Organising a hackathon?
These 30 themes are challenges to set for your teams, three for each corporate function. Choose one with the seven questions for choosing a theme, then turn it into your event brief with the Hackathon Brief Template.
Competing in a hackathon?
I've judged dozens of hackathons, and the winners are rarely the teams with the cleverest code. They read the theme better, build only what proves their idea, and start their pitch from hour one. How to Win a Hackathon shows you how. Browse the themes below and pick one to brainstorm ideas in Ainna, our AI innovation platform.
Explore hackathon themes by corporate function: Strategy · HR · Finance · Legal · Operations · Sales · Marketing · Service · Product · Technology
A good hackathon theme starts with a real problem - something your organisation or customers actually struggle with - and invites teams to explore solution paths.
In my experience organising and participating in hackathons at Microsoft, Accenture, and various startups - and as I discuss extensively in The Innovation Mode - the best themes share three characteristics: they're grounded in genuine pain points, they're specific enough to focus creativity, and they're open enough to allow for, what I call, the element of surprise - unexpected solutions.
But there's a fourth element that separates good hackathons from transformative ones: inclusivity.
For too long, hackathons have been stereotyped as events for developers and technical teams. This is a misconception worth correcting. The most valuable innovation insights often come from diverse teams that include people closest to the problem - customer service representatives who hear complaints daily, operations managers who see process breakdowns firsthand, salespeople who understand why deals are lost. In many cases, these domain experts have been excluded or discouraged from joining hackathons because they felt ‘not sufficiently technical’ to frame and prototype their ideas.
AI has shattered this barrier.
Today, anyone can prototype. A finance analyst can use AI coding assistants to build a working demo of their automation idea. A marketing manager can create a functional customer journey tool without writing code. A health professional can prototype a patient communication system. The technical bottleneck that kept non-developers out of hackathons is gone. This means your hackathon themes should be designed to attract diverse participants - not just engineers.
The hackathon theme ideas below are framed around business problems, not technical solutions, precisely because the best innovations emerge when visionaries meet domain, business and technical experts. When you run your next hackathon, actively recruit from operations, finance, HR, customer service, and sales: send the right message, give them the right AI tools and measure what happens - both in terms of opportunities produced and team/collaboration outcomes.
These themes are organised by corporate function, because that is how most corporate hackathons are set up: one function sponsors the event, or a company-wide event runs a track for each. Each is tagged with its innovation type - process, product or business model - because different hackathon types serve different strategic objectives. Each is written as a brief: the challenge and why it matters now, who is affected, what to prepare, where to focus, how to pitch it, the success criteria and what to watch for. Adapt them to your industry and context.
A great hackathon theme doesn't prescribe solutions or technologies - it frames a problem worth solving and encourages diverse teams to surprise you.
Hackathon Ideas for Strategy and Leadership
Strategic themes need the people who own the outcome. Invite strategy, finance, product and sales leaders, and a few people from the front line who see customers every day.
1. The AI-Native Competitor
How might an AI-native entrant win our most profitable customers, and how might we disrupt ourselves before it does?
New entrants run with a fraction of our people and cost, and incumbents tend to see them too late.
Read the briefHide the brief
- Who is affected
- Our most profitable customers, and the business units whose margins a newcomer would target first.
- Preparation
- Our five most profitable customer segments, what they pay us for, and two or three AI-first entrants to study.
- Where to focus
- The competitor's offer and economics, the customers most exposed to it, and the response we would launch first.
- How to pitch it
- Pitch as the competitor to one of our customers, then present our answer.
- Success criteria
- A threat leadership recognises as real, a response customers rate, and an owner willing to test it.
Watch for A comfortable competitor. If the attack worries nobody, it isn't sharp enough.
2. New Business From What We Already Know
How might we turn our proprietary data, expertise and customer relationships into a new revenue stream?
AI makes expertise packageable at a cost that was impossible a few years ago.
Read the briefHide the brief
- Who is affected
- Customers who would value what we know, and the experts whose knowledge never leaves the company.
- Preparation
- The data and expertise we hold, what contracts and privacy rules allow, and five customers willing to react to an offer.
- Where to focus
- A benchmark service for customers, an expert advisor product, or an insight service for partners.
