FAQ
Over the years we have collected the questions clients ask us most, from AI and pricing to how we run projects. Search below or browse by topic.
We design, build and run AI products end to end. That includes:
See AI development and AI agent development for more detail.
Yes, and it is often the fastest way to get value from AI. We start by reviewing your product, data and workflows to find where AI will genuinely save time or improve the experience. Most AI features connect to your existing system through APIs, so there is rarely any need to rebuild what already works.
A chatbot answers questions. An AI agent can also take action: it can look things up, fill in forms, update your CRM, draft replies or hand work to a person when needed.
Agents are worth it when a process is repetitive, rule-heavy and spread across several tools, for example triaging support tickets, qualifying leads or processing documents. If the task is simple, a plain automation is usually cheaper and we will tell you so.
We are model-agnostic and choose the model per use case, balancing accuracy, speed, cost and privacy. We regularly work with models from OpenAI, Anthropic (Claude) and Google (Gemini), as well as open-source models such as Llama and Mistral that can be hosted privately. We build so the model can be swapped later without rewriting your product.
Your data stays yours. We use enterprise AI APIs whose terms exclude training on customer data, or open-source models hosted in your own cloud when data must not leave your environment. We only send the model what it needs, apply access controls and logging, and follow the same security practices as the rest of our work.
We ground the AI in your own content rather than letting it guess. In practice that means:
Usually not. Modern AI models already understand language, so most projects only need your existing documents, knowledge base or records, not a huge training dataset. Custom training or fine-tuning is only worth it for specialised tasks, and we will tell you up front if that applies to you.
We normally start with a short proof of concept on your real data, measured against agreed success criteria, so you can see it working before committing to a full build. From there, the cost and timeline depend on scope, integrations and how much accuracy the use case needs. We give you a clear estimate once the proof of concept shows what is realistic.
There are two main running costs: model usage (you pay the AI provider per request) and hosting. We design to keep these predictable, using smaller models where they are good enough, caching repeated answers and setting usage limits. You get a cost estimate per month before launch.
Yes. We connect AI to the tools you already use, such as CRMs, helpdesks, ERPs, databases and document stores, through their APIs. That is what turns a clever demo into something your team actually uses every day.
We agree the measures before we build, for example hours saved per week, tickets resolved without a person, response time or answer accuracy. We test against them before launch and keep monitoring after it, so you can see the return rather than take it on trust.
You do. Just like our other work, you own the code, prompts, configurations and data pipelines we create for you. The only exceptions are the third-party AI models themselves, which are licensed from their providers.
O2SOFT is a UK-registered product engineering partner building web, mobile and custom software since 2012, with offices in Bradford and London and a delivery centre in Lahore.
We started out building standalone mobile and web apps. We made mistakes early on, and when we did we covered the cost of putting things right, which is a big part of why clients trusted us. From 2013 we introduced strategic management processes and moved to building integrated, business-first solutions.
Since then we have solved over 200 real-world business problems for everyone from billion-dollar brands to early-stage startups, and we still work with clients we first worked with years ago.
We manage projects from start to finish and are known for on-time delivery. We work best on:
We do not place staff on-site at client offices, and we do not do contract-to-hire. Those are better handled by a staffing agency.
We are a team of around 40 professionals, each selected through a rigorous hiring process.
We are registered in the UK with offices in Bradford and London, and our delivery centre is in Lahore, Pakistan. You contract with a UK company and get the capacity of a delivery centre that has been running since 2012.
Every candidate goes through a technical phone interview, an interview with the most relevant person on the team, a coding test that filters out around 80% of applicants, and an in-person interview with reference checks.
We look for people who would write code even if it were not their job, who pick up new ideas quickly under pressure, and who have strong computer science fundamentals.
Yes. We pay people for the time they spend learning, reading, attending conferences and speaking at meetups. We run regular lunch-and-learn sessions and cover the cost of training in Scrum, project management, engineering practice, design and business strategy.
Unless it is needed for an official security clearance, no. Requests for CVs usually mean a staffing arrangement, which is not how we work. Instead, you will meet the people on your team before the project starts, and if you are not excited about them, we are probably not the right fit.
To make sure we are a good fit for each other. We listen to what you want to achieve and look for a clear vision, a realistic budget and real commitment to the project. The first conversation is free and useful either way.
Yes. We sign NDAs as standard practice before any detailed discussion. Just ask during the first conversation and we will send ours or sign yours.
Absolutely. Checking references is an important part of choosing a partner. Once we agree we are a good fit, we will put you in touch with past clients.
You can see many of our recent clients in our case studies. For confidentiality reasons we share deeper project details, a full portfolio and references once we have established a good fit and signed a mutual NDA.
Rarely. RFPs are usually very specific about features but light on business context, and the best feature set is rarely the one written down before anyone has talked to users. We would rather help you discover the right product through interviews, prototypes and real feedback.
