AI for your business

We set up AI that answers customers, runs repetitive work, creates content, gets you found in AI search and turns your data into decisions. It works inside the tools you already use, and it is measured like everything else we do.

Get a free audit

Why now

AI has moved from experiments to everyday tools. Businesses use it to work faster, and customers use it to find and compare businesses.

The evidence is clearest in customer service. In a large study, support agents with an AI assistant resolved 14% more issues per hour, and newer staff improved by 34%1. McKinsey estimates current technology could in principle automate activities that take up 60 to 70% of employees' time2. In principle is the important phrase. The value comes from choosing the right work.

Customers have moved too. 73% prefer messaging when they contact a business3, and 74% expect service around the clock4. An AI agent that answers at 2 am, in the customer's language, closes that gap.

Nearly 9 in 10

organizations now use AI regularly in at least one business function.

From McKinsey, The State of AI in 2026

14%

more customer issues resolved per hour by support agents using an AI assistant. Newer staff improved by 34%.

From Brynjolfsson, Li and Raymond, Generative AI at Work, NBER

73%

of consumers prefer messaging when they contact a business. 72% are more likely to buy from brands that offer it.

From Kantar for Meta, Business Messaging Usage Research, 2025

Five ways we put AI to work

AI chat agents

Assistants on your website and WhatsApp that answer questions, take bookings and qualify leads, day and night.

Plays the Advancement ace

  • Answer common questions from your own information
  • Take bookings and enquiries, then pass them to your team
  • Ask the right questions to qualify a lead
  • Hand the conversation to a person when needed
Example
A clinic gets questions at night about prices and opening hours. The agent answers from the clinic's own information, offers open time slots and sends the booking to the front desk.
What to watch
AI agents can sound confident when they are wrong. We limit them to your own information, set rules for what they must not answer, and hand difficult conversations to a person. Every agent says it is an AI.
Built with
Website chat and WhatsApp Business, connected to your booking system or CRM.

Automation

Workflows that move information between your booking system, CRM, email and spreadsheets, so nobody copies it by hand.

Plays the Advancement ace

  • Send new leads to your CRM and team chat automatically
  • Reply to every enquiry within minutes
  • Create invoices, reminders and reports on a schedule
  • Connect tools that do not talk to each other
Example
A new website enquiry is added to the CRM, the sales team gets a message and the customer receives a reply within minutes, with nobody touching it.
What to watch
Automation repeats mistakes as fast as it repeats good work. We start with one workflow, test it with real cases and add alerts for when something fails.
Built with
n8n, Zapier and the connections of the tools you already use.

AI content and creative

Copy, images and short video produced faster with AI, then checked and finished by people.

Plays the Advancement ace

  • Product and page descriptions at scale
  • More ad variations to test more ideas
  • Image and video edits for each platform
  • First drafts of translations, checked before publishing
Example
An online shop with 800 products gets clear, search-friendly descriptions for every product, each one reviewed before it goes live.
What to watch
Unreviewed AI content tends to sound generic and can include errors. Every piece is edited and checked by a person, in your brand voice.
Built with
AI tools for text, image and video, with a person reviewing every piece.

AI search (GEO)

Getting your business described correctly and recommended in ChatGPT, Gemini, Perplexity and Google's AI Overviews.

Plays the Advancement ace

  • Check what AI tools say about you today
  • Structure your website so AI tools can read it
  • Build mentions on sites AI tools trust
  • Track your visibility every month
Example
A hotel wants to know why ChatGPT recommends a competitor. We find the gaps, fix its website and listings, and track how often it is mentioned each month.
What to watch
No one controls what an AI tool says. We improve the information these tools rely on and measure the result, without promising mentions.
Built with
Structured data, an llms.txt file, content and reputation work.

Data insights and dashboards

Your sales, ads, website and booking numbers in one dashboard, with clear answers on what to do next.

Plays the Advancement ace

  • Combine data from your website, ads, shop and bookings
  • One dashboard the whole team understands
  • Alerts when something changes
  • Monthly insights written in plain language
Example
An owner sees revenue, ad spend, bookings and cost per customer on one screen every morning, instead of logging into five tools.
What to watch
A dashboard is only as good as the data behind it. We check tracking first, so the numbers can be trusted.
Built with
Connected data sources and a live dashboard you can open on any device.

