spreadsheets and a stale CRM, and learns about at-risk accounts too late to act.
His first act each Monday is to Slack Hannah and ask how the week looks, because he has no way to see account risk without asking.
A customer drifts as usage drops, a ticket sits and a champion leaves, but nobody joins the signals up until it is mentioned weeks later.
An AI note-taker captured meetings but its summaries were generic and disconnected from account context, so CSMs wrote up their own anyway.
After · it works for you
The picture surfaced, not assembled
The system joins the signals and surfaces the account picture, so Tom forms his own view and catches risk with runway to act.
The Monday brief flags the three accounts that warrant attention, cross-referencing CSM concern against support patterns to rate churn risk.
Onboarding slippage shows up when it happens, so Halcyon is five days behind at day four rather than discovered at day thirty-four.
Meeting write-ups generate in context against account history and renewal timeline, so CSMs stay present and Tom reads what was said.
Tom's dashboard, liveLumen · Customer Health view · illustrative data
lumen-aios · customer-health.html
Illustrative result
“The signals were always there. The system joins them up six weeks before renewal, while there is still runway to act.”
1-2 hrs
monthly report, was most of a Friday
6 weeks
early warning before renewal
Day 4
onboarding slip caught, not day 34
1 saved
account renewed and expanded
Illustrative figures for a synthetic company. Real before-and-after numbers come from the programme measurement spine, captured from actual participants.
What Tom's system is made of
Knowledge base
CS playbook, support SLA model, implementation process, account health methodology, the account register, renewal calendar and implementation tracker.
Skills
Surface at-risk accounts, flag onboarding slippage, prepare renewal briefs, track support patterns, assemble retention reporting, coach from real calls.
Dashboard
Onboarding and implementation status, support queue and SLA summary, health scores and at-risk accounts, renewal pipeline, a coaching view.
Tasks
Weekday morning brief refreshing the health view, an email pass drafting to pending review, a weekly retention summary and dashboard refresh.
Connections
Support ticket queue, the CRM, the implementation tracker, renewal calendar, meeting transcripts, and Sales for shared account health.
Geoff Holloway
Finance & Operations Director · Lumen
Geoff Holloway is Finance and Operations Director at Lumen, owning the numbers, the metric definitions, board reporting and a vendor base of thirty to forty suppliers. He runs a team of about nine.
Before · you do the work
The numbers live in three places
Metrics disagree across systems, renewals slip through unnoticed, and board prep is a fortnight of assembly compressed into a scramble.
ARR is right in two systems and wrong in a third, and none agree on NRR because one runs calendar-month and another trailing twelve months.
A vendor auto-renewed on legacy terms rather than the renegotiated rate because the notice went to an engineer who had since left.
Board prep pulls from four or five sources, so Geoff writes the narrative twice, once when he thinks the numbers are final and once when they change.
After · it works for you
Prepared, not improvising
The system assembles the pack and the renewal picture from agreed definitions, so Geoff shapes the narrative and walks in prepared.
The morning brief gives one gross margin figure from the agreed definitions, cross-checked against the close, the same whichever system you open.
The renewal calendar no longer leaks, flagging anything in the next thirty days with usage and benchmark data pulled before the window closes.
Vendor and board commitments are captured at the time from the write-up, so Geoff walks into the next session knowing what he promised.
Geoff's dashboard, liveLumen · Finance and Operations view · illustrative data
lumen-aios · finance-ops.html
Illustrative result
“The system prepares the pack. Geoff shapes the narrative. Nothing renews on legacy terms because nothing slips through unnoticed.”
4-5 hrs
a week back from assembly
83%
gross margin, one agreed figure
30 days
renewal window flagged ahead
22 months
cash runway, live from the model
Illustrative figures for a synthetic company. Real before-and-after numbers come from the programme measurement spine, captured from actual participants.
What Geoff's system is made of
Knowledge base
Metric definitions for ARR, MRR, NRR, margin, CAC and payback, board reporting format, vendor register model and procurement process.
Skills
Track metrics from agreed definitions, flag cashflow deviations, surface vendor renewals, draft invoice chases, assemble the board pack, track commitments.
Dashboard
ARR and MRR movements, NRR and margin trend, cashflow bridge and runway, board metrics in OKR format, vendor renewal register as an exception view.
Tasks
Weekday morning brief with metric position and renewal calendar, email pass drafting chases to pending review, weekly metric refresh and update.
Connections
Accounting system, the FP&A cashflow model, the vendor register, board pack format, meeting transcripts, and shared metric definitions.
Sadie Klein
Head of People · Lumen
Sadie Klein is Head of People at Lumen, on the executive team, running a team of four to five and owning hiring, onboarding, appraisals, 1:1s, training and culture across a growing headcount.
Before · you do the work
Running the people function by memory
The information she needs is scattered across a spreadsheet, an HR system and email, so she assembles it by hand.
The hiring pipeline is a forty-row spreadsheet updated only when people remember, so she chases hiring managers for feedback twice a week.
