Pricing

AI agent pricing that starts with one workflow

Every engagement starts with a two-week, fixed-scope diagnostic on one request type. You see your own policy in the Rulebook and a sealed decision record before you commit, and production is priced on the scope you have already proven.

Clutch Top Chatbot Company, United States 2025Global Generative AI Award 2026 WinnerDeveloped by Sphere, named a Top Chatbot Company and a Top AI Company in the US by Clutch

Readiness diagnostic · week 2 of 2Ready
  1. 1
    Kickoff

    Request type and owner agreed. Fee waivers on chat and voice.

  2. 2
    Policy mappedfee-waiver-eligibility

    Your thresholds written as Rulebook rules.

  3. 3
    Rules tested and replayed

    Sample requests run through your rules. Sealed records ready for compliance.

  4. 4
    Decision

    Systems: core and CRM. Deployment: co-managed. Next step: production proposal.

The short answer

Priced on your scope, not on a price list

Palladium does not publish a price list. Each engagement is priced on three things: the request types you automate, the channels and systems involved, and where the platform runs.

Every engagement starts with a two-week readiness diagnostic at a fixed scope, with the fee agreed before kickoff. You know what you are buying before you buy it.

How AI agent pricing works

Five pricing models. Five ways to share the risk.

Conversational AI and AI agent vendors price in a handful of ways. Each one moves risk differently between you and the vendor. Here is how to read them as a bank.

ModelHow it worksWhat to watch for in a bank
Per seatA monthly fee for each employee who uses the tool. Fits internal assistants.Says nothing about customer volume, so it rarely fits customer-facing AI.
Per conversationA fee for each conversation the AI handles, resolved or not.You pay for conversations that end in a handoff. Ask how a conversation is counted across channels.
Per resolutionA fee only when the AI resolves the request without a person.Ties cost to outcomes. Check who defines “resolved” and whether escalations are excluded.
Usage-basedA platform fee plus model usage, often per token or per call.Hard to forecast. Model price changes and long conversations go straight to your budget.
Scoped engagementA fixed-scope pilot, then production priced on the request types and channels in scope.Easier to budget and to defend in procurement. This is how Palladium is priced.
What drives cost

Six things set the price of a deployment

01

Request types in scope

Each request type, such as Reg E disputes or fee waivers, brings its own rules, data and escalation paths.

02

Channels

Voice, chat and in-app, SMS and WhatsApp, and email each need their own setup and testing.

03

Systems of record

Core banking, CRM, ticketing, payment rails and identity providers the agents read from and act on.

04

Volume

How many requests each workflow handles, and how they peak during fraud waves or seasonal spikes.

05

Deployment

Hosted by Sphere, co-managed, or fully private in your VPC with private model options.

06

Who runs it

Sphere manages policy and operations, or your team writes the rules and Sphere operates the platform.

The pilot

A working session, not a sales process

The two-week diagnostic covers one request type. At the end you have evidence your compliance team can review and a scoped proposal you can take to procurement.

Week 1

Map the workflow

We walk the request type with its owner, map your policy thresholds to Rulebook rules, and review the channels and systems involved.

Week 2

Replay the decision

We run your rules on sample requests and produce a sealed decision record your compliance team can retrace step by step.

Then

Go live in weeks

If you continue, the diagnosed workflow goes into production on one or more channels, with Coach / QA review from day one.

After that

Scale by request type

Add request types, channels and teams on the same Rulebook and audit trail, or move to a private deployment.

Check if a workflow is ready

Engagement options

Same architecture. Three ways to run it.

What changes is who runs the platform and where it runs. The rules, agents and Ledger are the same in every option.

Fastest start

Managed by Sphere

Sphere runs the platform. You bring the policies and systems of record.

  • Fastest path to production
  • Fully managed guardrails and updates
  • Best for a first request type
Most common

Co-managed rollout

Your team owns policy authoring. Sphere owns platform operations and audit.

  • Direct control over rules and thresholds
  • Sphere-operated infrastructure
  • Scales across request types
Custom

Private deployment

Fully private in your VPC, with private model options and your own operating team.

