Q
ApisQ
Orchestrated Intelligence

AI capability for
growing businesses

The AI and automation capability of a much larger company, without hiring the team — delivered as a service, on your own infrastructure where it matters.

Initial capability pack · August 2026

Why now

The day disappears into admin,
and nobody on staff can build the fix

Every growing business we meet has the same four problems, in the same order.

The day disappears

Meetings, email, documents, formatting, chasing. The work that actually earns money happens in the gaps.

Everyone says use AI

Nobody on the payroll can turn that into something that runs reliably on Monday morning, on your data, without breaking.

Shadow AI is already here

Staff pay for chatbot subscriptions themselves. Usage is scattered, unmeasured and occasionally puts your material somewhere you would not choose.

Hiring does not solve it

An AI engineer is slow to find, expensive to keep, and one person leaving takes the whole capability with them.

Capability without headcount is now a buildable thing.
Our thesis

AI as a workforce multiplier,
embedded in your workflows, never above them

What AI does

  • Listens: meetings recorded, transcribed, summarised, turned into assigned actions
  • Reads: contracts, invoices, reports and decades of files, made answerable
  • Sorts: email triaged, requests routed, follow ups tracked
  • Drafts: newsletters, briefs, reports and replies, with sources attached
  • Watches: your market, your mentions, your competitors, around the clock
  • Files: everything landing in one governed place instead of six

What stays yours

  • Every decision, commercial and operational
  • The judgment about what matters and what to ignore
  • Every client, supplier and staff relationship
  • Final sign off on anything that leaves the company
  • The systems themselves: documented, transferable, yours to keep
Every AI output is evidence for a human decision, never the decision itself.
The map — and this deck's table of contents

One working day, six stages.
AI helps at every one.

CAPTUREmeetings · email · documents
KNOWLEDGEyour files, answerable
WATCHmarkets · mentions · content
DECIDEevidence · guards
AUTOMATEworkflows · publishing
RUNone panel · costs · resilience

Everything compounds

The meeting transcript feeds the knowledge base. The knowledge base grounds the drafts. The drafts are watched by the same monitors, run by the same panel, on the same governed data. Each stage makes the next one cheaper.

How we will use it

Every capability slide that follows highlights the stage it serves, and each stage in this diagram is a link. The point is not any single feature: it is that the six stages share one foundation.

Why integrated

Point tools solve slices.
The glue is where the cost and the errors live.

The subscription sprawl pattern

  • A note taker here, a chatbot there, an automation tangle nobody owns
  • Each with its own login, its own copy of your data, its own invoice
  • Your team becomes the integration layer: re keying, reconciling, reformatting
  • Nothing knows what the others know, so nothing compounds
  • And every competitor can rent exactly the same tools — no lasting edge is built on software anyone can buy

The ApisQ pattern

  • One foundation: your documents, data and workflows in one governed place
  • One knowledge layer: everything the business knows, indexed together
  • One governance model: access, audit and backup handled once, for everything
  • Every new capability plugs into the last, and compounds
  • The build encodes your way of working, and the intellectual property stays yours

Plenty of vendors sell a feature. We build the connected capability — an in house asset your business owns outright.

Not a concept deck

This is the estate we run our own company on.

~3 min
from a recorded meeting to owned action items, down from ~50 minutes Appendix A1
18,000
files of email and notes made answerable — 68,464 searchable chunks, people and organisation profiles Appendix A3
25,000+
documents machine transcribed through benchmarked engines Appendix A4
6
monitoring services in scheduled production Appendix A5
~30
services and jobs under one health panel Appendix A7
700M+
rows of governed data under automated audit, across 44 stores Appendix A8
3
published content feeds produced and shipped hands off Appendix A6
$0
in AI fees per meeting: it runs on local models, and every run is cost instrumented Appendix A9

This is our own working estate — the tools we run our own company on, daily — built and operated by the same small team that would build yours. Nothing here is a client deployment, and we do not present it as one. The gold chips link to the appendix slide that justifies each claim.

