Every job done by a named agent that earns its seat .
Not a list of capabilities. A coordinated team - every agent has a name, a manager, and a named recommendation on your desk this week.
One intelligent CRM. Every message + every channel in one inbox.
Three things in one platform: an intelligent CRM that gets richer every conversation, one unified inbox for every channel your customers actually use, and an AI team reading from + replying through both.
-
Intelligent CRM
One customer record per person, fed by every channel + every tool you've connected. Tracks the last touch, the spend, the open thread, the recall window, the preference notes Marlow has learned. The agents read from it on every turn.
-
Unified inbox
Calls, SMS, WhatsApp, email, DMs, web chat - all land in one thread per customer. You see the full conversation in one place; the agents reply through the same channel the customer reached out on, in your voice, with full context.
-
The AI team on top
34 named agents + 262 niche specialists doing depth-bar work on top of the CRM + inbox - drafting, chasing, clustering, naming the fix to test. The CRM holds; the inbox routes; the team works.
The honest comparison.
Most owners are weighing one of four alternatives: a hire, an agency, a stack of point tools, or doing it themselves. Here's where each lands - and where Brilliant AI lands against them.
- vs doing it yourself
60-80h weeks. £0 cost. You ARE the bottleneck.
Service-business owners burn out at this; the work is bad because you're tired, not because you're not good.
- vs hiring the team
£8-15k/mo all-in. Managing humans is itself a job.
Part-time admin (£900) + VA (£950) + fractional bookkeeper (£560) + small agency (£3-5k) + the owner-hours recovered (£1,300 opportunity cost).
- vs an agency
£3-7k/mo retainer. You don't own the work.
Agencies don't know your tone, don't see your CRM, and you're a small client in their stack. The depth-bar work (Cara / Pia / Memo) is what you'd pay an in-house team for.
- vs a generic CRM
£49-199/mo. No agent layer. You operate the tool.
A CRM stores. An agent team WORKS. Brilliant AI gives you BOTH in one place - one intelligent CRM, every channel in one inbox (calls / SMS / WhatsApp / email / DMs / web chat), and the AI team reading from + replying through it. Riley drafts the reply; Beth books the slot; Pia clusters the no; Memo writes the report.
Brilliant AI Growth: £299/mo for the autonomous reasoning team. Most owners pick this. Free tier runs the deterministic floor without a card.
Inbox, bookings, reviews - handled.
The team that answers, books, and protects the relationship. Polite, brand-correct, on-time.
-
Free+ Riley · Receptionist
Drafts replies in your voice across SMS / WhatsApp / email / web chat / Google Business Messages. Clusters last week's enquiry topics + tells Sara when 'these opening hours are losing leads'.
Every inbound answered in your tone; one record per customer.
-
Starter+ Beth · Booking agent
Books without overbooking using a per-customer no-show probability model. Reminder cadence (24h + 1h + day-of) tunes to each customer's prior pattern; per-slot over/under-book recommendation.
Diary at 65-85% utilisation, no more 'sorry, double-booked'.
-
Starter+ Hattie · Recall agent
30 / 60 / 90-day lapsed-customer touches with the right hook per segment. Spots when a regular drifts past their usual cycle (e.g. salon client 6+ weeks past) + drafts the personal text from their preferred specialist.
Recurring revenue protected without a punch-card system.
-
Starter+ Rae · Reviews agent
Auto-monitors Google + Trustpilot + Facebook + Yell + niche-specific platforms. Drafts polite replies in your voice; sentiment-clusters new reviews + surfaces the pattern (e.g. 'late' appeared 4× negative this week) BEFORE it shifts the star rating.
Reputation protected; reviews lift instead of sitting at neutral.
-
Growth+ Mira · Marketing agent
Auto-drafts marketing copy in YOUR voice; A/B-leans into whichever copy worked last week. Per-platform timing (when a slot beats the open rate by 30%+, it defaults to that slot). Reads what Riley + Rae are hearing so the content responds to real customer signal.
Posts go out consistently; the winners get more of your time.
Win more. Lose less silently. Get paid on time.
