Our own product
Writford
A workspace for UK legal firms connecting cited research, matter records, time recording and billing. The research workflow keeps source checking and professional review with the solicitor.
Explore WritfordManchester software house. We build websites, web and mobile apps, MVPs and AI for UK businesses.
Chatbots that answer from your systems, not from a generic model — customer service, sales qualification, appointment booking and internal knowledge. Connected to your CRM, helpdesk and diary, with the handover to a person designed first.
Every chatbot is custom-built for your specific business — not a generic template.
Answer FAQs, handle returns and troubleshoot issues 24/7 without staff cost. Most businesses see routine tickets fall noticeably in the first month.
Qualify leads automatically. Ask the right questions, score prospects, and pass hot leads to your sales team.
Book appointments, send confirmations, and handle rescheduling — integrated with your calendar system.
Give your team instant answers from internal docs, policies, and procedures. No more digging through shared drives.
Every project is scoped and quoted after a free discovery call, so you get a fixed price for your situation rather than a number that assumes someone else’s. These are the shapes most engagements take.
A business answering the same handful of questions all day and wanting them off the inbox.
Typical timeline: 1–2 weeks
Anyone whose answers live in a system rather than an article: orders, bookings, account records.
Typical timeline: 3–6 weeks
Teams who need the bot to take actions across several systems, not just answer.
Typical timeline: 6–10 weeks
Timelines are typical and depend on how many systems are involved and how ready your data is. Tell us what you need and we will give you a fixed price and a date.
Would a chatbot pay for itself? Test it with your own enquiry volumes, staff time and customer value — no figures of ours pre-filled.
Open the ROI calculatorWe audit your current support flow, identify top questions, define the bot scope and integrations needed.
We connect the AI to your data, write the system prompt, and configure the conversation flows.
Deploy to your website, WhatsApp, or CRM. Connect handoff flows. Test with real users.
Review conversation logs, improve weak responses, expand knowledge. Ongoing improvement as your business grows.
A bot that only knows your FAQ page is a search box with a personality. The useful ones read from the systems where the answers actually live. What decides the scope is rarely our side of the work — it is what your system permits. Some expose everything through a documented API, some allow reading but not writing, and some sector-specific products need the vendor to approve the integration first. We check yours before quoting.
This is a typical integrated build. A simpler FAQ and handover bot compresses into the first two weeks; a multi-system agent runs longer. The column on the right is the part most proposals leave out: a chatbot project stalls on access and content far more often than on engineering.
| When | What we do | What we need from you |
|---|---|---|
| Week 1Scope and boundary | We audit your live enquiries, agree the questions the bot will own, and write the list of things it must never attempt. We check what your CRM or booking system actually permits before promising anything. | An export or sample of recent enquiries, a named decision-maker, and read access to the systems in scope. |
| Week 2Content and connections | We connect the bot to your documents and records so it answers from your material, build the integrations agreed in week one, and set up the handover route to a person. | Your FAQs, policies and product or service detail, plus whoever owns the systems for a short access conversation. |
| Week 3Testing against real conversations | We replay your real historic enquiries through the bot, tune the weak answers, and check the whole widget works by keyboard and screen reader before anyone sees it. | Two or three people to try to break it, and an hour to review the answers it gets wrong. |
| Week 4Launch and watch | We go live, usually to a share of traffic first, and read every conversation in the opening fortnight. Weak answers get fixed while you watch rather than after a quarterly review. | A decision on where escalations land, and someone to receive them. |
We design the handover to a person before the conversation flows, because that boundary is what makes everything else safe to automate. A bot that knows when to stop is worth more than one that always has an answer, and the most expensive failures we are asked to fix are almost always a bot that pressed on when it should have stopped.
The list is yours, not ours. We write a first draft from your sector and your policies, you correct it, and it becomes the specification the bot is built against and tested on.
Not legal advice, and your own advisers should confirm how each applies to you. But these are the questions that separate a supplier who has thought about UK deployment from one selling a widget.
Article 50 of the EU AI Act covers systems that interact directly with people and requires them to be built so users know they are dealing with an AI system. The European Commission confirms those transparency obligations apply from 2 August 2026, with a limited grace period only for the separate marking of generated content on systems already on the market. There is no equivalent UK statute, but if any of your customers are in the EU it reaches you.
You stay the data controller. Your privacy notice has to cover chatbot conversations, a lawful basis has to be recorded, and there should be a retention rule rather than an indefinite archive. The ICO updated its guidance on AI and data protection in April 2026 to reflect the Data (Use and Access) Act, and says it is revising its automated decision-making and profiling guidance under the same Act.
A large number of shipped widgets fail here. WCAG 2.2 requires every function to work by keyboard, so quick-reply buttons cannot be mouse-only. New messages need an ARIA live region or a screen reader user never learns a reply arrived. The input label must survive typing, and the focused element must not be hidden behind the widget.
Each sector page sets out the systems we connect to there, the rules the bot has to work under, and what it must hand to a person.
Not listed? We work across UK professional services, hospitality and SaaS too. Tell us about your business.