Capability · Capture
CAPTURE
KNOWLEDGE
WATCH
DECIDE
AUTOMATE
RUN

A meeting ends. The actions are already assigned. Demonstrated

Recording

The call recording is collected automatically after the meeting ends

Transcribe

Transcribed on your own hardware, with speaker attribution — nothing leaves the building

Summarise

A structured summary: decisions, open questions, who said what

Actions

Action items with owners and due dates, deduplicated against what is already on the list

Your task system

Uploaded before people are back at their desks

~3 minutes, $0

Measured against the ~50 minute manual pipeline it replaced. It runs on local models, so the AI fee per meeting is $0.

A gate, not a hope

Transcript quality is scored first. If it fails, every downstream step is blocked — you get a warning, never a confident summary of a bad recording.

Owners, resolved

Nicknames, initials and full names map to real people, so an action lands on the right list rather than in a paragraph nobody owns.

This runs on our own meetings, every week. Appendix A1

Capability · Capture
CAPTURE
KNOWLEDGE
WATCH
DECIDE
AUTOMATE
RUN

An inbox that triages itself Demonstrated

What it does, every few minutes

  • Every incoming message classified: client inquiry, billing, partnership, admin, noise
  • Genuine requests become tasks with an owner, a due date and a priority
  • Categories route to the right person or channel automatically
  • Spam and bulk mail suppressed rather than deleted, so nothing is lost
  • A morning digest, plus a follow up radar for anything that went quiet

The guard that makes it safe

The agent is structurally prevented from sending mail as a person. Allowed and blocked addresses are checked when the service starts, not when it is about to send, so a misconfiguration fails at boot rather than in your client's inbox.

Nothing is sent on your behalf without a human deciding to send it. The system's job is to make the decision take ten seconds instead of ten minutes.

Running as a daemon on our own mailboxes across email and messaging channels. Appendix A2

Capability · Knowledge
CAPTURE
KNOWLEDGE
WATCH
DECIDE
AUTOMATE
RUN

Your files, finally answerable Demonstrated

Decades of documents, email exports and notes sit in folders nobody searches. We turn that into a knowledge base your team questions in plain English, with every answer citing the file it came from.

Ask your knowledge base
Illustrative reconstruction of a real query pattern — every claim carries the document it came from.

Proven on a genuinely private archive: 18,000 files of email and notes processed on premises into 68,464 searchable chunks, with people profiles, organisation profiles and reconstructed email threads. Accuracy is measured against a fixed benchmark question set before anyone relies on it. Appendix A3 Appendix A4

Capability · Watch
CAPTURE
KNOWLEDGE
WATCH
DECIDE
AUTOMATE
RUN

Eyes on the outside world

Machinery that never sleeps Demonstrated

  • News and public attention tracked continuously at zero subscription cost
  • Social chatter read by language models for what it actually says, not counted by keyword
  • Thresholds measured against each subject's own normal, so an alert means something
  • Digests a manager reads in thirty seconds, with the evidence one click away

Pointed at what you watch In build on engagement

The machinery is demonstrated on investment watchlists in our own estate — 6 monitoring services in scheduled production. Retargeting it is configuration, not a rebuild.

Your brandYour competitors Your suppliersYour market Regulatory changesHiring signals

One person can genuinely watch three things. The monitors watch everything you name, around the clock, and tell you where to look first. Appendix A5

Capability · Automate
CAPTURE
KNOWLEDGE
WATCH
DECIDE
AUTOMATE
RUN

Content and publishing, on autopilot Demonstrated

Sources in

Newsletters, articles, videos and audio fetched on a schedule, transcribed, and summarised into structured digests you can actually skim.

Finished output

3 published content feeds produced and shipped hands off, plus a video channel watcher that briefs us every morning before anyone opens a laptop.

Validation that fails loudly

The publish step checks its own output and refuses to ship a broken feed. A pipeline that fails silently is worse than one that does not run.

The same machinery drafts your newsletters, client briefs and monthly reports — in your voice, grounded in your own documents, with citations attached for the human who signs it off. Appendix A6

Everything published still passes a person. The pipeline removes the assembly, not the judgment.