Yes, when the scope is clear. A fixed-price project gives you a defined scope, a fixed cost and a delivery date, and it works best after a discovery workshop. We also offer a dedicated team on a monthly basis and an offshore delivery centre model. We will recommend the right one on the first call.
Absolutely. We start with a walkthrough of the live system, followed by a code audit. You then get a detailed report with our analysis, recommendations and, if it makes sense, a takeover plan.
You do. Before we start, we sign a services agreement that gives you ownership of all work and deliverables once our invoices are paid.
Yes, and that is completely normal. Most clients are not ready for development when they first come to us. We help you work out what to build and how much to budget for it through our pre-project consulting.
Yes, we are always evaluating new projects. Pre-project consulting takes around two weeks, and the Discovery, Design and Planning phase takes four to six weeks. Development starts after that.
We start with pre-project consulting to size up your idea, then run a Discovery, Design and Planning (DDP) engagement. Development begins after that, and you are involved at every stage. See our process for the full picture.
A typical project team has two to three developers, supported by UI/UX designers, testers, a project manager and cloud engineers. Each team has a Project Lead, a full-time developer who plans the work and is your main point of contact. Sharing support roles across teams keeps costs down and lets developers focus on production code.
Anyone on the team can talk to you directly, and calls and emails during the working week are normal. On top of that, every project has a fixed rhythm:
Poor communication is one of the most common reasons projects fail, so we build steady routines: a weekly one-hour review of progress, blockers and plans, and a 15-minute daily stand-up to keep momentum. Everything runs on a "no surprises" principle: you always know where things stand, and risks are dealt with as soon as they appear.
An active role. You bring the industry and market knowledge; without it we cannot know if we are building the right product. In practice that means a weekly call or meeting and being reachable during normal business hours.
Yes. That flexibility is often what makes a project succeed. Large changes after kick-off can affect budget or timeline, so we will always show you the impact before you decide.
Always. You get weekly updates, regular releases to a staging site you can try, and direct access to your team. Open, transparent communication is how successful projects are run.
Software is rarely finished. As your business evolves, your product needs to evolve too. After the first release, it often pays to pause new features and polish what is already there based on real user feedback. We can continue with ongoing support and improvements for as long as you need.
QA specialists are involved from the start and test plans are written early. Testing happens at every stage, not just at the end, and code reviews and pair programming keep code quality high throughout.
We make good use of proven open-source tools, libraries and frameworks, as well as our own internal libraries, so we never reinvent the wheel. That leaves our time for the parts that make your product unique.
We cannot give an accurate price until we understand your idea, business, market and users. What we can do is tell you what is typical for projects like yours, based on 200+ projects since 2012, and help you set a sensible budget before you hire us.
Most projects are delivered in phases. The first phase focuses tightly on the features that matter most and sets you up for what comes next. Whatever the budget, we aim for the value we deliver to far outweigh the cost.
We offer three models: a dedicated team (monthly), a fixed-price project (a defined scope for a fixed cost) and an offshore delivery centre (headcount-based). The right one depends on how certain the scope is and your timeline. We will recommend one during the first call.
Many clients start with a focused version 1, learn from real users, and come back for the next round of design and development with a much clearer direction.
A small team is usually two to three full-time developers; a large one is four to six. As a rough guide, a 1,200-hour project with a team of three takes about three months (1,200 hours divided by 3 people at 32 hours a week, divided by 4 weeks a month). On larger projects we can add people and move faster.
With custom software you are paying a team to design and build the first version of a product, and managing risk is part of that work. Bugs found during a sprint are fixed in the next one alongside other backlog items. We also have standards and review practices in place so that each team gets it right the first time as often as possible.
For hourly work, we invoice monthly in arrears: for example, hours logged in January are invoiced on 1 February, with payment due within ten days. For a dedicated team, each month is paid in advance.
We work hard to de-risk projects up front so estimates are accurate. If we get it wrong, we cover the extra cost of getting back on track, because it is the right thing to do. You can also cancel at any time, which keeps us accountable at every stage.
We choose the right tools for each project. Around 80% of our work uses:
Yes. Where the app does not need heavy background processing, we often use React Native or Flutter to target both platforms from one codebase, which cuts development time and cost significantly. See mobile app development.
Yes. Every app we build includes an API layer, and we regularly integrate with third-party APIs too.
Retail and e-commerce, travel, healthcare, education, hospitality, real estate, sports, distribution, risk analytics and IoT. See our case studies for examples.
Yes. We have handled HIPAA compliance on large healthcare projects, and PCI compliance is part of our everyday work because most of our projects process, transmit or store card data. Our full security and compliance pack is available on request.
Yes. We host and manage client applications on AWS and Azure, and thousands of people rely on the products we host every day.
Yes. We are ISO 27001 certified, and our full security and compliance pack, covering data handling, security and hosting policies, is available on request.
We back up every night, and every few hours for some systems, so we can recover quickly. Our recovery response time is under 24 hours.
Try a different word, or ask us directly.
Whether you are starting from an idea or rescuing something already in flight, the first conversation is free and useful either way.
