Company Profile

BattleBottle helps commanders make better and faster counter-drone decisions. It is an AI-assisted counterUAS (C-UAS) decision-support capability, built in the UK, for planning, training and experimentation.

The problem 

Low-cost drones now arrive in numbers that outpace the people who have to decide what to do about them. Under time pressure, operators and commanders must work out which sensors and effectors to employ, where, in what order and at what cost, while protecting personnel, platforms and critical assets. The limiting factor is the decision, not the equipment.

The user need 

Operators require decision support that enables them to identify, prioritise and orchestrate the most appropriate counter-UAS response using available sensors, effectors and operational constraints, in order to protect personnel, platforms and critical assets against evolving drone threats.

What BattleBottle does 

BattleBottle takes a scenario (terrain, threat composition, the counter-UAS and uncrewed assets available, and a budget), generates and compares the response options, and returns ranked, cost-aware force packages with deployment positions and engagement priorities, in seconds rather than hours. It is advisory: it recommends, a human decides. Outputs export in Cursor on Target (CoT) format for use alongside TAKbased command and control tools.

The capability is fully working, interactive software. It can be demonstrated today and used, tested, piloted and iterated against a sponsor’s own scenarios. 

Who it is for 

• Commanders and counter-UAS operators: decision support before committing to a course of action, cutting cognitive load, decision time and wasted effectors.

• Training establishments: operators and recruits learn counter-UAS employment by repetition in a synthetic environment, scored on outcome and cost, with structured progression from first exposure to assessed proficiency.

• Engagement and the talent pipeline: the same simulation is available as a video game, giving young people an accessible route into uncrewed systems and tactical thinking.

Operational advantage 

Response options are compared against the threat picture before any asset is committed (mission effectiveness). Recommendations prioritise the protection of people, platforms and critical infrastructure (force protection and survivability). The training loop builds operator proficiency without consuming live effectors (readiness). Ranked options replace raw triage under time pressure (decision speed). Every option carries its cost, so cost-per-effect is visible in a war of attrition (resource efficiency).

Modelling, simulation and data 

Behind the capability sits a synthetic environment: a simulation engine with real terrain, representative uncrewed and counter-UAS platforms, electronic warfare zones, weather and a cost for every asset committed. Every completed simulation, whether played by a person or generated by self-play, is captured as structured engagement data: force composition, placement, engagement order, spend and outcome. More than 1,600,000 completed simulations, combining live and synthetic play, have been processed to date.

Scenarios can be built from open-source reporting, sponsor-supplied data or synthetic generation, and sponsor-owned datasets can be used under NDA. A library of more than 90 reconstructed drone engagements from Ukraine (2022 to 2026), each held as a single Cursor on Target file with a source-confidence grade, is available for analysis, planning and experimentation. The models are fine-tuned openweights models, trained on BattleBottle’s own data and hosted in the UK.

Security and sovereignty 

UK company and UK-hosted AI. Full encryption end to end. Delivered as compiled executables; source code and model weights never leave BattleBottle’s secure environment. Own collected datasets. Patent pending at the UK Intellectual Property Office.

What we offer 

• Software capability: the decision-support tool, for evaluation, pilot and licence.

• Modelling and simulation service: scenario build from a sponsor’s terrain, threat picture and inventory, course of action comparison and after-action data.

• Decision-support and data exploitation services: analysis of a specific defended site or operation, and structured engagement datasets for training, experimentation and model development.

• Training and advisory: the simulation as a training tool with scoring and progression, and AI and synthetic-environment expertise for Defence programmes.

Validation 

Every build passes internal verification. Independent validation is the next step: we are seeking a sponsored trial that compares BattleBottle’s recommendations with subject-matter-expert plans against the same scenarios, scored blind, with the results written up as after-action data. If you can host that trial, we want to talk.

Engagement to date 

• DSET 2026 Government Audience Award, voted by MoD and NATO delegates.

• Invited exhibitor at the UK MoD Counter-UAS Conference, Larkhill, and the NATO Allied Air Command Counter-UAS Conference, Ramstein.

• Finalist and UK representative, I/ITSEC 2026 Serious Games Showcase and Challenge, Orlando.

• Pilots being set up in Ukraine with UK MoD & NATO support. 

• Advisory board drawn from UK government defence and security, the British Army, the Royal Navy, the US Army and exited DefenceTech founders. 

Founders 

Francis Jones, Chief Executive Officer: former STEM teacher and school leader, AI and education specialist, part-time PhD researcher at Brunel University London on knowledge transfer mechanisms, founder of prior AI education ventures, and King’s College London Distinguished Alumnus of the Year 2025 (STEM) for work on AI in education.. 

Bram Janssen, Chief Technology Officer: one of the original VC-backed blockchain video game developers, co-founder of Duckland Games (Forgotten Playland), and a speaker and exhibitor at GDC, Gamescom and AVAX Summit. He builds the simulation engine. That pairing is the point: one founder studies how knowledge transfers, the other builds engines people want to keep playing. The result is a decision tool that trains its users and learns from them at the same time.

Not an incremental upgrade 

Wargaming and unit-laydown tools let a planner place units and run a simulation. They do not generate and rank the options for the planner, price each option, or learn from every engagement played. BattleBottle does all three, and the gamified layer is what makes the third possible at scale.

Working with BattleBottle 

• Pilot sponsors in the UK MoD and allied forces who want to evaluate the capability against their own scenarios.

• Primes and system integrators who want to test it within an existing counter-UAS command and control system.

• Training organisations looking for a synthetic environment for counter-UAS employment.

• Investors: the pre-seed round is open to defence and dual-use investors; the investor summary is available on request.

Contact Francis Jones at francis.jones@battlebottle.tech to arrange a live demonstration or a scoped pilot.

Photo gallery