Capability · Decide — the trust slide
CAPTURE
KNOWLEDGE
WATCH
DECIDE
AUTOMATE
RUN

What our automation refuses to do

The clarity gate Demonstrated

An agent asked to do something ambiguous asks a specific question instead of guessing. Demonstrated in our ticket to pull request pipeline, which runs unattended around the clock.

The send authority guard Demonstrated

No agent can send mail as a person. The allow and block lists are checked when the service starts, so a mistake fails at boot.

Human review queues Demonstrated

Low confidence outputs are parked for a person rather than pushed through. Confidence is a number the system reports, not a feeling.

Read only by default Demonstrated

Anything that only needs to read cannot write. Enforced by the database and the operating system, not by the application asking nicely.

An automation you cannot describe the limits of is not an automation you should be running.
Capability · Run
CAPTURE
KNOWLEDGE
WATCH
DECIDE
AUTOMATE
RUN

One panel that runs the shop Demonstrated

Everything on one screen

~30 services and scheduled jobs: live status, last run, logs, start and stop, and a run now button. No terminal, no guesswork.

Freshness you can see

Backup tiles turn the banner red when a copy goes stale. The absence of a success is itself checked, every day.

The same on both platforms

Identical discipline on Mac and Windows, so the estate does not depend on which machine somebody happened to use.

You always know what ran, what failed, and what it cost. Appendix A7 Appendix A9

This is the layer most AI projects skip, and the reason so many of them quietly stop working three months after launch.

Capability · Run
CAPTURE
KNOWLEDGE
WATCH
DECIDE
AUTOMATE
RUN

Your data, governed Demonstrated

The boring superpower. One governed database instead of scattered spreadsheets and files nobody can join together.

One place

700M+ rows across 44 stores in a single governed database, so a number in one report can be joined to a number in another.

A catalog that cannot lie

Generated from the live system, not hand written. Documentation that disagrees with reality fails its own check.

Audits that caught real defects

Automated audits run against the estate and have found 2 real defects that no person had noticed.

Recoverable by routine

Least privilege enforced by the server, immutable history, twice daily backups with a verified copy shipped off site nightly.

We learned the hard way that sync is not backup — and we wrote up why. Appendix A8 Appendix B3

What it is worth

Three ways this pays

Value leverWhere it landsIllustrative before → after
Cut costMeeting minutes, document intake, routine reporting, chasing follow upsA meeting: ~50 minutes of write up → ~3 minutes, at $0 in AI fees. Supplier documents: manual re keying → structured data with exceptions flagged.
Scale capacityWhat one person can watch, read, draft and produce in a weekWatching the market: three things by hand → everything you name, around the clock. A monthly report: assembled → regenerated.
Raise qualityProvenance, institutional memory, consistency, fewer manual errors"Where did this figure come from?" → answered with the source attached. A leaver's knowledge → stays in the business.

The same lean team, with materially bigger reach. We deliberately quote no percentage saving: the honest number depends on where your hours actually go, which is what the first conversation is for.

No pricing in this pack. Commercial structure is shaped in conversation once the first use case is chosen.

What it feels like

A day in the augmented business

WhenWhat happens
07:30The morning digest is waiting: the inbox already triaged, two mentions worth knowing about, one competitor announcement, each with a thirty second summary.
09:15Someone asks the knowledge base "what did we quote this client last time, and on what terms?" — a cited answer in seconds, from your own files.
11:00Supplier documents land as structured data instead of attachments. Anything that does not reconcile is flagged for a person, never guessed.
14:00The team meeting ends. Decisions and action items are on the right people's lists before anyone is back at their desk.
16:30The monthly report regenerates from live data. Checked and sent inside the hour, rather than blocking a day.
17:30Everyone goes home. The estate keeps collecting, transcribing, watching and filing, and it is all in tomorrow's digest.

Nobody's job changed. Everybody's reach did.

How we would work together

Three doors. One foundation.

① Consulting sprint

A scoped, embedded engagement — typically four weeks, two consultants working alongside your team. The use case is chosen together in week 1, and you finish with a working first version, not a slide pack.

Training and build in one. Designed explicitly against the documented reasons most AI projects fail. Appendix B2

② Embedded retainer

Your AI function, without the headcount. A weekly working session, a visible request board, and everything documented inside your own estate as it is built.

