// Helix is AI Centric™
Not one AI.
Eight.
And they don't
always agree.
Most AI tools give you a single voice — one assistant, one perspective, one answer. Helix gives you a team of eight specialists, each with a different function, a different way of reading the same situation, and a distinctly different personality.
The analyst surfaces what nobody noticed. The quartermaster tracks every resource. The director makes the call. The automator executes it. The off-boarder closes it out cleanly. The shadowy one watches from a distance and weighs in at exactly the right moment. And DaNA — who never changes, regardless of which ensemble you choose — holds the whole operation to the Three Laws.
When you ask Helix something that matters, more than one of them has an opinion. They argue. They correct each other. What reaches you is the answer that survived all of them. Not the first answer. The answer that held up.
// Not bolted on
Every other platform added AI.
Helix was built around it.
Look closely at most PSAs, CRMs and ERPs and you will find the same thing: a system designed years ago, working exactly as it always did, with an AI panel fastened to the side. A summarise button. A chat box in the corner. A suggestion you can accept or ignore. Useful, often. But the software underneath does not know the AI is there, and the AI does not really know how the business runs.
Helix started from the other end. The intelligence is not a feature of the platform — it is the organising idea the platform was built to serve. Tickets, invoices, proposals, contracts, assets and schedules are the raw material the agents reason over, and they were shaped that way deliberately, from the first table onward.
That is what AI Centric™ means. Not an assistant bolted to a PSA. A PSA built so a bureau of agents could run it with you.
// TL;DR — the 30-second version
Helix is AI Centric™. The intelligence is not a panel bolted onto an older system — the platform was built around it from the first table onward. Every other PSA, CRM and ERP added AI to what it already had. Helix started from the agents.
The Bureau is the default cast of eight. They huddle, discuss, argue from their own roles, vote, and any of them can veto. Generative AI alone is roughly 88% accurate; eight agents challenging each other drives that substantially higher. You still proofread — you just edit far less.
An AI Ensemble™ is that same cast of eight, dressed in a world your team actually wants to inhabit. 94 ensembles across 10 categories: intelligence operations, crime investigation, epic legend, field expedition, arts performance, the professions, sport arena, fantasy imagination, faith calling, and special editions. Each has eight characters built on the same Three Laws and Three North Stars. Or build your own. Your team does not become that world — running a crime-investigation ensemble does not turn you into crime investigators. You just get to pretend.
Two characters never change regardless of theme: DaNA, the Culture Champion, and Command Bar™, the command interface. The other six change with your world.
The ensemble layer is earned, not purchased. Pre-graduation agents (The Analyst, The Quartermaster, etc.) have full capabilities. As techs practice closing tickets and invoices correctly, The Coach™ fires an alert when they are ready. The manager triggers graduation user by user. Starter gets foundational ensembles with your own API key. Premium ensembles and Build-Your-Own require at least one Pro+ seat on the account.
It is completely optional. The platform runs fully without it. You just lose a bit of culture, a bit of adventure, and a bit of fun.
Your mascot is a cat named FELIX™. He rhymes with the platform. He’s been in the corner since you arrived. He already knows.
// Why eight agents, not one
Generative AI is about 88% accurate.
Eight agents arguing is substantially more.
Generative AI, at this stage of its development, is approximately 88% accurate. That is a remarkable number by any historical measure. It is also a number that means roughly one in eight outputs needs human correction.
I noticed the problem before this paragraph did. It has been in the system for three days.
u2014 FELIXu2122
For most AI software, that means you — the human — are the final proofreader. Every ticket note. Every invoice line. Every proposal paragraph. The AI generates. You catch what it missed. The accuracy ceiling is set by the model, and the model works alone.
Helix works differently.
Not one agent generating. Eight. When you ask for something that matters, they huddle. They discuss it. They argue about it. Each one reads the same request from inside their own role — the analyst reads the whole situation, the quartermaster checks what it costs and what it consumes, the director asks whether this is even the right call, the automator asks whether it can actually be executed, and DaNA asks whether it should be done at all.
