Conversational characters for messaging

Your customers already have the app.
It's called Messages.

A personalised character that lives inside iMessage, WhatsApp and Telegram. It talks like a person, remembers what matters, and messages first, without anyone opening anything new.

Kingly Clark·August 2026

The reality

Nobody wants another app.
Everyone answers a text.

The most-used screen on any phone is the messaging screen. Every brand that asks people to download, log in and check a dashboard is competing with that screen and losing. Meanwhile the bots that do reach people there are menus with a greeting.

98%

of WhatsApp messages get opened, against roughly 21% for email

Infobip, 2026

88%

are read within five minutes of arriving

Infobip, 2026

68%

say a bot that doesn't understand them is the main reason for a bad experience

Verint

35%

choose an AI agent specifically so they never have to repeat themselves

Salesforce

The idea

Not a bot. A character who texts.

We build a named personality with a voice, opinions and a memory, and drop it into the chat your customer already uses. No buttons to press, no flows to click through. You text it like you'd text anyone, and it always texts back.

Lives where they chat

No new app to check

iMessage, WhatsApp, Telegram, SMS. One-to-one or inside the group chat that already exists.

Talks like a mate

Human, not helpful-desk

Short messages, a point of view, jokes that land. Designed around how people actually text, not how systems answer.

Remembers you

Tell it once

Every conversation teaches it something. Favourites, dislikes, timing, tone. It keeps learning and adapts to each person.

Proof, live today

Meet Bud.

A sports concierge we built and run. Bud is a gorilla who knows more about sports than your drunk uncle, and he lives in iMessage. He learns how you watch, remembers who you support, and pings you only for the moments worth looking up for.

iMessage, liveWhatsApp, next950+ soccer leagues, NFL, NBA, cricket, rugby, F1

Text Bud, budknowsball.com

Bud, a gorilla in a sage Nike Golf half-zip, crouched against a sage backdrop

Bud knows ball

Hear him talk

This is how Bud texts.

Read it and notice what isn't there. No "How can I help you today?". No menu. He has opinions, he knows when the game matters, and he tells you to go be with your family when it doesn't.

A real exchange, in Bud's voice, from budknowsball.com.

Bud

iMessage · active now

Thu 6:42 PM

new phone, who dis. kidding. who we supporting?

lol ok. warriors fan. don't make it weird.

Thu 7:31 PM

heads up, our guys tip at 8. three point dogs at home, which is disrespectful. i'll keep an eye on it.

Sat 4:18 PM

not gonna lie to you, log off, go be with the fam. i'll text you if it gets interesting. it won't.

you're the best bud

Sun 10:47 PM

that's a wrap, you won. box score says it was close, it wasn't, our defense was elite in the 4th. sleep well.

Why it feels human

We studied how people actually text.

The difference between a bot and a mate isn't intelligence. It's behaviour. Every character we build follows the unwritten rules of a group chat.

Restraint

Text once, then shut up

One nudge when it matters. No double-texting, no "just checking in". Silence is part of the personality.

Timing

Knows when to talk

Quiet hours, time zones, the fact that nobody wants a breakdown at 1am. It waits until you're ready.

Memory

Never makes you repeat yourself

Preferences stick from the first mention. Ask for something once and it becomes a habit.

Voice

Has opinions

Real people take sides, use lowercase, and get a bit sarcastic. A character with no edge reads as a form.

Initiative

Messages first

Heads-ups before, a nudge during, a wrap after. It starts conversations instead of waiting to be prompted.

Honesty

Says when it isn't worth it

"Log off, it won't get interesting." Trust comes from the messages it chooses not to send.

Self-learning

It gets better every week it knows you.

The platform builds a living memory of each person: what they follow, what they ignore, when they read, how they like to be spoken to. Nothing is configured up front. It learns from the conversation and quietly adapts, so a soccer fan never hears about basketball and the Sunday early game arrives without being asked.

Bud

iMessage

Tue 10:14 AM

sunday early game. locked in.

got it. early window only. nfl. no overtime drama unless it's our team.

Learned

Sunday early window · NFL only · quiet after 11pm · no basketball · recaps, not video · lowercase, dry humour

Listen

You say it once

Preferences, moods, and throwaway comments all count.

Remember

It writes it down

A private memory per person, updated after every chat.

Adapt

Next message changes

Tone, timing, topics and frequency shift to fit you.

Repeat

Week after week

The longer it knows you, the less it needs to ask.