Yours to keep. If we disappeared, a competent successor could pick it up from the documentation alone. Appendix B1

③ Productised tools

The meeting pipeline, email triage, the knowledge base, the monitors — stood up as services on your infrastructure.

These are honest generalisations of tools we already run internally, adapted to your systems. That is a real build, not a shrink wrapped product. In build per client

Or name the workflow that hurts most, and we will tell you honestly whether we can reach it. Every door leads onto the same shared foundation, so wherever we start, the next step is already half built.

How this is kept safe

Private by architecture,
auditable by default

1 · Identity at the edge

Zero trust access for anything browsable: an allowlist, a one time code, identity checked on every single request. Nothing sits open on the internet waiting to be found.

2 · Read only applications

Dashboards and viewers hold no credential that could write, so a compromised page cannot corrupt anything.

3 · Least privilege data access

Enforced by the database server itself. Collectors write only their own area; everything else reads. Secrets live in a vault, never in code or files.

4 · Sensitivity tiered AI

Private workloads run on local models on hardware you control — nothing leaves the building. Cloud models touch only non sensitive work, and only under a policy you agree in advance.

5 · Immutable history and alarmed backups

Collected data is never edited in place; every structural change is versioned and logged; twice daily backups with a verified copy off site nightly, and an alarm if the chain ever breaks.

This is how our own estate already runs. Governance was not added for this deck.

Why us

Why ApisQ

We built it, and we use it daily

Not a proposal to integrate other people's products. A working estate: capture, knowledge, monitors, publishing, one ops panel, one governed database — operating together today. This deck was produced and shipped by the estate it describes.

We show our failures

Published negative results and written post mortems, including the publishing bug in Appendix A6. A partner who only shows wins is a vendor.

Built for firms that cannot hire this team

You do not need to find, afford and retain an AI engineer to have the capability. We provide it as a service, one workflow at a time, with no enterprise rollout.

With you, not for you

We build alongside your people and document everything in your own estate. Transferable by design, so the capability outlasts any individual, including us.

Next steps

Let us find your first hour back.

1 · Conversation

A working session with the people doing the work: where do the hours actually go, and which of them would you most like back?

2 · Pick a door

A consulting sprint, an embedded retainer, or a productised tool stood up on your infrastructure. Whichever fits how you prefer to buy.

3 · First result

A first working outcome within weeks of the build starting, with go and no go gates throughout, so you are never committed beyond the last proven step.

ApisQ
Orchestrated Intelligence
info@apisq.co

Initial capability pack for discussion — not a formal engagement proposal. Prepared in the United Kingdom. ApisQ provides technology and workflow capability; nothing here constitutes legal, financial or professional advice.

Appendix

Deeper detail, and the failures we publish

For the person who wants to see how it actually works before recommending it. Everything described here is our own estate, running today.

A · Capability deep dives

The nine capabilities from the main pack, taken apart: pipelines, gates, engines, numbers and the post mortems

B · How we work

Phasing and gates, the embedded engagement, build versus buy, and what a first engagement needs from you

C · FAQ & glossary

The questions everyone asks first, and plain language definitions of every term used in this pack

Appendix A · Capability (1 of 9)

From recording to assigned actions

1 · Collect

The recording and any auto transcript are pulled once the meeting ends

2 · Transcribe

Local speech model on your own hardware — no upload, no per minute fee

3 · Diarize

Who spoke when, separated into distinct speakers

4 · Merge

Words and speakers reconciled into one attributed transcript

5 · Summarise

Two passes: structure first, then decisions, risks and open questions

6 · Deduplicate

Actions checked against what is already on the board before anything is created

7 · Upload

Tasks created with owner, due date and the transcript link attached

The quality gate

Every transcript is scored before the summary stage. Below the bar, the pipeline stops and says so. A confident summary of an unusable recording is the worst possible output, so it is the one thing the design makes impossible.

Owner alias mapping

A maintained map turns a nickname, a set of initials and a full name into one person, so an action reaches a real list. Unmapped speakers are flagged rather than assigned to nobody.