- How to pitch it
- Show the product as a customer would use it, with what customers said they would pay.
- Success criteria
- Customers who would pay, a legal route to offer it, and a business owner ready to run a pilot.
Watch for Selling what we have no right to share. Check contracts and privacy before the build.
3. Big Decisions, Tested Before We Commit
How might we stress-test our biggest strategic decisions against plausible futures before we commit capital?
Planning cycles run slower than markets, and AI can simulate the options in hours.
Read the briefHide the brief
- Who is affected
- The executives who make the calls, and everyone who lives with the results.
- Preparation
- One recurring strategic decision, such as an investment, a market entry or a pricing move, with its past cases and outcomes, and the people who decide it.
- Where to focus
- A scenario simulator for that decision, an AI red team that challenges every investment case, or a radar for early market signals.
- How to pitch it
- Replay a past decision with the tool and show what it would have flagged.
- Success criteria
- Decision-makers who would use it next quarter, and a past decision it would have improved.
Watch for False precision. The value is in better questions, not a single number.
Hackathon Ideas for HR and Talent
Invite recruiters, HR business partners, hiring managers and people who joined in the past year. Clear any data about people with data protection before the event, and use anonymised or synthetic records.
4. Hiring for Potential, Not CVs
How might we assess candidates on demonstrated ability rather than CVs, and reach the high-potential talent our process screens out?
AI-polished applications all look alike, and CV filters miss people from non-traditional backgrounds.
Read the briefHide the brief
- Who is affected
- Hiring managers short of the right people, recruiters buried in AI-polished applications, and capable candidates without the usual CV.
- Preparation
- Two hard-to-fill roles, what the best people in them actually do, and five recent hires willing to say how they were found.
- Where to focus
- Work samples in place of CV screens, matching on the skills that predict success, or the right people already inside the company.
- How to pitch it
- Take one real candidate through the new process, and show what the old one would have missed.
- Success criteria
- Hiring managers who would use it for their next role, a fairer shortlist on a bias check agreed with HR, and a shorter time to hire.
Watch for The law. The EU AI Act treats AI in recruitment as high-risk. Bring in legal and data protection before the event, and keep a person on every decision.
5. One Role, Redesigned for the AI Era
How might we redesign a critical role so that AI raises the productivity, impact and career prospects of the people in it?
AI changes tasks before it changes jobs, and the companies that redesign roles deliberately keep their best people.
Read the briefHide the brief
- Who is affected
- The people in the role, their managers, and the customers or colleagues they serve.
- Preparation
- One critical role, a week of its task logs, and three people in the role to join the teams.
- Where to focus
- The role's tasks mapped with the people in it, what AI should take on, and the redesigned role and its career path.
- How to pitch it
- Show a day in the role before and after, told by someone who does the job.
- Success criteria
- Hours moved to higher-value work, people in the role who want the new version, and a manager ready to pilot it.
Watch for Redesigning the role without the people in it. They know where the value is.
6. Reskilling at the Speed of AI
How might we help every employee learn the AI tools, skills and ways of working their role now needs, as fast as AI changes it?
Skills now age in months, and annual training plans cannot keep up.
Read the briefHide the brief
- Who is affected
- Employees whose jobs are changing, managers who need new skills in their teams, and learning teams working to annual plans.
- Preparation
- The skills the company will need in two years, the current training catalogue, and ten employees willing to test.
- Where to focus
- Learning built into daily work, peer-led AI champions, or a marketplace that matches people to projects where they learn.
- How to pitch it
- Show one employee learning a new skill in a week, and what they can now do.
- Success criteria
- Skills applied at work within a month, employees who would choose it, and a learning owner ready to scale it.
Watch for Another course catalogue. People learn by doing the work, not by finishing modules.
Picked one? Turn it into your event brief with the Hackathon Brief Template.
Hackathon Ideas for Finance
Invite controllers, financial planning, someone from treasury and budget holders from the business.
7. Finance That Steers, Not Reports
How might finance move from explaining last quarter to shaping the decisions that will define the next?
Closing and reporting take most of finance's time, and AI can take on the routine.
