For investors
down arrow
The problem
Primary school and Early Years teachers face an overwhelming administrative workload due to the important yet time-consuming task of tracking student progress. This involves systematically recording and evaluating students’ academic performance against targets determined by the National Curriculum.
We empower teachers to flexibly evaluate student progress by automatically sorting feedback into targets and proficiency levels using natural language processing
Competitive landscape
In a stagnant industry marked by a lack of innovation, market leaders provide rigid  tracking requiring manual and tedious data entry. They focus on ticking objectives, basic data analytics and often follow a “one size fits all” approach. This inflexibility adversely affects students on different learning paths, such as those with special educational  needs or disabilities (SEND), English as an additional language (EAL) or those under pupil premium, leading to unfair assessment despite developing at their own pace.

Progress trackers are an essential part of learning management systems (LMS). Classify addresses the persistent demand for a more holistic approach to student tracking. Using natural language processing for smart classification, we deliver more meaningful insights into children’s individual learning paths.
Table 1: Competitors Analysis
Classify will provide the features above directly to schools in a B2B model. This enables centralized decisions in schools, fostering widespread adoption among teachers and facilitating seamless integration across subjects. Schools can now pay for one platform with all of their requirements rather multiple fragmented subscriptions.
“We find ourselves juggling a variety of software solutions—ranging from formative or summative assessments, mental health tools to intervention tracking.

Currently, our school utilizes approximately five different software subscriptions, each serving a distinct purpose with separate logins and payments, and they don’t communicate with each other”
Headteacher of William Ford C of E School
Market potential
Our Total Addressable Market comprises of 32,163 schools in the UK. Focusing on primary schools in England, our Serviceable Addressable Market consists of 16,791 schools, with an average revenue of £1,967 per school per year.

Leveraging the revenue from the first year we are profit making, our Serviceable Obtainable Market positions us at a 4% of our SAM. This demonstrates initial success and potential for market share growth as Classify expands.
Market Size
Validation: our experiments
We mapped our riskiest assumptions in terms of the product, the business and the technology, to assess desirability, viability and feasibility respectively.

Below, we outline the experiment designed to test each hypothesis, as well as their category, risk level (considering their importance and uncertainty), key learnings and next steps.
Stakeholders involved in our validation
Product
High Risk
Hypothesis
Inputting data with Classify is more 
time-efficient for teachers than 
our competitors
Method
Wizard of Oz: teachers wrote lesson summaries, which were manually classified behind the scenes, then presented on a mock-up platform as if it was automated.
Key Learnings
  • Teachers found writing reflections to be an additional, time-consuming task and were unwilling to do it regularly.
  • Audio/dictation-based input was suggested as preferable for live marking & capturing key moments especially in primary education, saving time.
  • Teachers appreciated the meaningful, classified information provided for both academic and personal development.
Next Steps
  • Identify features to prioritise based on teacher feedback for streamlined development.
  • Develop a mock-up focusing on dictation-based input to validate its desirability.
Early prototypes were shown to teachers as a “Wizard of Oz” experiment.
Feature Sorting Cards
Product
High Risk
Hypothesis
The classifying algorithm and dictation feature will be the most useful to teachers
Method
Feature sorting cards: teachers ranked 7 features to assess their priority. Features were initially positioned randomly to mitigate biases.
Key Learnings
  • Easy-to-use navigation was more critical than anticipated, addressing a key pain point of existing software, making it a high risk/high importance assumption.
  • Most desirable features were dictation (with new use cases identified e.g. recording artwork, PE lessons) and automatic classification of reflections into targets.
  • Filters and automatic report generation was not as useful for teachers, but would be of higher importance to senior leadership.
Next Steps
  • Develop a high-fidelity prototype to use in a classroom with dictation features to assess usability and on-the-go tracking.
  • Test the technical feasibility of an AI algorithm being able to classify the teacher’s unfiltered reflections against school curriculum goals.
Product
Medium Risk
Hypothesis
Classify can be used in a classroom setting, on the go without disrupting a class.
Method
Focus group: teachers in a school using an interactive prototype during a lesson with children, to observe usability.
Key Learnings
  • Teachers were able to use easily it while teaching and live marking, and did not find it to be a hindrance.
  • The UI was perceived as friendly and more intuitive than existing platforms.
  • Flexibility to each school is an integral component we hadn’t previously considered in depth. There are differences in success metrics, objectives, and “assessment ladders” (different ways of assessing to which extent the target is met).
Next Steps
  • Share existing prototype with teachers to recruit beta testers.
  • Show refined prototypes to teachers and headteachers assess the extent schools are willing to pay for the software.
A second MVP was developed to be used in the focus group
Focus groups with teachers from Becket Primary School
“A tool which uses AI to take random observations as well as assessment marks and collates them against learning objectives, dropping them into all the right places would be, I think, invaluable. If all data could be in one central place, so that results in one subject could cross-fertilise the objectives of a different subject that would make tasks punch above their weight”
Surveyed primary school teacher
Technical
Medium Risk
Hypothesis
The capabilities of the Chat GPT API can be leveraged to accurately classify teachers’ reflections
Method
Expert interviews: Consulted with Dr Sam Cooper, Dr Freddie Page,  Lei Ge, (PhD student) about LLM capabilities and the feasibility of embedding AI into our platform.