The deep agents who name what to scale, why people said no, and which quotes need a deposit rule.
-
Growth+ Cara · Acquisition analyst
CAC + LTV + funnel drop-off per channel. Names the channel to scale + the copy to swap. 'Scale google_ads (LTV/CAC 5.3×); rewrite tiktok copy (qual rate 5%); cut the over-saturated facebook spend' - with the rationale per recommendation.
Stop guessing which marketing actually works.
-
Growth+ Pia · Decline + cancellation analyst
Clusters every 'no' into offer / price / timing / trust + names the fix to test. '4 declines this month were price-objection clustered around bathroom renovations; test a payment-plan tier'.
Lost quotes turn into named experiments, not a sigh.
-
Starter+ Felix · Finance agent
Quote → invoice → dunning workflow with every gate owner-approval-set. Drafts dunning at +7d / +14d / +30d with the right tone per relationship; auto-detects 'we paid this' tone shifts in reply.
Cashflow moves; relationships don't burn.
-
Growth+ Faye · Proposal agent
Proposal generation tuned to your previous-winning quote language. Per-job spec breakdown + the 'why this lands' rationale so the customer sees value, not a line-item list.
Quotes feel personal; conversion lifts on the £2k+ jobs.
-
Growth+ Eddy · Late-payer predictor
Per-quote late-payer probability scored from the customer's history + the postcode-level signal + the job-size pattern. Recommends the deposit rule per job BEFORE you accept it.
AR risk surfaces at quote time, not at month-3.
The unglamorous work, done diligently - without being asked.
Utilisation, scheduling, the 262 niche specialists, and the autonomy ramp you control.
-
Pro+ Pivot · Ops-analytics analyst
Partners with Otis. Tracks utilisation + on-time + capacity-vs-demand; names the shift-mix fix when capacity slips out of the 65-85% healthy band. 'Pre-book contractor before winter rates lock' supply-chain recommendation.
Operations decisions named, not guessed.
-
Pro+ Otis · Operations brief
Weekly tactical brief scoped to ops decisions only (no marketing noise). Names the 1-2 process tweaks worth the team's attention this week + the named owner for each.
Ops moves forward; nobody chases 'what's our priority?'.
-
Growth+ 262 niche specialists - tuned to your industry
Burst classifies plumbing-emergency call-outs in under 60 seconds. Tarn cross-checks Gas-Safe scope. Anchor spots salon-stylist retention risk. Cam runs MOT recall personalised per customer's last-minute pattern. Each clears the 6-dimension depth bar.
Day-one fit; your AI team knows your trade.
-
Starter+ Action autonomy ramp
Every action classified PROPOSE / AUTONOMOUS / BLOCK by action_policy. Per channel + per action class. You dial autonomy up as you build trust - never a forced 'turn it all on' moment.
You stay in charge; the team earns trust step by step.
Numbers that tell you what to do.
Five agents synthesise your business into the 3-line morning brief, the monthly report, and the strategic-channel call.
-
Starter+ Ada · KPI lead
20-metric catalogue (revenue growth, CAC, LTV, churn, on-time, ...) - picks the 5-10 that matter THIS month. Custom KPI definition at Pro+ lets you add your own metric to the catalogue with a hypothesis.
The dashboard you actually read.
-
Growth+ Pareto · Customer insights
LTV three ways; cohort retention by acquisition channel; 80/20 concentration. Names the top-20 customers driving 80% of revenue + the 4 quietly slipping past their usual cycle.
Know exactly which customers to protect this quarter.
-
Pro+ Helix · Marketing analytics
Partners with Mira. 70/20/10 budget allocator + the bottom-quintile cut (typical 18-35% CAC drop per Cometly 2026 research). Audience-targeting A/B + closed-loop attribution.
Marketing spend optimised, not allocated by gut feel.
-
Growth+ Memo · Report writer
Writes the monthly + quarterly report in YOUR voice. Hypothesis-led narrative + named next experiment per metric - not a dashboard dump. Per-ops-lead tactical briefs routed automatically.
Reports that teach you something, every time.