Before and after

~50 minutes of listening back, writing up and chasing, per meeting. Now ~3 minutes end to end, at $0 in AI fees, because transcription and summarisation both run locally.

Demonstrated on our own weekly meetings. The pipeline is not a demo path: it is how our action lists get written. Demonstrated

Appendix A · Capability (2 of 9)

Inside the triage service Demonstrated

StageWhat happensWorth knowing
ClassifyEvery message labelled: client inquiry, billing, partnership, supplier, admin, newsletter, spamThe taxonomy is yours to edit; categories are configuration, not code
ExtractGenuine requests become tasks with an owner, a due date and a priority, linked back to the threadA message that is not a request creates nothing — silence is a valid outcome
RouteEach category goes to the right person or channel automaticallyWorks across email and messaging channels, same rules
Digest and radarA morning summary of what arrived and what was created, plus threads you sent that went quiet, resurfaced on a scheduleThe radar is the part people notice most: nothing slips because everyone was busy

The send authority guard

The agent cannot send mail as a person. Allowed and blocked addresses are validated when the service starts, so a bad configuration stops the service at boot rather than surfacing in a client's inbox.

Token lifecycle

Mailbox credentials sit in a vault, refresh automatically, and are scoped to the minimum permission the job needs. A token about to lapse raises an alert on the ops panel Appendix A7 rather than failing silently at 3am.

Appendix A · Capability (3 of 9)

An archive of 18,000 files, made answerable

18,000
files of email and notes ingested
68,464
searchable chunks in the index
100%
processed on premises — nothing uploaded

1 · Organise

Raw files reorganised into a consistent structure with clean text, not left as a folder dump

2 · Extract

People, organisations, dates and commitments pulled out of the text

3 · Connect

A knowledge graph plus a vector index, so related facts surface together

4 · Profile

A profile per person and per organisation, assembled from everything that mentions them

5 · Answer

Questions answered only from retrieved passages, each claim citing its file

Threads, reconstructed

Scattered replies and forwards are stitched back into coherent conversations, so "what did we actually agree" has one place to look rather than eleven.

Measured before it is trusted

A fixed benchmark question set with known answers is re scored after every change. Retrieval quality is a number we can show you, not a claim.

Our own private archive, processed on our own hardware. The same pipeline pointed at your archive produces the same result. Demonstrated

Appendix A · Capability (4 of 9)

The document factory

25,000+ documents machine transcribed to date. Any format in, clean searchable text out, with a path back to the source page.

What comes inHow it is readWhy that route
Digital documentsDirect text extractionExact, instant and free — never send a machine readable file to a model
Scans and photographsTwo vision engines plus mechanical extraction, results fusedEngines disagree; taking the best of three beats trusting any one
Slides and spreadsheetsStructure aware extractionA table read as prose is worse than not reading it at all
Audio and videoLocal speech model, then the summariserSame engine as the meeting pipeline Appendix A1
HandwritingVision engine, low confidence flaggedAnything uncertain is marked for a person, never quietly guessed

Provenance in the filename

Every output records which engine produced it and from which source page. When a downstream answer looks wrong, you can find out why in one step instead of re running the batch.

The failure modes post mortem

We benchmarked the engines against each other and wrote up where each one fails: dense tables, multi column layouts, faint scans, rotated pages. Choosing a route per document is data, not a preference.

Appendix A · Capability (5 of 9)

Anatomy of a monitor

StepWhat happensWorth knowing
CollectScheduled collectors pull news, public attention and social chatter for each subject you nameThe news tier runs at zero subscription cost
UnderstandA language model reads each item: is it really about you, and what is it actually saying?"Great, another outage" is not praise
AggregateWeighted by reach, per subject, against a rolling baseline of what normal looks likeMeasured against that subject's normal
AlertThreshold breached, alert raised with a written narrative and links to the underlying itemsWritten for a thirty second read
Everything published about your subjectsnews, public attention and social, collected on a schedule
Actually about you, scored and weightedtarget aware reading drops the coincidental matches
A handful of alertseach with a thirty second narrative and the evidence attached
6 monitoring services in scheduled production, scored on local models. Demonstrated
Appendix A · Capability (6 of 9)

Content pipelines, and the bug we published

What runs today

3 published content feeds produced and shipped hands off, a newsletter and article digest pipeline, and a video channel watcher that briefs us each morning. Demonstrated

The shape of a pipeline

Fetch → transcribe → summarise → assemble → validate → publish. Every stage is restartable, and running it twice produces the same result rather than a duplicate.