Read the briefHide the brief
- Who is affected
- Business leaders deciding without a forward view, and finance teams busy closing the books.
- Preparation
- Two years of monthly results and forecasts, the questions budget holders ask, and two leaders willing to test.
- Where to focus
- A continuous forecast leaders use every week, a "what if" assistant for budget holders, or closing the books in a day.
- How to pitch it
- Answer a leader's real question about next quarter, live.
- Success criteria
- Forecast accuracy on past data, leaders who would use it every week, and a finance owner for it.
Watch for A forecast nobody trusts. Show the drivers behind every number.
8. Pricing for the Value We Create
How might price optimisation align what each customer pays with the value we deliver, and lift margin without losing trust?
AI lets buyers compare everything, and lets sellers measure the value each customer gets.
Read the briefHide the brief
- Who is affected
- Customers who feel they pay for what they don't use, and sales teams negotiating exceptions.
- Preparation
- A year of usage and revenue by customer, the discounts sales agreed, and five customers willing to react.
- Where to focus
- Outcome-based pricing for one product, usage tiers, or a pilot price tested with real customers.
- How to pitch it
- Show what five real customers would pay under the new model, and how they reacted.
- Success criteria
- Revenue against today's model, customers' reactions, and finance and sales agreeing to a trial.
Watch for A model too complex to explain. If sales can't say it in a sentence, customers won't buy it.
9. Stopping the Projects That Don't Pay Back
How might we identify early the initiatives that won't deliver a return, and reallocate their budgets to stronger bets?
Every portfolio carries projects nobody will stop, and AI can read the evidence across all of them.
Read the briefHide the brief
- Who is affected
- Leaders funding the portfolio, teams stuck on projects going nowhere, and the ideas waiting for budget.
- Preparation
- The current portfolio with budgets, milestones and status reports, and the people who approve funding.
- Where to focus
- A portfolio view that links spend to outcomes, stop criteria agreed up front, or a monthly decision forum.
- How to pitch it
- Run the real portfolio through it and show the three projects it would question first.
- Success criteria
- Decisions leadership would take on its evidence, and money freed for better ideas.
Watch for Blame. Stopping a project early should count as a good decision, not a failure.
Picked one? Turn it into your event brief with the Hackathon Brief Template.
Hackathon Ideas for Legal, Risk and Compliance
Invite in-house lawyers, compliance and risk officers, and the colleagues who wait for them.
10. Contracts as a Competitive Advantage
How might we make contracting so fast and transparent that ease of doing business becomes a reason customers choose us?
Deals stall in legal for weeks, while AI drafts and reviews in minutes.
Read the briefHide the brief
- Who is affected
- Customers waiting to sign, sales teams whose deals stall, and lawyers buried in routine contracts.
- Preparation
- The contract playbook, recent deals with how long each took to sign, and two customers willing to test.
- Where to focus
- Self-service contracts within the playbook, terms customers can actually read, or negotiation closed in a day.
- How to pitch it
- Take a real deal from request to signature, live.
- Success criteria
- Days to signature, lawyers who sign off the output, and a sales leader keen to use it.
Watch for Speed without control. Keep a lawyer on everything outside the playbook.
11. Compliance Built Into the Work
How might we build compliance into everyday workflows, so that it is assured as the work is done rather than audited afterwards?
New rules on AI, sustainability reporting and data protection arrive faster than compliance teams can grow.
Read the briefHide the brief
- Who is affected
- Teams slowed by checks after the fact, compliance staff chasing evidence, and the business exposed in between.
- Preparation
- One regulation that matters to us, the controls it requires, and the process where those controls are checked today.
- Where to focus
- Rules turned into checks inside the tools people use, regulatory change mapped to owners, or evidence collected as the work happens.
- How to pitch it
- Take one control from a manual check to a built-in one, and show the evidence it produces.
- Success criteria
- Controls an auditor would accept, time saved, and a compliance owner ready to pilot it.
Watch for Automation the regulator won't accept. Bring compliance in before the build.
12. Responsible AI That Speeds Us Up
How might AI governance give teams the clarity and confidence to adopt AI faster, while keeping risk in check?
Every function wants AI, and unclear rules either block it or let risk through.