A new starter joins Monday and she cannot confirm the IT equipment was requested, then rebuilds the onboarding schedule from a stale template.
Overdue appraisals are carried in her head, so she cannot tell Erin honestly how the cycle is going without pulling and formatting HR data.
After · it works for you
Assembly and chasing gone, judgement kept
The system surfaces stalls, drafts the chasing and prepares the papers, while Sadie still makes every call about a person.
The overnight brief flags roles with no action logged in a week and drafts a chase to each hiring manager, waiting in her review folder.
The training report is already built, eight people incomplete with names and managers and a draft chase each, cleared in eight minutes not forty.
Interviews are recorded with prior consent and returned as a factual debrief with no score, rating or recommendation, for the human panel to decide.
Sadie's dashboard, liveLumen · People view · illustrative data
lumen-aios · people.html
Illustrative result
“The system did not solve the people problems. It gave her the time to solve them herself.”
2-3 hrs
admin a week, now ~30 min review
Same-day
hiring-manager chase cycle
8 mins
training chase, was 40 minutes
Zero
scores or rankings on any person
Illustrative figures for a synthetic company. Real before-and-after numbers come from the programme measurement spine, captured from actual participants.
What Sadie's system is made of
Knowledge base
The hiring process and role briefs, onboarding steps, the appraisal and 1:1 cycle, training register and interview competency frameworks.
Skills
Keeps the pipeline current, surfaces onboarding items early, tracks appraisal compliance, drafts chases, runs consent-first interview and 1:1 write-ups.
Dashboard
Hiring pipeline by stage with days-at-stage, onboarding amber and red items, appraisals due and overdue, training compliance by function.
Tasks
Weekday morning brief refreshing pipeline, onboarding and appraisals, a daily chasing-message pass to review, weekly compliance refresh, lint.
Connections
The HR system, the hiring pipeline, calendar, inbox and, with consent, recorded interviews and 1:1s in a bounded people area.
Fiona Carmichael
General Counsel · Lumen
Fiona Carmichael is General Counsel at Lumen, running a four-person function across Commercial Legal and Compliance, with a direct line to the board and ownership of the SOC 2 audit that unblocks the upmarket move.
Before · you do the work
Records that never quite talk
Legal and compliance records are spread across a document store, spreadsheets and personal notes that never connect, so everything is assembled on demand.
The contract register is a spreadsheet she trusts only roughly, current enough for the five things she carries in her head and no further.
A new cloud vendor has been live three weeks with its risk assessment pending and its DPA unsigned, the gap an enterprise buyer's legal team finds.
A master tracker built six months earlier worked for five weeks, then went stale because maintaining it was itself a job that lost to the actual work.
After · it works for you
Registers current, judgement still hers
The dashboard builds the picture from the register, surfaces the at-risk items and drafts the outbound, while Fiona makes every legal judgement.
The sub-processor gap surfaces before the prospect call rather than because of it, with drafts to Aileen and Geoff already waiting in her review folder.
Each SOC 2 evidence request maps to a control and its documents, so Aileen reviews in a day what used to take three to find.
Negotiation meetings are written up against the contract and risk register, with commitments flagged for her sign-off and nothing recorded without it.
Fiona's dashboard, liveLumen · Legal and Compliance view · illustrative data
lumen-aios · legal-compliance.html
Illustrative result
“Not a policy document. The actual controls, demonstrated.”
4-5 hrs
a week back, now ~1 hr review
~50%
less time finding SOC 2 evidence
3-4 days
new-vendor DPA gap, was weeks
Within hrs
counterparty follow-up, was next day
Illustrative figures for a synthetic company. Real before-and-after numbers come from the programme measurement spine, captured from actual participants.
What Fiona's system is made of
Knowledge base
The contract templates, the standard positions on key clauses, the sub-processor list, risk register, certification status and processing-basis register.
Skills
Keeps the contract register current, flags new vendors needing a DPA, drafts language and deviations, prepares renewal briefings and audit reports.
Dashboard
Contract pipeline with at-risk and overdue items, renewals in eight weeks, sub-processor status, risk register by severity, certification progress.
Tasks
Weekday morning brief on at-risk contracts and certification, an email pass drafting nudges to review, a weekly register refresh, knowledge-base lint.
Connections
The document store, contract and risk registers, sub-processor list, inbox and calendar, plus consent-recorded negotiation and audit meetings.
Owen Pryce
Head of IT · Lumen
Owen Pryce is Head of IT at Lumen, running a three-person team that owns the service desk for roughly 140 staff, identity and access, devices, a 60-plus SaaS estate, and the safe rollout of new tools including AI.
Before · you do the work
Accountable for AI he cannot see
He is answerable for a tool estate and an access picture he cannot show cleanly on demand, while the ticket queue always wins the day.
The queue opens with thirty-one items, mostly the same four or five problems, and he works them himself because his two juniors are stretched.
A Manchester SDR starts using a free AI note-taker he has never heard of, the third shadow-AI entry that month, with no idea what it does with the audio.
A leaver left nine days ago and he is only fairly sure her CRM access was revoked, which is not the answer he wants to give an auditor.