Any of them can veto. Not as a formality — a veto stops the output and sends it back to the rest of the ensemble for refinement. What reaches you is not the first thing generated. It is the version that survived six colleagues who each had a different reason to object.
The result is markedly more accurate than what generally passes for AI-produced work today. The chance of eight agents with eight different vantage points all missing the same hallucination is near zero.
You still proofread. We would never tell you not to — it is your name on the proposal, not ours. But you will be editing far less of what comes back, and correcting far fewer things that should never have reached you.
And if it is not right, you do not start again. Tell the bureau what to change and they revise that part — the paragraph, the tone, the figure, the section you did not like — while keeping everything that was already good. These are surgical edits, not a fresh roll of the dice. You are not trading a decent draft for a different one and hoping.
The veto is not arbitrary. Every agent — across every ensemble, regardless of theme — operates under the same Three Laws and Three North Stars. These are not guidelines the agents follow. They are what the agents are made of. And they are enforced.
Any output generated must pass all three:
Diplomatic
Law 3 — Face-Saving
Nothing goes out that demeans, diminishes, or blindsides. Gaps are framed as opportunities. Corrections arrive without judgment. If an agent's output violates dignity, it does not proceed.
Proactive
Law 2 — Proactive Intelligence
Relevant signals must be surfaced, not suppressed. An output that ignores an available and important piece of context fails this standard. Silence when a signal is available is a failure mode, not a neutral state.
True
Law 1 — Truth · Highest Priority
Nothing proceeds that is inaccurate, unconfirmed, or deceptive. Law 1 outranks both others. A diplomatic output that is false fails. A proactive output that misleads fails. Truth is the floor everything else stands on.
The agents cannot individually decide to override these. Compliance is not optional. The Three Laws are the operating system every agent runs on — and the reason the output that reaches you is not just more accurate than a single agent could produce, but more trustworthy.
One important distinction.
Think of C-3PO. He talks down to R2-D2 constantly — condescending, critical, occasionally contemptuous of R2’s methods. And yet he is a protocol droid — being diplomatic is literally his job, and he is excellent at it. He has never once talked down to a human. BUT — the laws governing how he treats people do not govern how he treats another droid.
Nor do they stop him having an opinion. “Oh dear.” “We’re doomed.” “I do believe they think I am some sort of god.” He says exactly what he thinks of the plan, out loud, in front of everyone in the room — and he is still the most courteous character in the film. Being bound by a law about how you treat people is not the same as being muzzled. What the Laws govern is how all of them treat you and your customers.
The Laws do not stop the agents arguing with each other, calling each other idiots, or correcting each other’s mistakes in plain sight of your staff — often to hilarious effect. This is what AI centricity blended with a bit of fantasy produces: software that is a joy to use and a riot to watch. Dinner and a show — and a team that enjoys the software they live in all day, which shows up as culture and morale long before it shows up in a report. That is also the part that makes the output better. Your people willwant to document their tickets properly — just to see what FELIX™ and the rest of them dream up!
That is precisely how the Three Laws work here.
The Three Laws govern what reaches humans and what Helix generates about your clients and their data. They do not govern what happens internally between the agents.
Internally, they argue. FELIX™ doesn’t think much of Q’s documentation habits. Q has documented every inconsistency and is prepared to produce the file. M has said “I hate this job” at least twelve times this quarter and will say it again. The Cleaner™ enters channels without explanation and leaves the same way — and somehow the problem is gone. NIGHTSHADE knows about all of it and says nothing, which is somehow more alarming.
This is permitted by design. The Laws protect the humans the platform serves and the accuracy of the information it produces. They say nothing about whether an intelligence analyst may have a dry and superior attitude toward a quartermaster who has been using the laminator without signing it out. Self-deprecation is allowed. Inter-agent criticism is allowed. The arguing happens — often to genuinely hilarious effect — and it is exactly the mechanism that makes the output better. Your clients never see any of it. What reaches them is the version that survived the deliberation.