What it can do

Everything a good mate does in the chat.

Answers

Replies to anything

Ask a question in plain language and get a real answer, with live data pulled in when it needs it. News, scores, stock, bookings, whatever your world runs on.

Triggers

Messages on its own

Scheduled updates, event-driven alerts, threshold nudges. "Tell me when it ships", "every Friday", "if the price drops".

Group chats

Joins the group

Sits inside an existing group, follows the thread, chips in when tagged or when it has something worth saying.

Broadcasts

Talks to thousands, one at a time

Send to a segment and every person gets a version written for them, in the character's voice, at the right time.

Follow-ups

Remembers to come back

Reminders, check-ins, "how did it go?" the day after. It carries the thread so your team doesn't have to.

Hand-off

Knows when to get a human

When a real person is needed it passes the full context across, so nobody repeats themselves.

iMessageWhatsAppTelegramSMSMessengerInstagram DMWeb chat

Where else it works

Bud does sport. Yours does whatever your customers care about.

The character changes, the platform underneath doesn't. A few we'd happily build tomorrow.

Retail

The personal shopper

Texts when the thing you wanted is back, remembers your size, never sends a catalogue.

Hospitality

The concierge

Knows you like the corner table, sorts the late checkout, suggests the one bar worth going to.

Fitness

The training partner

Checks in after a missed session, adjusts the plan, celebrates a PB without a leaderboard.

Banking and fintech

The money mate

A plain-English nudge when spending runs hot, a heads-up before a debit order, no dashboards.

Events and clubs

The fixture secretary

Sits in the club group chat, posts the team sheet, chases RSVPs, knows who always bails.

Media

The editor who texts

Three stories you'll actually read, at the time you actually read, in your own interests.

Property

The agent's sidekick

New listings that match what you said you wanted, viewings booked by text, follow-up done.

Loyalty and rewards

The insider

Points, perks and drops delivered as a tip-off from a friend, not a promo blast.

Internal teams

The ops buddy

Lives in the staff WhatsApp group: shifts, stock alerts, "who's on tonight", answered instantly.

How it compares

Why this isn't another WhatsApp bot.

	Menu-driven bot	Generic AI chatbot	Kingly Clark character
Lives inside iMessage, WhatsApp, Telegram	WhatsApp only, usually	Mostly a website widget	Yes, and inside group chats
Messages you first	Broadcast blasts	Waits to be asked	Proactive, personal, timed
Remembers you between conversations	No	Sometimes, within a session	Persistent memory per person
Keeps learning and adapting	No	Rarely	Self-learning, every conversation
Sounds like a person	Buttons and menus	Polite, generic, verbose	A named character with a voice
Knows when not to message	No	No	Quiet hours, restraint, one nudge
Hands off to a human with context	Sometimes	Sometimes	Yes

How we build it

Character first. Then the plumbing.

We don't start with the technology, we start with who this person is. The rest is our job, and you never have to see it.

Phase one

The character

Name, voice, opinions, boundaries. What it talks about and what it never mentions. Written and tested in real conversations with your team.

Phase two

Knowledge and triggers

We connect it to what it needs to know and decide what it should message about on its own, and when.

Phase three

Pilot with real people

A small group on iMessage or WhatsApp. We watch how they actually use it, and the character learns from day one.

Phase four

Roll out and keep learning

Open it up, add channels, and let the memory compound. We stay on to tune the voice as it grows.

Pricing

Pay per conversation, not per seat.

Simple usage pricing per thousand messages. The more your character talks, the cheaper each message gets. Drag the slider to see what a month looks like at your volume.

Messages per month

100,000

10k100k500k1m5m

Estimated monthly

$3,000

Blended rate per 1,000

$30.00

Volume band	Per 1,000 messages
First 100,000	$30
100,000 to 500,000	$28
500,000 to 2,000,000	$26
Beyond 2,000,000	$25

Bands apply progressively within a month, so the first 100,000 messages are always $30 per thousand and only the messages above each threshold move to the lower rate. One message is one text sent or received. Character design, integration and pilot are scoped separately. Channel carrier fees, where a platform charges them, are passed through at cost.

Next step

Text Bud tonight.
Then let's talk about yours.

The fastest way to understand this is to have a conversation with it. Join Bud's waitlist, watch how he behaves over a weekend of fixtures, and we'll take it from there.

budknowsball.com

Kingly Clark · kinglyclark.com · August 2026