  • Full data sovereignty
  • Bring-your-own-model support
  • Best for the strictest mandates
Managed by SphereCo-managed rolloutPrivate deployment
Who writes the rulesSphere, from your policiesYour teamYour team
Who operates the platformSphereSphereYour operating team
Where it runsHosted by Sphere, strict data isolationHosted by SphereYour VPC: AWS, Azure or Google Cloud
ModelsModel-agnosticModel-agnosticPrivate model options, bring your own
Audit trail and guardrailsIncludedIncludedIncluded
What to measure

The numbers that justify the spend, from a $2B regional bank

62%Of tier-1 volume automated. Fee-waiver and dispute questions moved off the call-center queue.
9 daysTo compliance sign-off on the audit trail.
7.2sAverage decision on a fully autonomous Reg E dispute.
100%Of interactions reviewed by the Coach / QA agent.

Agree these measures before the diagnostic starts: autonomous resolution rate, escalation rate, time to compliance sign-off, and handle time on the request type in scope.

Included every time

Controls come standard, not as add-ons

Every engagement, at every size, ships with the controls your compliance and vendor-risk teams ask for first.

Explore trust & security

Replayable audit trail

Every decision sealed in the Ledger and exportable for examiners as a digitally signed PDF.

Limits enforced in code

Dollar and risk thresholds live in the Rulebook, never in the model.

Screening in and out

Prompt-injection screening and PII redaction on every message, in both directions.

Coach / QA agent

Every interaction reviewed, with root causes and policy improvements surfaced.

SOC 2 Type II

Independently audited controls. No training on your customer data, by default and by contract.

Due-diligence kit

The answers your vendor-risk team will ask for, ready before the first call. Open the kit

FAQ

Pricing questions

How much does Palladium cost?

Palladium is priced per engagement, based on the request types in scope, the channels and systems involved, and the deployment model. Every engagement starts with a two-week readiness diagnostic at a fixed scope, with the fee agreed before kickoff.

Why is there no public price list?

Two banks automating the same request type can have very different volumes, channels and core systems. The diagnostic puts your scope in writing, so the production proposal is priced on your actual situation, not a generic tier.

What do we provide for the diagnostic?

The policy for one request type, an owner who can answer questions about it, and a contact in compliance who will review the sealed decision record at the end.

How long until we are in production?

The diagnostic takes two weeks. A first production workflow typically ships in weeks after that. A $2B regional bank had its audit trail signed off by compliance in 9 days.

Can we start hosted and move to our own cloud later?

Yes. The architecture is the same in every option, so rules and records carry over when you move from a managed or co-managed setup to a private deployment in your VPC.

Is pricing different for credit unions and community banks?

Pricing follows scope, not institution type. A community bank or credit union starting with one request type on one channel scopes a smaller engagement than a regional bank automating several request types across voice and chat. Palladium for banks and credit unions.

Developed by Sphere

Built by an award-winning AI engineering team

Palladium is a Sphere product. The engineers who design, run and support it are the same team Clutch ranks among the top AI and chatbot companies in the United States, and the team behind production systems for more than 300 clients over 21 years.

That is why Palladium ships with what regulated teams ask for first: rules in code, a record of every step, and people on every exception.

21 years in production engineering300+ clients28 countriesSOC 2 Type II

About Sphere

  • Clutch Top Chatbot Company, United States 2025
  • Clutch Top Artificial Intelligence Company, United States 2025
  • Global Generative AI Award 2026 Winner
  • Global AI Awards 2026, Best in AI for Documentation Validation
  • Clutch Top AI Text Generation Company, Florida 2025
  • Global Tech Award Winner 2026
  • Corporate Vision 2026, Most Innovative Technology Consulting Firm
  • Global Recognition Awards 2026

Bring one policy. We’ll replay it.

In a 30-minute discovery call we load one of your thresholds into the Rulebook and show you the sealed record it produces. Then we scope your diagnostic.