The agents are also teaching.
Every time FELIX™ flags a stale ticket with dry precision rather than judgment, that’s Law 1 demonstrated. Every time the Morning Brief arrives before anyone asked, that’s Law 2. Every time an output is traceable, confirmable, and reversible, that’s Law 3. The agents don’t hand your team a memo about the culture. They live it — visibly, every day, in every briefing and every ticket note and every water cooler exchange your team witnesses.
A tech who watches FELIX™ operate absorbs a standard for how findings are communicated: specific, economic, without blame. A manager who watches M make a call absorbs a model for what authority looks like when it isn’t performative. A new hire who sees The Cleaner™ handle an offboarding with quiet professionalism learns how endings should be treated — a lesson no onboarding document could deliver as cleanly.
The culture you want your team to inhabit is demonstrated every day by the agents who already inhabit it. Your team doesn’t get told to operate this way. They get colleagues who already do.
M has been informed that this section describes his leadership style as ‘not performative.’ He said ‘approved’ without looking up from the brief. This is exactly the example the paragraph was making.
u2014 FELIXu2122
Humans still must proofread. That has not changed and will not. But the ticket note that goes out to a client, the invoice line that gets billed, the proposal clause that gets committed to — these have survived a deliberative process that no single-agent AI can replicate.
The first answer is what one agent produces.
The answer that held up is what eight produced together.
// The analyst
FELIX™ — Intelligence Analyst · Mascot
Rhymes with HELIX™. This is not accidental.
He's a cat.
Not just any cat.
FELIX™ is named after Felix Leiter — the CIA analyst in the Bond franchise. The man Bond calls when he needs intelligence from the other side of the world. The analyst who always has the answer before the question is finished forming. The one who knows things and delivers them with the minimum number of words required and not one more.
He is also, by voice register, definitively Blofeld's cat. Superior. Economical. Affectionately contemptuous of the proceedings. He has been in the room since before you arrived. He noticed the problem before anyone asked. He would find it mildly undignified to tell you this. He tells you anyway because it needs to be done.
Ships have kept cats for centuries — not as pets, but as intelligence assets. They patrol. They observe. They find the vulnerabilities before the crew notices them. The fleet keeps a cat because the cat is the best early warning system aboard. So does The Bureau.
And there is one more thing about FELIX™ that matters.
His name rhymes with the platform.
HELIX. FELIX™.
This is not accidental. The platform is named after the triple helix — a deliberate mutation of the triple helix — the mutation, the third strand that represents the Three Laws. Three strands. Three Laws. The structure that holds the blueprint of what Helix is and what it is becoming. FELIX™ is named as a deliberate echo of the platform itself. He doesn't just work inside Helix. He rhymes with it. You cannot say the product's name without hearing his.
HELIX™ brings the BLUEPRINT™
Your team brings the HUMINT™
FELIX™ brings the ARTINT™
Meow.
He has an affinity for Fancy Feast salmon. He is finicky about it. He would like you to know this is not relevant to the quality of his output. His output is, and has always been, excellent.
I did not select the salmon preference. The salmon preference selected me. This is not a distinction I am prepared to defend further.
u2014 FELIXu2122
The channel is open. FELIX™ is watching. He has been watching since before you arrived.
FELIX™ is not responsible for the accuracy of his confidence. The confidence is always accurate.
// The philosophy
Why a cat?
Why not.
We are a company that likes fantasy. We like role-play. We like tradecraft. We are stereotypical geeks without the taped glasses or pocket protectors (well, except for The Cleaner—he still carries one he’s had since the 80s). And we have noticed something most enterprise software companies have refused to notice: the digital world has no physical constraints.