Validation that fails loudly

The feed is checked against its own specification before publishing. A pipeline that ships a broken feed quietly is worse than one that stops.

The episode numbering bug — told honestly

Two machines running the same publishing job allocated the same episode number 105 times. Because the feed derived each item's identity from that number, podcast apps treated the collisions as the same episode and silently hid the duplicates. Nothing errored. Nobody noticed until a listener asked where an episode had gone.

The fix was not a patch: we made the collision structurally impossible by changing how identity is allocated, then added a validator that fails the publish if two items ever share one. We wrote it up rather than quietly correcting it — because finding the failure mode you would never have looked for is most of what this work is.

Appendix A · Capability (7 of 9)

The ops panel

~30 services and scheduled jobs on one screen, in a browser, with no terminal involved.

Services

Always on daemons: triage, monitors, collectors. Live status, uptime, start and stop, and the last hundred log lines without leaving the page.

Jobs

Every scheduled task with its last run, its outcome and a run now button — so "did the backup happen?" is a glance, not an investigation.

Health checks

Backup freshness tiles turn the banner red when a copy goes stale. The absence of a success is checked, which is the failure people miss.

Feature flags

Capabilities can be turned off without a deployment, so a misbehaving pipeline is stopped in seconds by whoever is at the desk.

The same panel and the same discipline on Mac and Windows. Cross platform parity is deliberate: an estate that only works on one person's machine is a liability, not a capability.

This is the layer that decides whether an AI project is still working in six months. It is also the layer nobody demos. Demonstrated

Appendix A · Capability (8 of 9)

The governed warehouse

700M+
rows under one governed database
44
stores, each catalogued and versioned
2
real defects caught by automated audit
verified backups per day, one shipped off site nightly

A catalog generated from the live system

  • The catalog is produced from the database itself, so documentation cannot drift from reality
  • A staleness check fails the build when it does drift, which is how you know it is true
  • It is readable by people and by AI assistants, which is what makes grounded answers possible at all

Audits that caught real defects

  • Automated audits run over the estate on a schedule, not on request
  • They have found 2 real defects that no person had noticed — the point of the exercise
  • Access is least privilege and enforced by the server: a collector can write only its own area, everything else reads
  • Collected data is immutable; corrections live in a separate layer, so nothing is silently rewritten

Our own estate, under exactly this discipline. Backup strategy and the resilience story: Appendix B3 Demonstrated

Appendix A · Capability (9 of 9)

What the AI actually costs

Per run cost logs

Every AI call records what it cost, which model served it and which job asked. "What are we spending on AI?" has a number behind it, broken down by workflow.

Routing by complexity

Simple, high volume work runs on local models at no per query fee. Cloud models are reserved for the genuinely hard queries, under a policy you set.

The metered key guard

A job that would silently bypass a flat rate plan and start billing per call refuses to start. The guard checks which credential is in play before any work begins.

Why the guard exists

The expensive failure in AI operations is not a big bill you notice. It is a misconfigured job quietly running on a metered credential for a fortnight while everyone assumes it is covered by the subscription. Making that condition impossible to start is cheaper than monitoring for it, and far cheaper than discovering it on an invoice.

The meeting pipeline is the clean case: transcription and summarisation both run locally, so the AI fee per meeting is $0 Appendix A1. Where local inference replaces a metered call, the saving is estimated from the measured per run cost rather than asserted.