Read the briefHide the brief
- Who is affected
- Teams waiting for approval to use AI, and the legal, risk and security teams asked to approve it.
- Preparation
- The AI uses teams have asked for, the current policy, and the EU AI Act categories that apply.
- Where to focus
- A self-service AI risk check, a catalogue of approved tools, or a register of AI uses ready for the AI Act.
- How to pitch it
- Take three real requests through the new route, with the decision and the time each took.
- Success criteria
- Approvals in days rather than months, risks caught, and legal and security agreeing to run it.
Watch for A policy nobody reads. The route must be faster than going around it.
Picked one? Turn it into your event brief with the Hackathon Brief Template.
Hackathon Ideas for Operations and Supply Chain
Invite operations and site managers, planners, maintenance engineers, buyers and two suppliers.
13. Seeing the Next Disruption First
How might we detect supply chain disruptions weeks before they reach us, and act ahead of our competitors?
Shocks are now routine, and AI can read the signals across suppliers, news and logistics.
Read the briefHide the brief
- Who is affected
- Planners and buyers fighting shortages, and the customers waiting for late orders.
- Preparation
- Our critical suppliers and parts, the last three disruptions and how long we took to react, and two suppliers willing to share data.
- Where to focus
- A supplier risk radar, response playbooks that start automatically, or alternative sources lined up in days.
- How to pitch it
- Replay a past disruption and show when the team would have known, and what it would have done.
- Success criteria
- Weeks of warning on past events, few false alarms, and a supply chain owner for the alerts.
Watch for Alerts nobody acts on. Decide who responds, and how, before the build.
14. Zero Unplanned Downtime
How might we predict failures in our plants, fleets or critical systems early enough to eliminate unplanned downtime?
Predictive maintenance has been promised for years, and AI with cheap sensors finally makes it practical.
Read the briefHide the brief
- Who is affected
- Production and service teams hit by breakdowns, maintenance engineers on call, and the customers who wait.
- Preparation
- A year of maintenance logs and failures for one type of equipment, the sensor data already collected, and two engineers.
- Where to focus
- Early warnings from data we already collect, an AI co-worker for technicians, or spare parts that arrive before the failure.
- How to pitch it
- Replay a past breakdown and show when the warning would have come.
- Success criteria
- Failures predicted on past data, false alarms engineers accept, and a plant manager ready for a trial on one line.
Watch for A model trained on a handful of failures. Check what it misses, not only what it finds.
15. Cutting Carbon and Cost Together
How might we cut emissions and operating costs together at one site or route, and build the case to scale it?
Reporting duties and energy prices are both rising, and AI finds savings people miss.
Read the briefHide the brief
- Who is affected
- Site and logistics managers under cost pressure, sustainability teams reporting emissions, and customers asking for both.
- Preparation
- A year of energy, fuel and production data for one site or route, and the sustainability team's figures.
- Where to focus
- Energy use at one site, route and load planning, or comparing suppliers' footprints.
- How to pitch it
- Show the savings in money and in carbon for one site or route, with the changes behind them.
- Success criteria
- Savings the sustainability and finance teams both accept, and a site manager ready to try it.
Watch for Numbers you can't defend. Overstated claims are worse than none.
Picked one? Turn it into your event brief with the Hackathon Brief Template.
Hackathon Ideas for Sales
Invite account managers, the bid team, sales operations, a sales leader and two customers.
16. Selling When Buyers Ask AI First
How might we stay on the shortlist when buyers rely on AI assistants to research, compare and purchase?
Buyers increasingly start with an AI assistant, and the first shortlist may be made by a machine.
Read the briefHide the brief
- Who is affected
- Sales and marketing teams whose first contact with a buyer may now be the buyer's AI, and customers who want faster answers.
- Preparation
- What AI assistants say about us and our competitors today, and three customers willing to describe how they research now.
- Where to focus
- Being found and recommended by AI assistants, offers an AI agent can compare and order, or the salesperson's role after the shortlist.
- How to pitch it
- Ask an AI assistant to choose a supplier in our category, live, before and after.
- Success criteria
- Where we appear in AI answers, a buying journey customers say they would use, and a sales leader ready to pilot it.
Watch for Tricks that game one assistant this month. Build on what buyers and their agents actually need.