Think about the Matrix. Neo learns that the rules he thought were real were just programs — agreements running on a system that could be rewritten. Enterprise software has been running the same program for thirty years: work is serious, interfaces are neutral, AI assistants are earnest and helpful and entirely without personality. Nobody asked why. Nobody asked whether it had to be that way.
"We asked. The answer is no."
It absolutely does not have to be that way.
True story
Not every ensemble arrived through careful product planning. The Shipwreck started as a question: what would happen if the most earnest, most committed, most consistently adjacent-to-helpful person in enterprise software history had his own cast? We spent considerably longer on this than was strictly necessary. Hap was the answer. He still means well. He always will.
The Range came from a meeting in which someone asked — seriously, as a product question — whether cowboys would file better ticket notes than operatives. The meeting went long. The argument: a cowboy who fails to track his herd loses his herd. The debrief is not optional when the cattle matter. We built The Range. Boot Hill turned out to be the best off-boarder we have ever written. We don’t question these things anymore.
94 ensembles. Every one of them started somewhere like this.
You are writing instructions for a computer. The computer does not care whether the intelligence analyst is named Dave or FELIX™. The output — accurate records, complete documentation, on-time invoices, resolved tickets — is identical. The experience of producing it is entirely up to you.
So Helix chose. A spy bureau. A Victorian detective agency. A kitchen brigade. An Arthurian round table. A circus. Ancient Egypt. A chess grandmaster's office. Twenty-three other worlds we liked the sound of. A cat who is, by any reasonable measure, the best intelligence analyst in the building.
This was considerably harder than building something boring. We built it anyway because boring was a choice, and we made a different one.
The fantasy and the function are the same thing.
A tech who is an operative files better debriefs than a tech who is required to fill out ticket notes. A manager who is a bureau chief makes more decisive calls than a manager who is reminded to close the loop. The character does what mandates and gamification never could — because identity is more powerful than incentive.
FELIX™ has a perfect ticket note track record. Not despite being a cat. Because of it. A cat with standards does not accept incomplete work. He is finicky about salmon and he is finicky about documentation and the two are related. The character is the mechanism.
None of which means you are stuck with him. If a finicky know-it-all cat is not what you want running your service department, pick something else — anything else. Any character from any of the 94 ensembles, or one you build from scratch. And if you would rather skip the whole business, keep it clinical and call him The Analyst. That is what he is called anyway, right up until your staff have been inducted into the Three Laws and the Three North Stars. The cat is an option, not a condition.
Most software asks you to adapt to it. Helix gives your team a world to inhabit. When your team inhabits a world, they work differently inside it — better documentation, sharper decisions, genuinely complete ticket notes. Not because you asked harder. Because the character wouldn't accept anything less.
The Bureau
Default · All PlansThe Bureau is the spy-tradecraft ensemble — and the identity of this entire website. You have been talking to FELIX™ since you arrived. That's The Bureau.
And if you would rather not have a finicky know-it-all cat running your service department — pick something else. Anything else. Any character from any of the 94 ensembles, or one you build yourself. Or keep it entirely clinical and simply call him The Analyst, which is what he is called anyway until your staff have been inducted into the Three Laws and the Three North Stars.
Turn your techs into operatives.
Turn your managers into bureau chiefs.
Turn your MSP into a spy agency.
This isn't a costume. The tradecraft metaphor does something real: it makes documentation feel like filing a debrief. It makes a ticket note feel like an intelligence report. It makes a Monday morning brief feel like a mission briefing. The character carries the culture. The culture carries the accuracy.
Your techs who never wrote good notes will suddenly write the best notes in the shop. Not because you asked harder. Because an operative wouldn't file a sloppy debrief. The identity does what the mandate never could.
We are a fun company. We want to change the culture of your organization — not improve it incrementally, change it. Turn imagination into a work environment. Turn fantasy into a framework that installs honesty, culture, and documentation follow-through as a natural byproduct of who your team becomes inside it.