Prompt engineering experiment:  to test if we could get desired classification using GPT-4.
Key Learnings
  • AI can accurately classify observations into correct targets and levels for students but requires caution for potential errors, generalisations, and biases.
  • Leveraging ChatGPT’s capabilities is feasible but to obtain a higher accuracy, the model should be fine-tuned with our own dataset.
  • Human verification of AI-generated data is necessary, with teachers responsible for review due to privacy constraints (Classify does not access student data).
Next Steps
  • Given that teachers will serve as the "human in the loop" for verifying generated content, guidelines will be provided for teachers to ensure thorough review, responsibility and  accountability when submitting reports.
Example input reflection used in prompt engineering experiment
Output results given by ChatGPT
Example data protection assessment that primary schools use to assess safety of new platforms
Technical
Medium Risk
Hypothesis
Classify will be sufficiently secure to not endanger children’s personal data.
Method
Reviewed school policies regarding data protection on softwares and consulted with a Web Security expert, Dr. Sergio Maffeis.
Key Learnings
  • Main threats for cloud-based platforms are standard across web applications with sensitive data, there are existing standards to protect against common attacks.
  • Implementation of API with local data storage is necessary to prevent data exposure to OpenAI.
Next Steps
  • Prioritise security and privacy from the start of development.
  • Further investigate NCSC tests for enhanced security measures.
Business
High Risk
Hypothesis
Classify can save schools money by combining multiple subscriptions into one
Case Study: to compare our offerings to existing competitors, we asked a school in East London to provide us with a breakdown of their existing tracking platforms. Like many schools, they subscribe to multiple in order top  meet all of their tracking needs. All the calculations are based on 335 students.
Method
Key learnings
  • We are able to offer Classify (consisting of standard and extra features) at a lower price than existing subscriptions.
Next steps
  • Test headteachers are willing to pay for our subscription price.
Business
High Risk
Hypothesis
Teachers will be willing to pay £7 per student for Classify.
Method
Pitching and showcasing product prototype to headteachers, followed by offering the opportunity to sign a purchase letter of intent (non-binding agreement) - featuring the price of the platform, to assess willingness-to-pay.
Key learnings
  • Many headteachers are willing and interested in transitioning to Classify, as indicated by 3 letters of purchase intent.
  • Headteachers emphasised the critical importance of security and privacy, given the sensitive personal data schools handle, especially concerning minors.
  • They highlighted the benefits of improved insights and analytics, enabling earlier interventions and better assessment of their effectiveness through the platform's algorithm.
  • One interested headteacher was unable to sign the form, due to the school ‘facing financial difficulty’ and their hesitancy to transition teachers to another platform after a recent change.
Next steps
  • Confirm feasibility of switching platforms - exporting & importing information, data acquisition.
  • Having received enthusiasm for the platform, we need to further develop it in preparation for beta testing with the verified and willing schools.
  • Continue to expand our network of schools willing to participate in our pilot scheme.
Business
Beta Testers and Pilot Schemes
We tested the use intent by offering teachers the opportunity to sign a non-binding letter of support. Overall we learned that teachers are willing to switch from their existing platforms to try out Classify if their school adopted it, and received 13 letters of support. 
Teachers excited to be in our beta testing cohort.
Primary schools excited to be a part 
of our pilot scheme.
Traction
We’ve received overwhelming support and excitement for our platform, with  three schools already on board for a pilot scheme. We’ve also received letters of endorsement from teachers as well as purchase intent letters from  headteachers.