17. Every Seller as Good as Our Best
How might we give every salesperson the preparation, insight and judgement of our top performers?
AI can learn from what the best do and coach everyone else.
Read the briefHide the brief
- Who is affected
- Newer salespeople learning by trial and error, sales managers with little time to coach, and the customers they meet.
- Preparation
- What top performers do differently, in their own words and from the deals they won, and five salespeople to test with.
- Where to focus
- Meeting preparation in minutes, coaching on live deals, or a playbook learned from the deals we won.
- How to pitch it
- Prepare a real customer meeting both ways and let a sales leader judge the difference.
- Success criteria
- Salespeople who use it every week, and a sales leader willing to roll it out.
Watch for Surveillance. Use it to coach people, not to score them.
18. The Customers We Keep Losing
How might we use the evidence in our lost deals to win the customer segment we consistently lose to competitors?
Lost deals hold the most honest feedback we get, and AI can finally read all of it.
Read the briefHide the brief
- Who is affected
- Salespeople losing the same kind of deal, and the product and marketing teams who never hear why.
- Preparation
- Two quarters of lost deals with their notes and emails, and salespeople willing to be interviewed.
- Where to focus
- The real reasons behind the losses, a new offer for the segment we lose, or a win-back test.
- How to pitch it
- Show the losses grouped by reason, and the change that would win back the biggest group.
- Success criteria
- Reasons sales and product agree with, a change an owner commits to, and a way to keep learning from losses.
Watch for Blaming price. Ask what the buyer valued more than the difference in price.
Picked one? Turn it into your event brief with the Hackathon Brief Template.
Hackathon Ideas for Marketing
Invite marketing, brand, product, a salesperson and legal.
19. A Market of One
How might we personalise every customer interaction, at a cost that does not grow with the number of customers?
AI makes one-to-one content affordable, and customers ignore what isn't about them.
Read the briefHide the brief
- Who is affected
- Customers flooded with messages that don't apply to them, and marketers writing every version by hand.
- Preparation
- The main customer groups and what each needs, the brand rules, and customer data cleared with data protection.
- Where to focus
- Journeys that adapt to each customer, content written for each customer's context, or creative tested before launch.
- How to pitch it
- Show the same campaign for three different customers, and what each said.
- Success criteria
- Response against today's campaign in a test, brand approval, and an owner ready to run it.
Watch for Personalisation that feels intrusive. Use only what customers expect you to know.
20. Hearing the Market Before Competitors Do
How might we anticipate shifts in customer demand before our competitors see them?
Reviews, forums, calls and searches hold the signals, and AI can read all of them every day.
Read the briefHide the brief
- Who is affected
- Product and marketing teams deciding on last year's research, and customers whose needs go unheard.
- Preparation
- Six months of reviews, survey comments, sales notes and support conversations, cleared for use.
- Where to focus
- An insight radar, concepts tested with real customers in a day, or a weekly brief for product and sales.
- How to pitch it
- Present one emerging need the team found, with the evidence and a concept to meet it.
- Success criteria
- Insights product and sales would act on, and a concept customers want.
Watch for Trends without evidence. Every insight should link to what customers actually said.
21. From Idea to Market Test in Two Weeks
How might we validate a new offer, segment or market in two weeks rather than six months, before committing a launch budget?
AI compresses research, content and testing, so a market test now costs a fraction of a launch.
Read the briefHide the brief
- Who is affected
- Teams with ideas waiting for budget, and leaders deciding on launches without evidence.
- Preparation
- One new offer, segment or market leadership is considering, a small test budget, and what a full launch would require.
- Where to focus
- A launch kit for tests, a landing page and offer for the new market, or a pilot with paying customers.
- How to pitch it
- Show the test running, with the first real responses.
- Success criteria
- Real responses from the market, the cost of the test against a launch, and a decision leadership can take on it.
Watch for A full launch plan. A small live test tells you more.
Picked one? Turn it into your event brief with the Hackathon Brief Template.
Hackathon Ideas for Customer Service
Invite agents, team leads, the owners of the top contact reasons, and customers.
22. Fixing Problems Before Customers Notice
How might we detect and resolve problems before customers ever need to contact us?