Appendix B · How we work (1 of 4)

How an engagement unfolds

Discover
Week 1

Map where the hours go; choose the first use case together

Gate
your call
First capability
Weeks 2 to 4

A working first version, visible to the whole team

Gate
your call
Compound
Months

Extend along the six stage spine, one gate at a time

Gate
your call
Institutionalise
Ongoing

Your team runs it; the capability is yours

Every gate is a genuine exit. You are never committed beyond the last proven step.
PhaseFocusExample deliverables
DiscoverWhere the hours go; what documents, data and tools exist; agree the success testWorkflow map of your business · prioritised opportunity list · scoped first build
First capabilityDeliver the chosen use case end to end, visibly, on your infrastructureMeetings to actions live · inbox triaged · a first corpus answerable
CompoundExtend along the spine, each capability reusing the foundation the last one builtMonitors on your subjects · publishing pipeline · ops panel · governed data
InstitutionaliseYour team runs it; we advise, extend and keep the estate healthyDocumentation and catalogs · runbooks · training · roadmap reviews
Appendix B · How we work (2 of 4)

Two audiences, one program

Leaders get literacy and governance

  • What these systems can and cannot do, in business terms and with worked examples
  • How to tell a real capability from a demonstration, and how to read the cost and the risk
  • An AI usage policy that replaces shadow AI with something sanctioned and measured

The team gets the build

  • Two consultants working alongside your people, on your systems, typically four weeks
  • The use case is chosen together in week 1, from where the hours actually go
  • Training and build are the same activity: they learn it by watching it get built, and everything is documented in your estate as it is written
Documented failure modeWhat we do about it
Technology before problemThe use case is chosen from your hours in week 1. No tool is selected before the problem is written down.
No measurementA success test is agreed up front: what, observed by whom, would make this a clear win.
Sponsorship decayOne named sponsor, a weekly session in the diary, and a visible request board anyone can read.
The scaling wallA pilot that cannot become production is a demo. Governance, the ops panel and backups are part of the first build, not phase two.
Appendix B · How we work (3 of 4)

Build vs buy, honestly

We buy commodities

Foundation models, storage, mail plumbing, meeting capture. No pride of authorship where a vendor is genuinely better value. Vendor evaluations are written down before spend, not justified after it.

We build the connective tissue

Your knowledge layer, your triage rules, your monitors, your publishing voice, your catalog. The things no vendor can sell you because they are yours.

Always yours to own

The integration layer and the documentation. That is the defence against subscription sprawl, and it is handed over rather than held hostage.

Resilience, and the day we learned sync is not backup

File sync copies your mistakes at the speed of light. Delete something, or corrupt it, and the healthy copy is overwritten before you have noticed — which is exactly what a backup is supposed to prevent. We learned this on our own estate and rebuilt around it.

What replaced it: twice daily verified dumps, a copy shipped off site nightly, a separate ledger recording that each link in the chain actually happened, and an alert when one does not. Restores are tested, because a backup nobody has restored from is a belief, not a backup. Appendix A8

Appendix B · How we work (4 of 4)

What a first engagement needs from you

A sponsor

One person who owns the engagement and joins the weekly session. Roughly half a day a week, mostly deciding rather than doing.

Stepwise access

Read only access to the systems in scope for the chosen use case, granted one step at a time. Never everything on day one, and never more than the job needs.

Sample documents

A representative sample of the real material — the messy scans, not the tidy ones — so quality is measured on your documents from the first week.

An agreed success test

Written down before we start: what, observed by whom, would make this a clear win. It is also what makes the first gate a real decision.

What we do not need: a data migration, a platform decision, a committee, or anybody's job description to change.

Appendix C · FAQ (1 of 4)

The questions everyone asks first

"Will this replace my staff?"

No, and we would not take an engagement framed that way. It changes what their hours buy. The system collects, transcribes, sorts, drafts and files; your people question, decide, and hold the relationships.

The scarce resource in a growing business was never headcount. It was attention.

"Can we trust what AI writes?"

Trust is engineered, not assumed. Answers cite the file they came from; knowledge bases ship with a benchmark question set re scored on every change; low confidence outputs go to a human review queue; and nothing leaves the company without a person signing it off.

"We are not technical. Can we even run this?"

That is precisely what the ops panel Appendix A7, the written documentation and the retainer exist for. One screen, plain language, a run now button, and someone to call.

If a capability needs a specialist on staff to survive, we have not finished building it.