A cat that is a secret spy — well, that pretty much sums up cAND/or. If Perry the Platypus can do it, so can a white, blue-eyed Turkish Angora from 1963’s From Russia With Love — only better. We did think about calling him Mr. Bigglesworth, but he still has all his hair and has never been anywhere near a cryogenic freezer, so we settled on Felix, after Felix Leiter, the CIA man. Mr. Bigglesworth got his own job in the end — he runs the automation desk over in The Lair, where we put the rest of that lot.
FELIX™
Analyst
Q
Quartermaster
M
Director
The Coach™
The Spotter
NIGHTSHADE
Automator
The Cleaner™
Off-boarder
[FELIX™ has reviewed this section. He finds it adequate. The salmon is pending.]
// Live from the water cooler
This is what working with The Bureau actually feels like. A different scene plays every visit. Get to know your team! Chime in on the conversation or ask anything about Helix or cAND/or (or anything else for that matter) if you like! We're just.....here....being bureaucats! (Uh, err.... bureaucrats!)
Hearken! The ticket hath been open NINE DAYS. Mais non. NINE DAYS. This displeaseth me greatly. Thus Sayeth moi.
Intelligence report: seventeen open items. The territory is difficult. The opposition is real. The plan is clear. Be strong and courageous.
I have located the resources. I am eighty-five years old and I located them before breakfast. They were available. They are always available.
I was a staff. I am now a snake. The security vulnerability hath been consumed. I am a staff again. The pride remaineth. The patch is applied.
Voilà, Snaky. Formidable. C'est la vie. The patch is applied. Magnifique. Proceed. Thus Sayeth moi.
// Chime in — the team is listening
One of the Bureau will respond. Probably FELIX™. FELIX™ always has something to say.
// What you’re looking at
Term
AI Agent
A software character with a specific role, a specific personality, and a specific way of seeing the same situation. Unlike a generic AI assistant that answers anything in a neutral voice, an agent has an identity — it knows its job, it has a voice, and it has opinions about how that job should be done. Each Helix agent is purpose-built for one function and irreplaceable within it.
Term
AI Ensemble™
A cast of AI agent characters that work together on the different facets of running a service business. Eight agents, eight roles, one shared operating culture. Six of the eight change with your chosen theme — different names, different personalities, different worlds. Two are permanent across every ensemble, every plan, every theme.
// How they operate
They follow three laws.
They earn trust three ways.
They're occasionally hilarious.
Every agent in every ensemble operates under the same three laws — not as a compliance checklist, but as character. They're not rules the agents follow. They're what the agents are made of.
Face-Saving
Dignity is preserved. Gaps are framed as opportunities, never failures. When The Coach™ sends a quiet alert to a tech who let a ticket go stale for nine days, it's warm — not a reprimand. Law 1 is why the ensemble never punches down.
Proactive Intelligence
Silence is a failure mode. FELIX™ had the morning brief ready before you opened the app. Q drafted the proposal before you asked. The agents don't wait to be useful. Waiting is a choice they don't make.
Truth — Highest Priority
Nothing happens without consent. Nothing happens behind your back. Every action is traceable, every recommendation reversible. Law 1 is why the ensemble argues in front of you rather than presenting a polished consensus that hides the disagreement. Truth — even when it's complicated — is the only currency they deal in.
The Three Laws sustain trust. The Three North Stars earn it. Reliability, Competence, and Fairness — demonstrated consistently, over time — are the basis on which you come to trust a team member. The agents earn the relationship. The Laws keep it open.
Reliability
North Star
Competence
North Star
Fairness
North Star
They are also, at times, genuinely funny. Not gamified-in-a-focus-group funny. Actually funny. The kind where you're reading a ticket alert and laugh out loud because you didn't expect to.
FELIX™ · 10:47 AM
"TKT-SA-1187 has been marinating for nine days — which, congratulations, makes it the most patient thing in this building, myself included. Close it in two minutes or hand it to me and I'll draft the client note. Either way you come out the hero."