How the work is done

Picture the owner of a small hotel answering the same five questions on WhatsApp every evening. That is usually where the work begins, with one task that eats time and follows rules a person could explain. OpenAI's own guide to building agents says to start small, test with real users and grow from there7. Gartner predicts that over 40% of agentic AI projects will be cancelled by the end of 2027 because of rising costs, unclear value or weak risk controls8, so we write down today's numbers before anything is built.

Next comes the knowledge the AI will work from. Your price list, booking rules, menus and the answers your best staff already give become its reference, which is how OpenAI suggests building these routines from existing procedures and scripts7. We agree who owns the system and which customer data it may touch, in line with Indonesia's Personal Data Protection Law of 20229. WhatsApp has strict rules here, as it does not allow general-purpose AI chatbots, only AI that supports a business's own service, and WhatsApp conversations may not be used to train AI models10.

Before a single customer sees it, the agent is tested against real questions, the easy ones and the awkward ones, and its answers are scored to set a baseline7. People are wary, and in a 2024 Gartner survey 64% of customers said they would prefer companies not to use AI in customer service, mostly out of fear that reaching a person gets harder11. So every agent says it is an AI, which 72% of customers in a Salesforce study say matters to them12, and it hands the conversation to a person whenever that is the safer choice13.

After launch we read the conversations it got wrong, fix the source information and report the numbers to you every month. On its first day, Lightspeed's AI agent handled 35 to 40% of conversations without a team member, and with steady tuning that later rose to upwards of 60%14. AI search takes the same patience. Google says no special AI files or markup are needed to appear in its AI features15, and McKinsey found that a brand's own websites make up only 5 to 10% of the sources AI search cites16, so clear pages, reviews, listings and press coverage all count.

How long it usually takes

  • An AI assistant at Telkomsel, IndonesiaDeployed in 2 weeks
  • Testing with real questionsBefore any customer sees it
  • Tuning after launchOngoing, reviewed monthly
  • A typical small business projectNo reliable survey yet

From Microsoft Customer Stories, Telkomsel customer support teams speak volumes with help from copilot based on Azure OpenAI Service, 2024

It has worked for others

These are other companies and their published results, shared by the companies themselves or by the software vendors they used. They are not TWACE's results, and vendor stories tend to show the best cases.

  • 45%

    of customer interactions became self-service, up from 19%, after the telco moved its Veronika assistant to Azure OpenAI, and daily calls to agents fell from 8,000 to 1,000.6

    TelkomselIndonesia

  • 15%

    higher average occupancy across 20 hotels, credited to a whole system of automated daily booking data and Power BI dashboards that one developer built in two weeks.17

    Laris HotelsPoland

  • 65%

    of support volume resolved by the AI agent at its best, with a range of 45 to 65% across teams, while customer satisfaction stayed stable.14

    LightspeedGlobal

  • 12,000

    hours a year saved by this fitness studio chain with an AI agent on chat and phone, though customers rated the AI 80% against 93% for human staff.18

    solidcoreUnited States

What it usually costs

The AI services themselves are sold by usage, in US dollars. Intercom's Fin starts from $0.99 per outcome23, Zendesk charges $1.50 for each committed automated resolution or $2.00 pay as you go24, and Salesforce lists Agentforce at $2 per conversation25. These are published price lists, and they change often.

Automation tools cost less. Zapier's Professional plan starts from $19.99 a month26, and n8n's cloud plans start from 20 euros a month billed annually, with a self-hosted version available on GitHub27. On WhatsApp, Meta has charged per message since 1 July 2025, but replies to a customer inside an open service window are free, while marketing messages are charged28.

We found no reliable survey of what agencies charge small businesses to set up AI, in Indonesia or anywhere else, so any price you hear is one firm's quote. As for results, Intercom says from its own customer figures that Fin's resolution rate averages 76%29, while a named customer such as Lightspeed reports 45 to 65%14. Across the whole United States workforce, the time saved by generative AI equals just 1.6% of all work hours30, a reminder that the gains come from a few well chosen tasks.