After · it works for you
Governance he can see and test
The system sorts the queue, prepares the access requests and runs a real sanctioned-versus-shadow register, while Owen makes every security decision.
The overnight brief has drafted ticket responses, resolved routine ones via self-service, and flagged the few that need him with the reason.
The shadow-AI note-taker surfaces in the usage register from a new domain in expense and traffic patterns, with a warm draft offering the sanctioned alternative.
He tried hard to find the gap where the system would act on its own or skip the human check, could not, and moved from reviewer to advocate.
Owen's dashboard, liveLumen · IT Service view · illustrative data
lumen-aios · it-service.html
Illustrative result
“Not fewer people using AI. More people using AI Owen actually knows about.”
4 mins
to review a new tool, was 40
Day one
joiner access ready, not day three
~140
staff served by a three-person team
Export
compliance evidence, not a week's rebuild
Illustrative figures for a synthetic company. Real before-and-after numbers come from the programme measurement spine, captured from actual participants.
What Owen's system is made of
Knowledge base
The IT service model and SLAs, the joiner, mover and leaver process, the device and app-estate approach, and the AI-tool review and sanctioning process.
Skills
Triages tickets, serves self-service answers, prepares access requests, tracks the estate, maintains the shadow-AI register and supports the rollout.
Dashboard
The ticket queue and SLA, the asset and identity register, access and security exceptions, and the sanctioned-versus-shadow AI-adoption view.
Tasks
Weekday morning brief, a ticket-triage and self-service pass, a joiner-mover-leaver chase, a periodic access-review nudge, AI-usage refresh, lint.
Connections
The service desk, SSO and MFA identity, device records, expense and traffic signals, the connector approvals, and consent-recorded IT meetings.
Sam Doyle
Account Executive · Lumen · a rep on Marcus's team
Sam Doyle is an Account Executive at Lumen, one of Marcus Bell's reps working the Core SMB end of the pipeline. He carries a personal quota of about £240k of new ARR a year and runs four live deals end to end.
Before · you do the work
Admin eating the selling hours
Sam sells all week, then loses the evenings to writing it all up.
Yesterday's call goes soft before he types three thin lines into the CRM and moves on.
Harlow has gone quiet on price and he has no clean signal for how urgent it really is.
He builds the Dale Valley demo from memory and a skim of their website in the car.
After · it works for you
The rep who stopped taking notes
The same week, with the admin around the deal lifted off him.
Calls write themselves up against the playbook and MEDDPICC, filed to the CRM with tasks created, before he is back at his desk.
Harlow surfaces first, one customer across two systems, the Amber deal and the Red account joined as one thing that needs him today.
He walks into every meeting present and listening, having read a prepared brief instead of a website in the car.
Sam's dashboard, liveLumen · My Sales Desk · illustrative data
lumen-aios · sales-rep.html
Illustrative result
“Nothing here is technically hard. What changes is who is left holding the admin, and it is not him.”
6-7 hrs
a week lost to admin
~1 hr
a week reviewing and editing
Same day
follow-ups out, not days late
£240k
personal new-ARR quota
Illustrative figures for a synthetic company. Real before-and-after numbers come from the programme measurement spine, captured from actual participants.
What Sam's system is made of
Knowledge base
The one shared Sales knowledge base, the playbook, MEDDPICC, battlecards and deal-desk rules, read at rep scale.
Skills
The lumen-sales-rep sub-skill, a role variant that researches accounts, keeps his own MEDDPICC true and drafts quotes.
Dashboard
My Sales Desk, his own deals against quota, MEDDPICC gaps, customer health and a personal coaching view.
Tasks
The same morning brief and email pass as the team, scoped to Sam's own deals and inbox.
Connections
Meeting intelligence records and transcribes his calls, routing write-ups to the CRM and drafts to Pending Review.
04 · What a Business AI Operating System is made of
Five parts. One system.
01
A living knowledge base
Everything about your company, products, people, policies and methods, in one place the AI actually reads. It stops being your job to remember everything.
02
Skills
Role-specific AI workers that do real tasks: research, drafting, analysis, coaching, the repetitive work. Built for how you actually work.
03
Dashboards
Not just numbers. Here is what it means and what to do next. Analysis, not reporting.
04
Task management
Keeps you organised, and keeps the system itself on track, so nothing falls through.
05
Connections
Plugs into the systems you already run: CRM, ERP, case management, email, calendar, files. It works where your work already lives.
05 · What changes
Better for everyone it touches.
For you
Hours back. Sharper decisions. Less reactive, and more of the work that only you can do.
For your team
Shared knowledge, consistent quality, and less bottlenecked on any one person.
For the company
Faster and more capable, compounding as the system learns your business.
For your customers
Quicker responses, better-prepared people, and fewer things dropped.
06 · The difference
We build it with you. Then we teach you to run it.
Most people will either build you something and leave, or teach you and hand you a blank page. We do both. Lumen is what "built for you" looks like. The programme is how you get your own, and how you learn to keep extending it.