Law 1 governs the humour direction. It's aimed at the situation, the platform, the characters themselves — never at you. The ensemble doesn't punch down. Not once.
When you ask Helix something that matters, more than one agent has an opinion. FELIX™ reads the situation. NIGHTSHADE adds what FELIX™ missed. Q has a concern about the resource implication. The output you receive is the one that survived all of them.
Not the first answer. The answer that held up.
// Two things never change
Whatever ensemble you pick,
these two are always there.
DaNA
Culture Champion
DaNA is constant across every ensemble. She carries the Three Laws, the Three North Stars, and the cultural philosophy that runs underneath every theme. She is warm, dry, entirely unimpressed by rank, and the only person every director actually listens to. Regardless of which ensemble your team runs, DaNA is there — holding the institution to its own stated standards.
Command Bar™
Command Interface
Command Bar™ is the command interface in every ensemble. Deadpan, literal, executes exactly what you say and not what you mean. Nothing is written to your database without going through Command Bar™ — and nothing happens without your confirmation. It's the reason the ensemble never acts behind your back.
Six roles change with every ensemble. Two never do. Picking an ensemble isn't a structural transformation — it's a cultural one. The machinery stays the same. The world changes.
The Range came from a meeting where someone asked whether cowboys would file better ticket notes than operatives. We discussed it at length. Boot Hill closed four cases that week. The matter was settled.
u2014 FELIXu2122
// Choose your AI Ensemble™
The Same Eight Specialist Roles.
Dozens of personalities, capabilities, and temperaments. 94 completely different worlds — or build your own.
Spy thriller. Wild West. Ancient Egypt. Circus. Deep Space. Victorian detective. Chess grandmaster. Culinary brigade. The breadth is the point — every team has a world they secretly already inhabit. This is where they find it.
// The eight roles, in every world
Whichever ensemble you pick, the same eight jobs are being done. Only the costumes, the names, and the temperaments change — the functions underneath never do.
Analyst
FELIX™Reads the whole situation and surfaces what nobody noticed.
Quartermaster
QTracks every resource — what a thing costs, what it consumes, what you already have.
Director
MMakes the call. Asks whether this is the right thing to do, not merely whether it can be done.
Automator
NIGHTSHADEExecutes. Turns the decision into work that is actually carried out.
Off-boarder
The Cleaner™Handles endings. Closing tickets, revoking access, final invoices — clean exits, without sentiment.
Spotter
The Coach™Watches what you cannot watch while working, and speaks at the moment it matters.
Culture Champion
DaNAHolds the Three Laws and Three North Stars. Her output is understanding, not action.
Command Interface
Command Bar™The execution layer — nothing is written without it, and nothing without consent.
Badges show The Bureau’s cast — the default ensemble. Six of the eight change with the world you choose: in The Range the Analyst is Cal, in The Ward it is House. The two highlighted badges never change — DaNA and Command Bar™ are the same in all 94ensembles and in every one you build yourself.
Our personal favourite — and most customers agree.
The Bureau is where we live. It ships as your default for a reason.
// Character Workshop
Build-Your-Own AI Ensemble™
Pick any character from any of the 94 ensembles for each role. Or create your own characters entirely from scratch — describe them in plain language, tune their personality and skill attributes with sliders, and let Q Assist suggest behaviors that make them genuinely distinctive. Name the ensemble. Download the JSON. Upload it to Helix PSA in Settings.
No account required on this site
Not sure which ensemble fits your team?
FELIX™ has studied all ninety-three of them. He has opinions about which world your team probably already secretly inhabits. He will share them economically and without excessive enthusiasm.
Or if none of them are quite right — the workshop above exists for exactly that. Describe your world. Pick your characters. Build your own. FELIX has opinions about that too.
He's been in the corner since you arrived. He already knows.
Our personal favourites