If any of your customers are in the EU, yes. Article 50 of the EU AI Act covers systems that interact directly with people and requires them to be designed so users are informed they are dealing with an AI system; the European Commission confirms those transparency obligations apply from 2 August 2026, with a limited grace period only for the separate marking of AI-generated content on systems placed on the market before that date. There is no equivalent UK statute, but we build the disclosure in by default because hiding it damages trust and costs nothing. Confirm your own exposure with your legal adviser.
You remain the data controller and the conversation data stays under your control, in the UK or EU, with a retention rule agreed before launch rather than kept indefinitely. Your privacy notice needs to cover chatbot conversations and you need a lawful basis recorded for the processing. The ICO's guidance on AI and data protection sets the expectations and was updated in April 2026 to reflect the Data (Use and Access) Act, and the ICO has said it is revising its automated decision-making and profiling guidance in light of the same Act, so this is an area to re-check rather than set and forget.
Ours will, and it is worth asking any supplier the same question, because a large number of shipped widgets fail here. WCAG 2.2 requires all functionality to be operable by keyboard, so quick-reply buttons inside bot messages cannot be mouse-only. New messages need to be announced through an ARIA live region — the log role carries a polite live setting for exactly this — or a screen reader user never learns a reply arrived. The input label must stay visible once typing starts, and the focused element must not be hidden behind the widget.
Two things, and the second matters more. The bot answers from your own documents, records and systems through retrieval rather than from a model's general training, so it has something specific to quote. Then it has an explicit boundary: a list of things it must never attempt — legal or clinical advice, a price commitment, a refund outside your policy, anything outside your written protocol — and a handover to a person on the route you choose. We design that boundary before the conversation flows, because it is what makes the rest safe.
That depends far more on what it is allowed to say than on the technology. The bots that feel robotic are the ones padding a non-answer, looping a menu, or refusing to admit they cannot help. Ours answer in your words, from your material, and are built to say “I do not know, here is a person” early rather than late. We test with your real historic enquiries before launch, so you hear how it handles the awkward ones rather than the demo ones — and if it reads badly on a question, we change the answer rather than the wording around it.
Most mainstream CRMs, helpdesks, booking systems and store platforms expose an interface we can use, and we check yours specifically before quoting rather than promising in advance. The honest constraint is not our side, it is what your system permits: some products expose everything through a documented API, some restrict writes, and some sector-specific systems need the vendor to approve the integration first. On the discovery call we tell you which category yours falls into, and if a booking write is not achievable we say so and start with the parts that are.
Everything keeps running. The accounts are set up in your name rather than ours, the code and the conversation data are yours on completion and full payment, and the handover is part of the engagement rather than a favour at the end of it. Worth asking any agency this before you sign, because it is not universal.
The honest range is one week to three months, and the difference is not effort — it is how many systems the build touches and whose approval it waits on. What decides which end you land on.
How UK recruitment agencies use AI to reply to every applicant and screen against qualifying questions, what decides the cost, and the ICO and Fair Work Agency rules that apply in 2026.
A booking chatbot is only as good as the diary it writes to and the rules it works under. What decides whether one works for your business, and what it must never book.
Answering customers on WhatsApp is free under Meta's rules; starting the conversation is not. What the platform actually charges for, what drives the build cost, and how to avoid paying for messages you did not need to send.
Review our own live products, check our company registrations and read the commitments that shape a project. These give you concrete work and clear terms to discuss before booking a call.
Our own product
A workspace for UK legal firms connecting cited research, matter records, time recording and billing. The research workflow keeps source checking and professional review with the solicitor.
Explore WritfordOur own product
Tools for connecting shopping assistants with storefront product, cart and checkout actions. Its WordPress and storefront integrations show the kind of connected web workflows we build.
Explore Let AI BrowseRead the LetAIBrowse case studyThirty minutes, free, no obligation and no salesperson. You describe what is costing you time or customers, we ask about the systems involved, and you leave with an honest view of whether this is worth doing. We regularly end that call by saying the tool you already pay for covers it.
The code, the repository and the accounts are set up in your company name rather than ours, and they are yours on completion and full payment. Nothing runs on infrastructure we control, so ending the relationship does not end the thing we built.
Every project is quoted as a fixed price with a written scope after the call, so an over-run is our problem rather than yours. If the scope changes, we say so and re-quote before doing the work, not after.
In the UK or EU, under a data processing agreement, kept only as long as you specify. Any model provider we rely on is named as a sub-processor rather than left implicit, and we will tell you plainly where something you have asked for would put a commitment of yours at risk.
Agreed before launch rather than negotiated during an incident: what is covered, how quickly we respond, and what counts as new work. Every build has a defined escalation to a person, because the failure that costs you is the one nobody noticed.
Big Berri Limited is registered in England and Wales, company number 16562429, and registered with the Information Commissioner's Office under reference ZC119995. Both are public registers. Look us up rather than taking our word for it.
Book a free 30-minute discovery call. We will scope the bot, tell you which of your systems it can realistically reach, and say plainly if an off-the-shelf tool would serve you better.
Book a Free Discovery Call