ACE · AUTONOMOUS CONVERSION ENGINE
Give ACE a goal.
Convert more leads.
ACE is an agentic conversion engine that plans, orchestrates and learns from customer interactions across channels — continuously optimizing toward your business goal.
Running in production with Hero Fincorp, PatangStory and Insta Astro — decisioning and delivery on live customer traffic.
Goal
Increase qualified conversions
+25%Opportunity Finder
standby
Strategy Planner
standby
Execution Dispatcher
standby
Events
Intelligence
Next best action ↺
Goal progress
81%
From campaign software to agentic execution
Campaign tools wait for instructions.
ACE works toward a goal.
The campaign is no longer the unit of work. The goal is.
Traditional campaign software
- 01Build audience
- 02Choose template
- 03Choose channel
- 04Schedule
- 05Send
- 06Look at report
Every step is a human decision. Then it stops.
ACE — a continuous loop
- Define Goal
- Understand Opportunity
- Plan
- Act
- Observe
- Evaluate
- Replan
- Act Again
ACE keeps working until the outcome moves.
Two products, three ways to buy
Take the half you need.
Or both.
ONTACE decides. COMACE executes. Neither requires the other, and most clients start with one.
ONTACEthe ontology layer
Reads your data as it is — across systems, with keys that do not match and records that disagree — and builds one graph per customer. The disagreement is an input, not a blocker.
COMACEthe goal engine
Take a goal — recover this book, renew these subscribers, convert these leads — and it works the goal until the number moves. Anyone who pays, disputes or opts out leaves mid-flight.
ONTACE alone
You keep your entire communications stack. ONTACE reads your data, decides, and hands the decision back for your own systems to act on.
What runs
- A customer graph built from your own records, however they are keyed
- Around twenty cited facts and ten beliefs per customer
- One decision, with the evidence attached
- Predictions that return insufficient_evidence rather than guess
What you keep
- Your martech, your CRM, your journeys
- Your vendor contracts and your rates
- How and when you send — unchanged
What you need
- A warehouse or source systems we can read
- Roughly a few thousand contactable customers
- Someone who owns a number the decisions should move
ONTACE · deciding with evidence
The system knows what it doesn't know.
And it tells you.
Any platform can return a number. ACE returns the number, the evidence under it, and the rung of the ladder it had to stand on to get there — or it refuses to guess. Pick a situation and watch the decision change.
The situation
A lender six months in, with payments integrated and outcomes attributing. The objective's own conversion metric has evidence, so nothing further down the ladder is needed.
Worked examples · the behaviour is the product
Metric ladder
most outcome-proximate first
- Payment rate (objective's conversion metric)4,812 matured outcomesused×1.0
- Reply rate12,004 sends×0.5
- Read rate9,340 WhatsApp sends×0.3
- Delivery rate12,004 sends×0.1
Rung 1 · payment rate
WhatsApp, Tuesday 18:00–21:00
Chosen on the outcome the goal is actually measured by, not on a proxy for it.
Confidence 0.87 — undiscounted
No uplift column exists
Recommendations that were followed and recommendations that were not are reported side by side and never subtracted. The groups were not randomised, so the difference is not an uplift — and a schema test asserts the column has never been added.
Confidence is sample adequacy
Not a p-value. Calling it one would claim an experiment that did not happen. A channel chosen on delivery rate stays a weaker claim than one chosen on payment rate, however many deliveries were observed.
Deterministic and reproducible
The same outcome renders to the same string a year later, so two episodes embedded months apart stay comparable. Decisions can be replayed and audited rather than re-asked.
COMACE · ACE Live
Watch an autonomous system operate.
A single opportunity, end to end: detection, strategy, policy checks, execution, events and the business outcome — all in one journey context.
ACE Live
streaming · demo dataGoal
● ActiveIncrease qualified conversions
Target
25%
Current
18.7%
▲ 4.8% this week
Opportunities
0
Actions
0
Conversions
0
Simulated demo data
Agentic architecture · PACE
An architecture designed to act, not just automate.
PACE is ACE's agentic core: three cooperating agents that find opportunities, plan engagement strategy and execute it — inside the policies you define.
Opportunity Finder
Where is the highest-value opportunity right now?
- Customer likely to convert
- Payment nearing default
- Abandoned lead
- Customer eligible for renewal
- Inactive subscriber
- Interested but unconverted prospect
Strategic Scheduler
What is the right move — and when?
- When to engage
- Which channel to choose
- What sequence to follow
- Which template strategy to use
- How long to wait
- Whether another attempt is appropriate
Executive Dispatcher
Turn decisions into executable actions.
- Coordinates WhatsApp, SMS, RCS
- Email and voice delivery
- Future channels
- Through the provider-agnostic Nudging Gateway
The autonomous loop
Every interaction changes what ACE does next.
ACE can use delivery, engagement and business outcome signals to maintain context around a journey and determine the next appropriate action.
COMACE · define a goal
You don't create a campaign.
You create a goal.
Describe the outcome, the audience, the constraints and the channels. ACE's agentic architecture determines how to work toward it.
- 01
GOAL
The measurable outcome your business needs.
- 02
TARGET / DATA
The customers or records the goal applies to.
- 03
CONSTRAINTS
Budget, deadline, contact limits, quiet hours.
- 04
AVAILABLE CHANNELS
WhatsApp, SMS, RCS, email, voice — what ACE may use.
- 05
BUSINESS RULES
Compliance, escalation and preference policies ACE must respect.
