In active development
Playbook is a second-screen agent assist tool. AI-powered, multi-tenant, and deployed in a browser tab alongside any telephony system.
No integration. No IT project. No surveillance.
What it does
Everything the floor needs, in one tab.
- Call logging, with outcomes defined per queue
- A callback calendar the agent owns
- Policy search across your own documents
- Scripts and starter prompts for the hard calls
- Shift summaries and handover notes
Why it holdsOutcomes belong to a queue rather than the company, so an agent picks from the short list their queue actually uses instead of hunting through one dropdown built for everybody.
- Support Bot, for the question mid-call
- Call Scribe, for writing the call up afterwards
- Objection Coach, for the push-back
- Answers grounded in your uploaded documents
- Per-tenant spend caps and usage reporting
Why it holdsEvery call is costed against the tenant at its own model’s rate, and an unrecognised model is refused rather than billed at a guess — so an AI budget cannot be quietly overspent by a model nobody priced.
- Live team activity and queue status
- Call log review with coaching notes
- Whisper coaching, during the call
- Abusive-call alerts
- Mood as a team aggregate, never a person
Why it holdsA whisper is never written down. There is no insert anywhere in that path, and the audit trail records that coaching happened while deliberately omitting what was said.
- Mood check-ins the agent controls
- Wellbeing signals surfaced as trends, not names
- XP, levels and missions
- Leaderboards, arcade and a coin shop
- Announcements and team recognition
Why it holdsA mood aggregate is withheld until at least three people stand behind it, so a team trend can never be narrowed back to the one person having a bad week.
- Vulnerability actions logged at the point of the call
- Disclosure statements set per queue
- Outcome and vulnerability reporting, aggregated
- Audited CSV and print-ready evidence export
- In every tier, not a compliance add-on
Why it holdsA figure is withheld unless five distinct agents stand behind it — and a suppressed row keeps its place in the table rather than vanishing, because even the position of a hidden row would leak the volume behind it.
- Brands, queues, teams and agents
- Documents, categories and knowledge posts
- Feature flags per tenant, and per agent
- Analytics, seat utilisation and call log export
- Subscription and seat management
Why it holdsThe audit log is append-only by construction: the code that could edit or delete one of its rows does not exist outside the tenant-erasure path, and a commit hook blocks anyone adding it.
What’s in the box
A complete contact centre platform.
Playbook ships with three distinct operational surfaces, each built for a specific role in the room.
Agent UI
Scripts, RAG-grounded answers, AI tools, callback management, mood and wellness self-reporting, and training games. Everything an agent needs, in one browser tab.
Manager UI
Team performance at a glance. Abusive-call alerts. Ephemeral whisper coaching that disappears after the call. Anonymised mood aggregates: team trends only, never individual.
Admin UI
Tenant configuration, user and team management, brand customisation, document upload, feature flags, and AI usage caps. Full control without a development team.
What Playbook isn’t
Missing.
By design.
Every line below is a choice, not a missing feature.
Not a telephony system
No softphone. No dialler. No IVR. Playbook sits next to your existing telephony in a browser tab. It does not replace it.
Not a CRM
No customer database, no lead pipelines, no opportunity tracking. Your CRM stays your CRM.
Not call recording
No call audio. No transcripts. No live call data of any kind. Playbook never touches the call itself.
Not a monitoring tool
No keystroke logging. No AI scoring of agents. No individual mood reports surfaced to managers. Playbook helps agents do better work. It does not police them.
Built for
Regulated.
Agent-heavy.
Sector-agnostic by architecture. Deployed first in financial services, expanding outward from there.
Financial services
FCA-regulated firms: insurance, lending, investment platforms, challenger banks. Consumer Duty alignment is the wedge, covering abusive-call detection, vulnerability flagging, agent override authority, and anti-surveillance posture.
Utilities & telecoms
High-volume, regulated, complaints-sensitive. Same Consumer Duty pressure as financial services, same retention-driven buying motion.
UK public sector
Local government, NHS trusts, central government agencies. Public procurement listing in progress. Anti-surveillance and data-residency posture map directly to public-sector procurement criteria.
Healthcare-adjacent
Private healthcare, dental groups, optical chains. NHS handled through a separate gated lane with additional clinical-data safeguards.
Mid-market retail
Customer service teams across DTC brands and retail chains. Seasonal spikes handled by per-agent pricing without forced commitment.
Pricing
Transparent tiers.
No surprises.
Every tier starts with a conversation — your trial is provisioned on the call.
Tier 1 Starter
Up to 10 seats, flat rate. No IT project. Live in a day.
Tier 2 Growth
Per-agent pricing, no feature gates. Agents assisted, never monitored.
Tier 3 Professional
RAG-grounded AI with cited sources. Multi-brand, unlimited queues.
Tier 4 Enterprise
Procurement-grade. Negotiated SLAs, principal support, EU West residency.
Founding programme
Be one of ten.
I’m selecting up to 10 founding customers across the first 18 months. Founding customers receive 30% off list price for the first 12 months in exchange for case-study rights, logo use, and quarterly feedback calls.
Selection is at my discretion, based on case-study marketability, vertical diversity, and reference willingness. Available on Tier 2 and above. If you want one of the ten, make your case.
Playbook is built by Apexify — a London creative studio and software house. The same expertise is available for client software and brand engagements.
Before you ask
The questions buyers actually open with.
No. Playbook opens in a browser tab beside whatever you already use. There is no telephony code in the product at all — no dialler, no softphone, no recording pipeline — so there is nothing to integrate and no vendor whose permission you need.
No, and it is built so it cannot quietly become it. No call audio, no transcripts, no keystroke logging and no AI scoring of individuals. Mood data is aggregated, and an aggregate is withheld until at least three people have contributed to it, so a team trend can never be narrowed back to one person. The regulatory evidence export holds a stricter line again, at five.
A day. There is no integration, no IT project and nothing to install — agents sign in through a browser. Configuration of brands, queues, teams and scripts happens in the admin surface rather than through us.
Vulnerability logging, per-queue disclosure statements and an audited evidence export are in every tier rather than sold as a compliance add-on. The export is designed to be handed to a regulator as it comes out.
Yes. Erasing a tenant is a built-in operation rather than a manual database job, and the audit trail it leaves behind is append-only by construction — the code that would let anything edit or delete those records does not exist.