28%

of businesses in Indonesia, about 18 million, have adopted AI, according to AWS research from 2025.19

22%

of smaller organizations are scaling AI agents, flat for a year while large ones rose to 40%.20

86.4%

of marketing teams worldwide use AI in some areas, and content creation is the most common use.21

8%

of Google visits with an AI summary led to a click on a regular result, against 15% without one, in a 2025 US study.22

How we bring AI into your business

  1. 01

    Find the work

    We look for the tasks that take the most time or lose the most customers.

  2. 02

    Start small

    We set up one AI process first, with a clear target.

  3. 03

    Measure

    We track time saved, leads answered and sales made, with a person reviewing the output.

  4. 04

    Scale

    When it works, we add the next process.

Read the guide on where AI saves time in a small business

Clear rules for AI

  1. 01

    You own it

    Accounts, data and workflows are set up in your name.

  2. 02

    People stay in charge

    A person reviews what AI sends to customers until you are confident in it.

  3. 03

    Honest by default

    Every agent says it is an AI, and we tell you which AI services handle your data and why.

  4. 04

    Measured results

    Every AI process has a target, and you see the numbers every month.

Questions

Will AI replace my staff?

In most businesses AI takes over repetitive work, such as answering the same questions or copying data between tools. Your team spends that time on customers and on work that needs a person.

Do I need a lot of data to start?

No. Chat agents and automation can start with the information you already have, such as your prices, menus, policies and booking system.

Is my customer data safe?

We set up AI tools in accounts you own, share only the data a process needs, and tell you which services handle it.

Can AI work in other languages?

Yes. Chat agents and content can work in any language your customers use, and switch language to match each customer.

How much does it cost?

AI projects start from IDR 10 million. The running cost of AI services depends on how much they are used. You get a clear estimate before we start.

Sources

  1. Brynjolfsson, Li and Raymond, Generative AI at Work, NBER
  2. McKinsey, The economic potential of generative AI, 2023
  3. Kantar for Meta, Business Messaging Usage Research, 2025
  4. Zendesk CX Trends 2026
  5. McKinsey, The State of AI in 2026
  6. Microsoft Customer Stories, Telkomsel customer support teams speak volumes with help from copilot based on Azure OpenAI Service, 2024
  7. OpenAI, A practical guide to building agents, 2025
  8. Gartner, Gartner Predicts Over 40% of Agentic AI Projects Will Be Canceled by End of 2027, 2025
  9. BPK RI, JDIH, UU Nomor 27 Tahun 2022 tentang Pelindungan Data Pribadi, 2022
  10. WhatsApp, WhatsApp Business Solution Terms, 2026
  11. Gartner, Gartner Survey Finds 64% of Customers Would Prefer That Companies Didn't Use AI for Customer Service, 2024
  12. Salesforce, State of the AI Connected Customer, 2026
  13. Intercom Help Center, Fin explained
  14. Intercom, How Lightspeed achieves up to 65% resolution rate with Fin
  15. Google Search Central, AI features and your website, 2025
  16. McKinsey and Company, New front door to the internet, Winning in the age of AI search, 2025
  17. Microsoft Customer Stories, Laris Hotels unlocks legacy data with Power Automate, 2024
  18. Intercom, AI at every touchpoint, How solidcore scaled personalized support across channels
  19. Amazon Web Services, New AWS research shows strong AI adoption momentum in Indonesia, 2025
  20. McKinsey and Company, The state of AI in 2026, On the road to ROI, 2026
  21. HubSpot, 2026 state of marketing, Data from 1,500+ global marketers, 2026
  22. Pew Research Center, Google users are less likely to click on links when an AI summary appears in the results, 2025
  23. Intercom, Pricing
  24. Zendesk, Pricing
  25. Salesforce, Agentforce pricing, 2026
  26. Zapier, Pricing
  27. n8n, Pricing
  28. Meta for Developers, Pricing on the WhatsApp Business Platform
  29. Intercom, Fin
  30. Federal Reserve Bank of St. Louis, The State of Generative AI Adoption in 2025, 2025

Are both of your aces ready for what comes next? The free audit tells you.

Get a free audit