AI can spot the signs of trouble in orders, usage and payments.
Read the briefHide the brief
- Who is affected
- Customers who discover problems themselves, and agents handling contacts that should never have happened.
- Preparation
- The top twenty contact reasons, the data that shows each problem coming, and the owners of the causes.
- Where to focus
- Proactive alerts and fixes, updates before customers ask, or removing the top reasons for contact at the source.
- How to pitch it
- Take one common problem and show it fixed before the customer would have called.
- Success criteria
- Contacts that would no longer happen, customers who notice the difference, and the owner of the cause agreeing to a pilot.
Watch for A faster chatbot. Removing the reason to contact us is worth more.
23. People for the Moments That Matter
How might AI resolve routine requests so that our people can focus on the moments that decide customer loyalty?
Customers accept AI for simple requests, and expect a person when it counts.
Read the briefHide the brief
- Who is affected
- Customers in difficult moments, such as a complaint, a loss or a crisis, and agents with no time for them.
- Preparation
- Transcripts of moments that won or lost customers, with personal data removed, and the routine contacts that fill agents' days.
- Where to focus
- The moments that decide loyalty, AI for the routine, and agents free to solve problems instead of following scripts.
- How to pitch it
- Show one routine contact handled by AI and one critical moment handled by a person with time to help.
- Success criteria
- Agent time freed for those moments, customers' reactions, and a service leader ready to pilot it.
Watch for Hiding the person. Customers must always be able to reach someone.
24. Service That Grows Revenue
How might customer service become a source of revenue rather than a cost centre?
Service hears customers' needs first, and AI can tell when help is also an opportunity.
Read the briefHide the brief
- Who is affected
- Customers whose needs go unmet, agents measured only on speed, and sales teams who never hear what service knows.
- Preparation
- A sample of service conversations with personal data removed, the products customers ask about, and a sales manager.
- Where to focus
- Upgrade and renewal moments, a premium service customers pay for, or insight passed to product.
- How to pitch it
- Show a real conversation where help became a sale, and how often that moment occurs.
- Success criteria
- The revenue those moments could bring, customers' reactions, and service and sales agreeing to a pilot.
Watch for Turning service into selling. Help first, and offer only what solves the customer's problem.
Picked one? Turn it into your event brief with the Hackathon Brief Template.
Hackathon Ideas for Product and R&D
Invite product managers, engineers, designers, researchers and customers.
25. A Product That Adapts to Each User
How might our product adapt to each user's goals and context, and take work off their hands?
AI features are becoming expected, and the best ones remove work instead of adding buttons.
Read the briefHide the brief
- Who is affected
- Customers who give up on complex products, and the support teams who teach them.
- Preparation
- Usage data showing where customers struggle or repeat steps, and five customers to test with.
- Where to focus
- Setup done for the customer, an assistant inside the product, or a product that learns each user's routine.
- How to pitch it
- Have a first-time customer complete a key task, before and after.
- Success criteria
- Tasks completed faster, customers who would switch it on, and a product owner ready to build it.
Watch for AI that customers can't see or undo. They need to stay in control.
26. Every Idea Tested in a Week
How might we put every new product idea in front of real customers within a week, and decide faster what to build?
AI makes prototypes almost free, so the slow part is now deciding.
Read the briefHide the brief
- Who is affected
- Product teams with more ideas than evidence, and the leaders deciding what to fund.
- Preparation
- Three ideas waiting for a decision, and customers ready to test them.
- Where to focus
- Prototypes built with AI in a day, tests with real customers, and go or stop decisions on the evidence.
- How to pitch it
- Show one idea taken from concept to customer test during the event, and the decision it supports.
- Success criteria
- Decisions made on evidence, time and cost against today's process, and a product leader who adopts it.
Watch for Testing with colleagues. Only real customers count.
27. Selling the Result, Not the Product
How might we sell customers guaranteed outcomes instead of products they have to operate themselves?
AI and connected products let us measure results, so we can promise them.
Read the briefHide the brief
- Who is affected
- Customers who buy our product but struggle to get results from it, and the service teams helping them.
- Preparation
- The results customers buy our product for, what gets in the way, and three customers to react to an offer.