Parents are eager about the product, appreciating the potential for more detailed feedback and insights into their child's development. This can foster collaboration with educators for early interventions and support at home.

We have a growing LinkedIn following, with 683 page views and 57 followers including teachers, pupil tracking consultants, heads of education and AI experts.

We’ve already experienced the positive word-of-mouth networking effect, with secondary schools and academics from Birmingham City University contacting us, being keen to pilot with us.
Risk assessment
Our previous product assumptions were de-risked with our experiments with teachers, but business risks relating to switching costs, and technical feasibility risks still remain. The highest scoring risks remaining in our risk register have been detailed below, along with their mitigation strategy below.
Risk Assessment
Next steps
Product Roadmap
Our next stage is to take the MVP further in the R&D phase, and iterate on the interface and usability. This will allow us to optimize the features to then build the platform on the cloud, which will require full stack platform development.

After development, the product will be ready for a beta release, where we will be co-designing with the schools on the pilot scheme.

By September 2025, we will have our software readily available, and will continue increasing our cloud infrastructure, data storage and customer support for continuous growth. During this period, we will also test new features to be released, allowing us to expand to secondary schools by September 2027, allowing us to increase our market share.
Hiring Strategy
We will initially have the CEO, CTO and CCO full time and COO and CPO part time, and we will be responsible for the development, marketing and sales until the full scale launch. By then, we will need a full stack software developer to help us scale up the operations, and a full time sales person to help us increase our leads and continue growing our customer base.
Route to Market
Our strategies for customer acquisition are sales people directly in contact with schools, Linkedin advertising (CAC = £100), Facebook ads (CAC = £21) and conference (CAC= £400). The combination of these inorganic sales will help us increase visibility, credibility and trust, thus building a more solid foundation for organic word of mouth.
We’re delighted to have been selected as finalists for the Mayor’s Entrepreneurship Competition.
The team
At Classify, we are more than a team of design engineers; we're a group of individuals deeply connected to the education system, motivated to change the status quo and alleviate the administrative burden teachers are under. With our rigorous approach to testing, experiments and validation, we are the right people to turn this vision into a reality.

Our team has expertise in AI, software development, consumer research and marketing, and together with our advisors, teacher supporters and school connections, we want to put Classify into teacher’s hands as soon as possible! We're not just building a platform; we're crafting a tool born out of a deep understanding and a sincere desire to uplift the teaching experience.

We are looking for school collaborations, beta testers and connections in the EdTech industry. Join us, and let's redefine education together!
right arrow
Join our mailing list
Our advisors
Classify has come a long way in a short space of time, thanks in part to the incredible support of our advisors who share our vision for what the future of education can be.
Shafina Vohra
Finalist in the Global Teacher Prize & A Level psychology teacher, our education advisor
Snigdha Mazumdar
CEO of Webtage and our marketing advisor
Lei Ge
PhD student in Generative AI and our technical advisor
© Classify 2024. All rights reserved.
Contact Us
Home
Why we exist
The team
Blog
FAQs
Investors
Follow Us
LinkedIn
Instagram
X/Twitter