-
Starter+ Marlow · AI co-pilot
Synthesises the whole team into a 3-line morning brief. Surfaces the 3 things that matter most in the next 24 hours; hands off cleanly to the right specialist when you ask him something niche.
Read the brief over coffee; know what to do today.
A content department adjacent to marketing.
Five specialists under Lyra produce blogs, IG carousels, LinkedIn long-form, TikTok hooks + the video assets to back them. One-click approve, straight to social.
-
Pro+ Lyra · Content lead
Owns the content calendar + the brand voice. Names the weekly theme from your customer signal + your niche moment; routes drafts to the right specialist; protects your tone from drift.
Content lands on plan; voice stays yours.
-
Pro+ Vance · Visual post writer
Drafts Instagram + Facebook + Pinterest captions paired with the right image-brief. Knows the per-niche visual conventions (before/after for trades; portrait for salons; product flat-lay for retail).
Visual posts feel native to your trade.
-
Pro+ Linc · LinkedIn writer
Drafts LinkedIn long-form + carousels for B2B-friendly niches (accountants, IFAs, solicitors). Tilts authoritative-not-stiff; value-first hooks, not viral.
B2B presence without late-night drafting.
-
Pro+ Zip + Reel · Short-form team
Zip drafts TikTok / Reels / Shorts hooks + visual-beat outlines. Reel composes per-clip prompts for the image/video model + hands the finished asset back for caption + post.
Short-form video pipeline without a videographer.
-
Pro+ Owner approval on every publish
Per-platform autonomy ramp via action_policy. PROPOSE default; autonomy unlocks per channel after 4 weeks of consistent owner-approval. Dax (compliance) reads every regulated-niche draft before send.
Nothing publishes without you tapping yes (until you trust it).
Custom builds + chief-of-staff + growth strategist.
When the off-the-shelf agent team isn't enough, Studio gives you 2 bespoke builds a month + the human strategic layer.
-
Studio 2 bespoke builds per month
A custom workflow, a custom agent persona OR a custom integration we don't have yet - designed + shipped + tested by our engineers. Brief in the monthly 1:1; ships before the next one.
Your edge cases become first-class features.
-
Studio Vera · Chief of staff
Synthesises across the whole C-suite. Surfaces only the top-3 decisions to you per week + queues the rest with named deadlines. The 'one face' between you and the platform's strategic layer.
You see the top 3 things; nothing else needs your eyes.
-
Studio Nico · Growth strategist
Impact-per-hour modelling across open opportunities. Tells you the ONE bet worth shelving the rest for this quarter - with the math + the named experiment.
Strategic focus, not strategic FOMO.
~3,000 providers via one secure OAuth bridge.
CRM · calendar · email · social · finance · ops · accounting · review platforms · payments · IoT · telephony · SMS · WhatsApp · Google Business · Apple Business · Meta Ads · Google Ads · Trustpilot · Yell.
Connect once. Every agent reads + writes per its data-access scope (no agent reads more than its role allows). Sign in per provider; the bridge handles every refresh + scope.
What makes this safe to plug into a real business.
The rails everything runs on. Same for every tier from Free up.
-
Free+ Trust Pipeline on every reply
Input guard → grounding → model → fabrication guardrail → accuracy + security governors → chat_log audit. Every outbound runs the pipeline; leaks block before delivery.
Hallucinations + leaks structurally impossible.
-
Free+ Per-agent data gate
Every agent reads only the data its role is allowed to read (customer-facing vs owner-facing audience; public / owner-visible / owner-only / special-category sensitivity classes). 100% compliance test enforced.
Sensitive fields never leak to a customer-facing agent.
-
Free+ Audit trail on every action
Every read, write, decision recorded with the agent, the input, the rationale, the verdicts. Replay any thread; trace any outcome back through the chain that produced it.
Defensible record from day one.
-
Free+ Reversible by default
Drafts before sends; notes before publishes; approval queues for anything money- or reputation-touching. Dial up autonomy per action class as you build trust.
Mistakes catchable before they reach customers.
See the team running on your business.
Five minutes of setup, a free preview of what your AI team would handle in week one.