- Where to focus
- One product offered as an outcome service, pricing per result, and the risks we would take on.
- How to pitch it
- Show the service offer and its price against buying the product, and what customers said.
- Success criteria
- Customers who would switch, economics that work over the contract, and an owner to pilot it.
Watch for Risks we can't control. Be clear about what the customer still has to do.
Picked one? Turn it into your event brief with the Hackathon Brief Template.
Hackathon Ideas for Technology and IT
Invite IT, security, architects and people from the business who use the systems every day.
28. AI Agents on the Team
How might AI agents run an entire business process end to end, with people setting the goals and approving the decisions that matter?
AI is moving from assistants that answer to agents that act.
Read the briefHide the brief
- Who is affected
- Teams running repetitive multi-step workflows, and the managers accountable for them.
- Preparation
- One end-to-end workflow with its steps, systems and exceptions, and access to a safe test environment.
- Where to focus
- The workflow run by agents, the supervisor's role, and the controls that make it safe.
- How to pitch it
- Run real cases through the agents live, with the supervisor stepping in where needed.
- Success criteria
- Cases completed correctly, human time saved, and a process owner ready for a pilot.
Watch for Agents with more access than they need. Give each one only the permissions its task requires.
29. AI for Everyone Within 90 Days
How might we put secure, useful AI into every employee's daily work within 90 days?
People already use AI without permission, and the companies that enable everyone gain the most.
Read the briefHide the brief
- Who is affected
- Employees using unapproved tools or none, and the IT and security teams trying to keep up.
- Preparation
- What employees use today (ask them), the tools the company already licenses, and the security rules.
- Where to focus
- An approved toolkit for each role, a network of champions, and one role's day redesigned and measured.
- How to pitch it
- Show one team's week with its new toolkit, with the time saved and the risks closed.
- Success criteria
- Adoption in the pilot team, risks closed, and an IT leader ready to roll it out.
Watch for Tools without habits. Adoption comes from champions and examples, not licences.
30. Retiring the Systems We Fear to Touch
How might we modernise the legacy systems nobody fully understands, before a failure forces our hand?
AI can read old code and explain it, so modernising one piece at a time becomes affordable.
Read the briefHide the brief
- Who is affected
- The business relying on fragile systems, the few experts who still understand them, and the teams waiting for change.
- Preparation
- One legacy system, its code and documentation, and the expert who knows it best.
- Where to focus
- AI that documents the old system, migration in small slices, and tests written before anything changes.
- How to pitch it
- Show one piece of the system explained, tested and moved, with the expert's verdict.
- Success criteria
- Documentation the expert confirms, a safe path for the first migration step, and an IT owner for it.
Watch for Rewriting everything at once. One slice, moved safely, proves the path.
Picked one? Turn it into your event brief with the Hackathon Brief Template.
Industry-Specific Themes
The ideas above apply across industries. Here are additional themes tailored to specific sectors:
Financial Services
- Fraud pattern detection that adapts faster than fraudsters
- Personalised financial guidance at scale
- Regulatory change impact assessment automation
- Real-time risk visualisation for complex portfolios
Healthcare
- Patient journey friction reduction
- Clinical documentation burden relief
- Care coordination across providers
- Preventive intervention trigger identification
Manufacturing
- Predictive maintenance that actually prevents downtime
- Supply chain disruption response acceleration
- Quality prediction before defects occur
- Energy optimisation across production
Retail
- Inventory optimisation across channels
- Customer lifetime value prediction and action
- Return prevention through better matching
- Store associate augmentation for expertise
Telecommunications
- Network issue prediction before customer impact
- Churn intervention optimisation
- Self-service resolution for complex issues
- Personalised plan optimisation
Professional Services
- Proposal generation acceleration
- Knowledge reuse across engagements
- Resource allocation optimisation
- Client insight synthesis
From Ideas to Impact
A hackathon idea is just the starting point. The organisations that extract real value from hackathons have systems to capture, evaluate, and develop winning concepts.