New goal
Illustrative interfaceWhat outcome do you want?
Target
Upload dataset / connect source
Success event
lead.converted
Target
+20%
Deadline
30 days
Template intelligence
Stop guessing which message works.
ACE compares how templates actually perform — per cohort, per channel, per time window — and feeds the winner back into the next decision.
Template A
- Delivered
- 94.2%
- Clicked
- 8.4%
- Converted
- 3.2%
Template B
Winner- Delivered
- 93.8%
- Clicked
- 14.7%
- Converted
- 7.8%
ACE recommendation
Use Template B for Cohort 18–24, between 6–9 PM, through WhatsApp.
Illustrative example of ACE's template-intelligence direction. Demo data.
Cross-channel orchestration
One customer. One journey. Every channel.
Channels aren't isolated silos. ACE runs a single state-aware journey per customer — a failure on one channel triggers the fallback, an engagement on any channel stops the rest.
Journey state · demo
- WhatsAppFAILED
- Fallback decisionSMS SELECTED
- SMSDELIVERED
- Tracked linkCLICKED
- Remaining nudgesSTOPPED
- OutcomeCONVERTED
State-aware execution,
not parallel blasts.
Failure-driven fallback
A failed WhatsApp delivery immediately re-routes the message through the next best channel.
Engagement-aware stops
A click or reply anywhere in the journey halts the remaining scheduled nudges.
Journey memory
Every attempt, event and outcome stays attached to one journey context per customer.
Provider-agnostic by design
Your strategy shouldn't depend on your provider.
Every decision ACE makes flows through the Nudging Gateway — a provider-agnostic execution layer. Change providers without rewriting your campaign and decision logic.
Solutions by industry
The same engine. A different argument every time.
What changes between industries is not the technology — it is what counts as evidence, and what a good decision looks like. Pick the one you are in.
Lending & collections
An EMI is twelve days overdue. The existing system fires generic reminder #4 at 9am, in English, to a borrower who has never once answered a morning SMS.
1What ACE reads
- Missed 2 of 12 EMIs — both in months the salary credit landed late
- Kept 3 of 3 past payment promises
- Reads WhatsApp within two hours; has never opened a morning SMS
2What it concludes
Intends to pay, and is constrained on cash flow early in the month. Pressure would cost a cooperative relationship that is currently worth keeping.
3What it sends
WhatsApp, in Hindi, a respectful reminder with a payment link, at 7:40pm. Timing wins the outcome here — the approved template stands unchanged.
What moves
Recovery rate ↑Contact cost ↓Audit trail per messageMeasuredAt a leading NBFC: ₹5 Cr due, ₹3.5 Cr collected in a three-day run — 70% recovered against 46% the month before. Live in seven days.
Worked examples. The lending result is measured; the others are built to the same pattern rather than drawn from one deployment. If your outcome is measurable, it fits this shape — insurance, education, healthcare, travel, real estate and B2B pipelines all run the same loop.
Autonomy with control
Autonomy doesn't mean losing control.
You choose how much ACE decides on its own — and every level runs inside the guardrails your business defines.
Always enforced
Whatever the autonomy level, ACE never steps outside the policies you configure — and a human can override, pause or stop any goal at any time.
Architecture
Built like infrastructure, because it is.
ACE sits one layer above communication infrastructure: your data and business systems feed it, intelligence and decisioning run inside it, execution flows out through providers — and every event flows back in.
- DATA
- CAMPAIGN / GOAL LAYER
- PACE
- NUDGING GATEWAY
- CHANNEL PROVIDERS
- CUSTOMERS
- DLR + ENGAGEMENT + OUTCOME EVENTS
- EVENT PIPELINE
- INTELLIGENCE
The event loop is the architecture: delivery receipts, engagement signals and business outcomes are first-class events that feed the intelligence layer — which feeds the next decision PACE makes.
Developers · Built to connect
POST /goals
{
"goal": "lead_conversion",
"target": 0.25,
"success_event": "lead.converted"
}
→ { "goal_id": "goal_12873",
"status": "active" }POST /events
{
"type": "lead.converted",
"entity_id": "lead_8392"
}
→ { "accepted": true }Conceptual examples of ACE's API-first design — request access for current API documentation.
Security & governance
Agentic doesn't mean uncontrolled.
ACE is built for enterprises that need autonomy and accountability at the same time.
Tenant isolation
Every tenant's data, templates and provider credentials stay strictly separated.
Encryption
Data protected in transit and at rest.
Auditability
Decisions, sends and outcomes leave a traceable record.
Policy enforcement
Contact rules, quiet hours and business constraints are enforced in the execution path.
Role-based access
Operators, approvers and admins see exactly what their role allows.
Rate controls
Throughput and frequency limits at every layer.
Consent & contact rules
Opt-outs and suppressions are honored across all channels.
Provider credential isolation
Channel provider keys are stored and scoped per tenant.
Outcome traceability
Every conversion links back to the journey that produced it.
Human override
Pause, stop or take over any goal at any time.
What outcome are you
trying to achieve?
>
Get connected
Tell us the outcome.
We will show you the engine.
A short form, a real reply. We will come back with what ACE would do for your book — the channels, the sequence, and what it would measure.
What happens next
- 1A person reads itNot a router. Usually within a working day.
- 2We map your outcomeWhat ACE would target, on which channels, measured how.
- 3You see it runningA walkthrough on real data, not slides.