Appendix C · FAQ (2 of 4)

Running cost, and whether you are locked in

"What does it cost to run?"

  • High volume work runs on local models on hardware you own — no per query fee, and nothing leaves the building
  • The meeting pipeline costs $0 in AI fees per meeting for exactly that reason Appendix A1
  • Many feeds are free by design: the news monitoring tier runs at zero subscription cost
  • Every run is cost instrumented, and a job that would silently start billing per call refuses to start Appendix A9
  • The largest line item in most estates is duplicated subscriptions, which is what consolidation removes

"Are we locked in?"

  • It runs on your infrastructure and your accounts, not ours
  • Standard, widely used tools throughout — no proprietary format, no bespoke runtime
  • Everything documented and catalogued as it is built, so a competent successor can pick it up from the documentation alone
  • Walk away is a design goal: we would rather be kept because the work is good than because leaving is expensive

The honest counterpoint: any capability you come to rely on creates a dependency. The question worth asking a supplier is whether the dependency is on them or on the thing they built you. We build for the second.

Appendix C · Glossary (3 of 4) — AI & systems

The glossary, in plain language

TermWhat it actually means
LLM (large language model)AI trained to read and write text. Useful as a tireless junior who has read everything, and who must always show its sources.
AgentAn AI given tools and a goal, which decides its own steps — "find every message about this order and summarise what was promised" rather than one question, one answer.
RAG (retrieval augmented generation)The AI answers only from documents retrieved for that specific question, and cites them. It is the main defence against confident invention.
Knowledge graphA map of who and what your documents mention and how they connect: supplies, invoices, employs, agreed. It answers questions a keyword search cannot.
Vector index / embeddingA mathematical fingerprint of meaning, so the system finds passages that say the same thing in different words.
Grounding & citationsForcing the AI to answer from retrieved source text and show where each claim came from. The single most important control on trustworthiness.
HallucinationAn AI stating something fluent and false. Grounding, citations and benchmark testing reduce it to something you can audit rather than something you hope about.
Local modelAn AI model running on hardware you own. Data never leaves the building and there is no per use fee.
Benchmark question setA fixed list of questions with known answers, re scored after every change. It is how retrieval quality gets proved instead of asserted.
MCP (model context protocol)A common plug standard that lets an AI assistant use your tools and data sources without a bespoke integration for each one.
Appendix C · Glossary (4 of 4) — operations & governance

The glossary, in plain language

TermWhat it actually means
Least privilegeEvery system gets the minimum rights it needs and nothing more, so a tool that only reads cannot write, even if something goes wrong inside it.
Immutable historyCollected data is never edited in place. Problems are flagged alongside it and corrections applied in a separate layer, so nothing is silently rewritten.
Audit trail & provenanceThe recorded path from a number on a screen back to the document it came from. It is what lets anyone ask "where did this come from?" and get an answer.
Zero trust accessEvery request verifies identity first. Nothing is trusted merely for being "inside the network", and nothing is left open on the internet.
Shadow AIStaff using AI tools nobody sanctioned, on their own accounts. Usually well intentioned, always unmeasured, and occasionally puts your material somewhere you would not choose.
Human in the loopA person reviews before an output has effect. Applied where the cost of being wrong is real, rather than everywhere, so the review still gets read.
Idempotent pipelineRunning it twice produces the same result as running it once. It is what makes a failed job safe to simply restart.
Daemon / serviceA program that runs continuously in the background rather than when somebody launches it. The triage and monitoring capabilities are services.
Quality gateA check that stops the next step when its input is not good enough — the reason a bad recording produces a warning instead of a confident, wrong summary.
ApisQ
Orchestrated Intelligence

Thank you

Tell us where the hours go, and we will tell you honestly what we can give back.

info@apisq.co

Initial capability pack, v1 · August 2026 · Prepared in the United Kingdom for discussion purposes only. This document is not a formal engagement proposal. Every capability described is drawn from ApisQ's own working estate and is labelled Demonstrated, In build or Roadmap accordingly; no client implementation is claimed or implied. ApisQ provides technology and workflow capability; nothing here constitutes legal, financial or professional advice.

1 / 39