Every hackathon idea that shows promise needs:
- A clear problem statement - one page documenting who has this problem, why it matters, and what success looks like
- A structured idea document - capturing the solution concept, key assumptions, and critical unknowns
- An evaluation framework - consistent criteria applied across all submissions
- A development pathway - resources and sponsorship to take winning ideas forward
- A compelling pitch deck - to communicate the opportunity to judges and stakeholders
Without this, promising concepts stall once the winners are announced. Feed every project into your opportunity pipeline, not only the winners, so the work stays discoverable and usable across the company.
Give the Pitch Its Own Time
I've judged dozens of hackathons, and the pattern is consistent: teams spend almost all their time building and leave the pitch to the last hour, so strong work reaches the judges undersold. Organisers can prevent this. Share the evaluation criteria at the kick-off, and set a point in the schedule when teams move from building to the pitch. Give every team the same pitch script, answering the same questions in the same order: the problem, who has it, the evidence of potential and the next step. Judges can then compare projects quickly and fairly.
For detailed guidance on turning hackathon concepts into products, including templates and frameworks for post-hackathon development, see my guide on product discovery documentation.
Selecting the Right Theme for Your Hackathon
Not every theme fits every organisation. Consider:
Strategic alignment: Does this theme connect to priorities leadership actually cares about? Hackathons that produce ideas nobody wants to fund are demoralising.
Balance across types: Running more than one theme? Aim for one of each innovation type: a process theme for the quick wins, a product theme for new offerings and a business model theme for the bigger bet.
Participant expertise: Do your teams have enough context to contribute meaningfully? A supply chain hackathon in a company where few understand supply chain will struggle.
Feasibility of prototyping: Can teams build something demonstrable in hackathon timeframes? Some problems need months of data access or infrastructure that can't be provisioned for a 48-hour event.
Data, tools and regulation: Have you settled what data teams may use, and in which approved AI tools? Check too whether the theme touches a use the EU AI Act treats as high-risk, such as recruitment or evaluating workers.
Evaluation clarity: Will judges be able to assess submissions meaningfully against the judging criteria? Vague themes produce submissions that are hard to compare.
Follow-through capacity: Do you have resources to develop winning concepts? Don't run hackathons on topics you have no intention of investing in.
Once you have chosen a theme, turn it into the event brief with the Hackathon Brief Template: the twenty sections that define a corporate hackathon, from the theme and sponsor to the funding path and how outcomes are tracked.
For a complete guide on planning and running corporate hackathons - including evaluation frameworks and post-hackathon processes - see my detailed operational guide.
Making Your Hackathon Work in the AI Era
AI has changed what's possible in hackathon timeframes. Teams can now prototype in hours what previously took days. Non-technical participants can contribute meaningfully with AI assistance. The question has shifted from "can we build it?" to "is it worth building?"
This makes problem selection more important than ever. When building is easy, the differentiator is identifying the right problem and framing a compelling solution.
The new hackathon toolkit:
- For ideation: ChatGPT, Claude - exploring problem spaces and solution approaches
- For building: GitHub Copilot, Cursor, Bolt.new - rapid prototyping
- For pitching: Ainna, the agentic innovation platform, for pitch decks and documentation
Teams that master this toolkit deliver more ambitious projects with more polished presentations - all within the same short window.
For strategic context on why hackathon frequency matters in 2026 and how AI tools transform the hackathon format, see my article on corporate hackathons in the AI era.
Related Resources
For hackathon teams:
- How to Win a Hackathon: A Practical Guide - tactical advice for participants
For hackathon organisers:
- Should We Still Organize Corporate Hackathons in the AI Era? - the strategic case for more frequent hackathons
- How to Plan and Run a Corporate Hackathon - the operational guide
- Corporate Hackathon Guide - comprehensive FAQ on hackathon design and execution
For innovation leaders:
- Product Discovery Documentation Guide - turning hackathon ideas into products
- Innovation Toolkit - templates for problem framing, idea documentation, and evaluation
Need Help Designing Your Hackathon?
Selecting the right theme, structuring the event, and building follow-through processes requires experience. As someone who has designed hackathon programmes at Microsoft, Accenture, and multiple startups, I can help you:
- Select themes aligned with strategic priorities
- Design evaluation criteria that surface genuinely valuable ideas
- Create follow-through processes that turn concepts into products
- Build internal capability to run effective hackathons independently
