What it can do
Things that took weeks, done in a few clicks.
Each of these was set up by the person who owns the process, on the live platform, and filmed from start to finish. No developer, no ticket in the IT queue.
- Send every customer their own letter, not a mail merge
- Turn a bad survey score into a recovery the same day
- Build a complex, multi-owner process with no code
- Serve customers inside their own AI assistant
- Interrogate your data
- Reach every customer before their contract ends
- Resolve billing enquiries in a few clicks
- Let customers change their address in chat
- Build a knowledgebase with no effort
- Make sense of thousands of journeys
- Know the moment a problem starts
- Coach every agent on every conversation
Send every customer their own letter, not a mail merge
Their own figures and tables, with whole paragraphs only where they apply.
Personalised mailings are harder than they look. A typical mail merge drops a name and a number into one fixed letter, so anything that really differs by customer turns into another version, another spreadsheet and more checking.
Here each letter is written from the customer's record: their own rent, the new figure worked out for them, and a table of their own payments. Whole paragraphs appear only where they apply. A tenant who owes rent is offered a plan, a tenant who asked for extra support is told someone will call, and a tenant who needs neither gets neither. Each one comes out as a finished PDF and stays on the customer's record.
It is one workflow in the no-code designer, and a campaign runs it for everyone in a segment. The paragraphs are written in an ordinary editor, with the customer's details dropped in.
Filmed on a demo landlord, so the homes, the people and the figures are made up. Email sending is off for this demo, so the film shows the letters as files rather than in anyone's inbox.
60 seconds, captions in the picture.
Turn a bad survey score into a recovery the same day
Sent from a case or a campaign, with low scores followed up from your console.
A survey score is only useful while you can still do something about it. Here the survey runs on the same platform as the service, so a low score is not a number in a report: it is a real case, and everyone who gave one is listed as soon as they answer.
An agent can send it from the case in two clicks, or any workflow or campaign can send it as a step. The customer gets a private link by email and answers with no login: ratings, multiple choice, grids, best-and-worst trade-offs and free text, all built with no code, and every answer lands on their case.
In the console, one click turns the low scorers into a campaign, and running it starts a follow-up journey for each of them.
Filmed on a demo council, so the people and the figures are made up.
70 seconds, captions in the picture.
Build a complex, multi-owner process with no code
Chat, forms, photos and task allocation across tenant, landlord and contractor, built in minutes.
A damp and mould report is a hard process to automate. It starts with a tenant in chat, where forms in the conversation find them and capture the problem, and the answers set the priority and allocate the job to the right contractor.
From there it has three owners, each on their own screen. The landlord sees it on a statutory clock. The contractor's dispatcher allocates an engineer in their own portal. The engineer works through the required checks and uploads before and after photos, then it goes back to the landlord to review and close.
All of it is one workflow in the no-code designer, built in minutes.
Filmed on a demo landlord, so the home, the people and the figures are made up.
105 seconds, captions in the picture.
Serve customers inside their own AI assistant
Claude shops, troubleshoots and runs a journey for your customer, using your Syntalyze assistant.
More customers now ask an AI assistant, such as ChatGPT or Claude, before they visit your website. Switch on your Syntalyze endpoint and the address can be added to their assistant as a connector. From then on it uses your answers, your products, your fixes and your journeys, not a guess from the open web.
Here a customer asks Claude for headphones to swim with and gets the one pair TechStore sells that works in the pool, with a link to buy. A phone that has slowed down is diagnosed with TechStore's own troubleshooter. And "Where is my repair?" runs TechStore's repair journey, identity check included, through to the status and what happens next.
How AI agents connect to your assistant
81 seconds, captions in the picture.
Interrogate your data
Ask your customer conversations anything, in plain English.
Every conversation your customers have with you is read in, so you can ask it the questions you would otherwise commission a report for. "What are customers dissatisfied with?" comes back with the reasons, drawn from what customers actually said.
Then drill into one: "What do customers dislike about the routers we supply?" No dashboard to build, no analyst, no SQL. Ask the next question as soon as you think of it, and the answers are kept so you can compare and re-run them.
43 seconds, captions in the picture.
Reach every customer before their contract ends
A segment, a campaign and a daily schedule, in a few clicks.
Describe the audience in plain English, "customers whose contract ends in exactly 30 days", and the AI builds the customer segment. Attach it to a campaign and set the campaign to run every morning. Each run emails the customers whose contract ends 30 days from that day, so each customer gets the email once, a month before their contract ends.
The same few clicks work for tips a month after joining, retention offers, new services or chasing late payers. No list to export, no mail-merge, nobody remembering to run it on a Monday.
92 seconds, captions in the picture.
Resolve billing enquiries in a few clicks
A bill-explaining journey, built from a plain-English brief, then live in chat.
The brief: verify the customer, fetch this month's and last month's bills from the billing system, and have AI explain the difference in a few short sentences. The AI assistant builds the journey and it is published.
A customer asking "Why is my bill higher this month?" then gets their own answer: both totals, the discount that ended and the roaming pass that was added. No code, no ticket in the IT queue, and no call.
90 seconds, captions in the picture.
Let customers change their address in chat
A change-of-address journey, built from a plain-English brief, then live.
The brief: check who the customer is, ask for the new address, send it to the customer system and confirm it back. The AI assistant builds the journey and it is published.
A customer who says "I've moved house" is verified, gives the new address and sees it confirmed. One of the most common reasons to call, handled in chat without anyone writing code.
91 seconds, captions in the picture.
Build a knowledgebase with no effort
Upload your store list, and customers can ask for the nearest one.
A spreadsheet of stores goes in: addresses, opening hours, phone numbers, services, parking. Seconds later a customer asks "Where is my nearest store to Kilkenny? Is there parking?" and gets the store, the address and the car park charges, from that sheet.
No mapping project, no developer, no new page on the website. Change the spreadsheet and the answers change with it.
A store list is one example. It could be your returns policy in a Word document, your products from your website, your product sheets in a PDF, or your HR rules for internal use.
39 seconds, captions in the picture.
Make sense of thousands of journeys
Chats, calls and transcripts on one map: volume, handle time, cost and mood.
Conversations from the Syntalyze chatbot, other chatbots, call recordings and transcripts, all read into one map. Where customers came from on the left, how it ended on the right, and the topics discussed on the way in between.
Click one topic, here billing enquiries, and only those journeys remain. Then see the same map by handle time, by cost to serve, and by mood: green where customers came out happier, red where they did not. It is the quickest way we know to see where the money and the goodwill go.
61 seconds, captions in the picture.
Know the moment a problem starts
Set a threshold on a risky topic. Pass it, and the alert and the email go out.
Mark a topic as a risk, here "Internet connection", and set an alert threshold. When conversations about it pass the threshold, the analytics page flags it and an email goes to the person who needs to know.
You hear about the outage from your own customers' conversations, before the complaints reach social media.
54 seconds, captions in the picture.
Coach every agent on every conversation
Every conversation checked against your script and your outcomes.
Every conversation is checked against your script, the phrases agents must say and the ones they must never say, and against the outcomes your team counts as success: issue resolved, customer happy, or a refund, a hand-off, a cancellation. Every conversation, not a sample.
Team leaders go straight to the conversations worth a coaching chat, with the reason next to each one.
It is the same for AI chatbots as for human agents: the chatbot is held to the same script and the same outcomes, and shows up as one more agent on the team.
56 seconds, captions in the picture.
Try it on your own customers.
Start with a 24-hour advisor this afternoon, and add the journeys, campaigns and reports when you are ready. Fourteen days free.