Skip to content
OppositeNumber

AI & digital product studio · Guildford, Surrey

We are your Opposite Number.

Guiding you through the age of AI to a more streamlined, cost-effective and powerful business. We find where AI pays back, prove it in a fixed-price sprint, then build and run the software behind it. No hype, no 40-slide decks. Working systems, measured results.

Trusted by

  • Homeconcepts
  • 1minus1
  • Thunderbolt
  • WBK
  • Homeconcepts
  • 1minus1
  • Thunderbolt
  • WBK
  • Homeconcepts
  • 1minus1
  • Thunderbolt
  • WBK
  • Homeconcepts
  • 1minus1
  • Thunderbolt
  • WBK
  • 0

    from idea to a working AI prototype

  • 0

    median eval score on shipped AI features (placeholder)

  • 0

    surprise invoices. Fixed prices, published upfront

Our impact in numbers

Numbers we are happy to be held to.

These are the figures we report against on every engagement. Replace the placeholders with your own once the first projects land.

See how we work

Showreel

The work, and the people behind it.

Video placeholder

Why us

Three promises most agencies will not put in writing.

01

Fixed prices, published upfront

Our audits and sprints are fixed price and our engagement models and rate bands are on the website. You know the number before the first call.

02

Working software in 10 days

Every AI engagement starts with a 10-Day Sprint that ends in a working prototype on your data and a go or no-go decision. Half the time of the usual 20-day sprint.

Framing workshopDay 1Eval set on your dataDays 2–4Working prototypeDays 5–8Go / no-go packDay 10
03

Security built in, not bolted on

Your data is your business, so we treat it that way. Least-privilege access, encryption in transit and at rest, no client data used to train models, prompt-injection and data-leak testing before every launch, and a full audit trail you can inspect. Controls aligned to ISO 27001 and Cyber Essentials Plus, written into every statement of work.

How it works

From your data to your systems, with a score at every step.

This is the shape of every AI system we ship. Evaluation before the model, guardrails before your users, monitoring forever.

Your dataCRM · docs · emailEvalsground truth setModel layerbest fit by scoreGuardrailspolicy · PII · auditYour systemsERP · portal · Slackmonitoring · evals re-run on every model update

What it looks like in production

Real numbers, not a demo.

The cards on the right are the kind of dashboard every client gets: evaluation scores, hours returned, cost per task and a live audit trail. If the number does not move, we tell you.

Eval score

0%212 cases

Every release is scored against your ground truth before it ships.

Hours returned to the team

0

+58% this quarter

Cost per proposal

Before£42.00
After£13.10

Automation with a human approval step, measured end to end.

Live activity

  • Invoice #4471 classified → Approvals
  • Policy check passed · 0 PII
  • Proposal drafted for Acme (2m 14s)
  • Model update detected · 212 evals re-run
  • Eval score 96.1% · no regressions

Delivered

Things we have shipped.

Drag to explore

All case studies

What we do

New processes that cut cost. New products that drive growth.

AI Services

Put AI to work in the business you already run.

  • AI Strategy. Find where AI pays back, prove it fast, and own the plan at leadership level.
  • AI Development. Build AI into your products and processes with evals, guardrails and observability.
  • AI Adoption. Make it stick with training, governance, security and visibility in AI search.
Explore AI services

Digital Services

Design and build the products your customers come back for.

  • Strategy. Decide what to build and de-risk it before the big spend.
  • Build. Apps, platforms and websites engineered to last.
  • Design. Design systems and digital brands that scale with you.
Explore digital services

Plugs into what you own

One governed AI layer. Every system you already run.

We do not replace your stack. We connect it. Policies, evals and an audit log sit between your tools and any model, so you can swap vendors without rebuilding.

  • CRM
  • ERP
  • Email & docs
  • Data warehouse
  • Slack & Teams
  • Customer portal
  • Vendor-neutral

    Anthropic, OpenAI, Google or open-weight models, chosen by evaluation score and swapped without touching your systems.

  • Governed by default

    Policy checks, PII redaction and a complete audit trail on every call. Aligned to ISO 42001.

  • Measured forever

    Evals re-run automatically when a provider updates a model, so quality never silently drops.

Sectors

Built for your industry, not a generic template.

The same engineering rigour, shaped around the regulation, data and customers your sector actually has.

  • Financial Services
  • Healthcare & Life Sciences
  • Retail & Consumer
  • Media & Entertainment
  • Professional Services
  • Property & Construction
All sectors

Kind words

What clients say.

“Placeholder: a short quote from a client about the outcome we delivered and how it felt to work with us.”
Client name
Job title, Company
“Placeholder: a second quote, ideally with a number in it. Revenue, hours saved, time to launch.”
Client name
Job title, Company
“Placeholder: a third quote focused on the team, the process and ongoing support.”
Client name
Job title, Company

FAQs

Straight answers.

How quickly can we start, and how long does a project take?

We can usually start a 10-Day AI Sprint or an Opportunity Audit within two weeks. A well-scoped MVP, app or website typically takes eight to sixteen weeks. Larger platforms run longer, and many clients keep a retained team afterwards.

Do you publish prices?

Yes. Our sprints and audits are fixed price, and our engagement models page shows day-rate bands for retained teams. You will never get a surprise invoice from us.

What technologies do you use?

The right ones for the job. For the web we favour Astro, Nuxt and Next with a headless CMS. For AI we work across Anthropic, OpenAI, Google and open-weight models, selected by evaluation rather than fashion. For apps we build natively or cross-platform depending on the product.

Who owns the IP and the code?

You do. Everything we build is delivered into your repositories and your cloud accounts, with documentation your own team or another supplier could pick up.

Can you support us after launch?

Yes. We offer monthly support retainers for security, updates and small changes, and retained product teams for ongoing roadmap delivery. We also help clients hire and hand over to in-house teams when that is the right move.

How do you handle data security and AI risk?

Every AI engagement includes a data and security review, evaluation against your success metrics, and guardrails against prompt injection and data leakage. We align governance work to ISO 42001 and can work inside your existing ISO 27001 controls.

Talk to us about AI and digital

A 30-minute call. No deck, no pitch. Bring a problem and we will tell you honestly whether AI, software or neither is the answer.