Community demand board
FixMyProblem
Real problems worth solving — from startups and small businesses across the EU, posted by the people who have them. Add yours, or click "Me too" on the ones you share.
Missed call, lost customer
I work alone. When I'm with a client, I can't pick up the phone — and the caller just books the next company on Google.
Missed call, lost customer
I work alone. When I'm with a client, I can't pick up the phone — and the caller just books the next company on Google.
What people want
An AI that answers every call, books the appointment into my calendar, and texts me a two-line summary.
Sources
- How are solo business owners handling missed calls?
- Has anyone tried automated phone answering services?
- How are small businesses handling missed calls and slow follow-ups?
Hlido's take
◐ Agents exist, but quality lags
Phone-answering agents exist, but the best we've independently tested scores 58/100. Demand is ahead of quality here.
Independently reviewed: klariqo 58/100 · usercall 40/100
300 receipts every month
Each month I photograph a pile of receipts and retype them into my accounting tool, line by line. It eats a full weekend.
300 receipts every month
Each month I photograph a pile of receipts and retype them into my accounting tool, line by line. It eats a full weekend.
What people want
An agent that reads receipts and invoices, books every transaction into the right category, reconciles monthly, and asks me only about the odd ones.
Sources
- Tell HN: AI coding is sexy, but accounting is the real low-hanging target
- Bookkeeping in 2025 feels stuck in the stone age — why?
- Where are the real AI tools in accounting?
Hlido's take
⬤ Open gap — nobody has built this yet
Open gap — among 845 independently reviewed agents, no credible bookkeeping agent yet.
My bot answers nothing
I installed a support chatbot for my shop. It sends customers help-center links but can't check their order — so 8 of 10 chats still end up with me.
My bot answers nothing
I installed a support chatbot for my shop. It sends customers help-center links but can't check their order — so 8 of 10 chats still end up with me.
What people want
A bot that actually reads the customer's order and history, solves the real problem, and hands over to me fast when it can't.
Sources
- Ask HN: Will low quality AI customer support be the new normal?
- Tell HN: Stripe seems to have suspended human customer support
Hlido's take
✓ An answer exists
Good ones exist — two support agents scored 90/100 in our independent testing. Most deployed bots are just badly chosen.
Independently reviewed: mavenagi 90/100 · sitegpt 90/100
Quotes die silently
I send offers, get busy, and forget to follow up. Weeks later I learn the customer went elsewhere because nobody chased the quote.
Quotes die silently
I send offers, get busy, and forget to follow up. Weeks later I learn the customer went elsewhere because nobody chased the quote.
What people want
An agent that watches my sent quotes and politely follows up every few days until there's a yes, a no, or a booked call.
Sources
- Looking for an AI that runs the entire sales workflow automatically
- How are you using AI for emails and follow ups?
Hlido's take
✓ An answer exists
Partial solutions rated 78–80/100 in our testing; the full hands-off follow-up flow is still stitched together by hand.
Independently reviewed: parakeeet-ai 80/100 · smartlead 78/100
Posts sound like AI
I tried every AI writing tool for our company LinkedIn. They all produce the same plastic voice — my customers can tell in one sentence.
Posts sound like AI
I tried every AI writing tool for our company LinkedIn. They all produce the same plastic voice — my customers can tell in one sentence.
What people want
A writer that learns from my last 100 posts and our real company facts, and drafts in MY voice — not template text.
Sources
- I've tried every AI content tool and they all have the same problem
- Am I the only one who thinks AI writing tools are way too expensive?
Hlido's take
✓ An answer exists
Top writing tools score 90/100 on capability in our tests — but 'sounds like me' (voice fidelity) is exactly the part nobody measures. We should.
Independently reviewed: jasper 90/100 · anyword 90/100
€35,000 surprise AI bill
One developer's coding agent silently skipped prompt caching — the cloud bill came to $38,000. Nothing warned him while it was happening.
€35,000 surprise AI bill
One developer's coding agent silently skipped prompt caching — the cloud bill came to $38,000. Nothing warned him while it was happening.
What people want
A hard spending cap per agent: track every model call live, show who spent what, and stop the workflow before the bill explodes.
Sources
- $38k AWS Bedrock bill caused by a simple prompt caching miss
- Which AI agent spent the money on your OpenAI/Anthropic bill
- Ask HN: How do you reduce token usage with Cline/Continue?
Hlido's take
✓ An answer exists
A spend-guard agent scored 90/100 in our independent testing — this problem is more solved than people know.
Independently reviewed: llm-spend-guard 90/100
Same correction, every day
I correct my coding assistant's mistakes every session — and next morning it makes the same ones again. It forgets everything as my project grows.
Same correction, every day
I correct my coding assistant's mistakes every session — and next morning it makes the same ones again. It forgets everything as my project grows.
What people want
Durable project memory: my rules and past corrections applied automatically in every new session.
Sources
- Ask HN: Best AI IDE right now? ('a lot of forgetting as my codebase gets larger')
- Ask HN: Is there an AI bot that works like a literate programming build step
Hlido's take
◐ Agents exist, but quality lags
Memory layers exist (one rated 90/100), but no coding agent ships durable project memory out of the box — a half-open gap.
Independently reviewed: mem-0-ai 90/100
Agent quietly got worse
Our production agent's answers degraded for two weeks before anyone noticed. Customers noticed first — that's the problem.
Agent quietly got worse
Our production agent's answers degraded for two weeks before anyone noticed. Customers noticed first — that's the problem.
What people want
A watcher that continuously tests my agent's outputs and alerts me the moment quality drifts — before customers feel it.
Sources
- Ask HN: How do you know when your AI agent's output quality degrades?
- Ask HN: How are you monitoring AI agents in production?
- Ask HN: What tools are you using for AI evals? Everything feels half-baked
Hlido's take
✓ An answer exists
Two eval/monitoring platforms scored 90/100 in our testing — the tooling matured faster than word got around.
Independently reviewed: braintrust 90/100 · helicone 90/100
Agent wants my passwords
To automate real work, every agent asks for full access to my email and admin accounts. One prompt injection and it can do anything I can.
Agent wants my passwords
To automate real work, every agent asks for full access to my email and admin accounts. One prompt injection and it can do anything I can.
What people want
Per-task, temporary, logged access — and a human confirmation before anything risky.
Sources
- Ask HN: How do you give AI agents access without over-permissioning?
- Ask HN: How do you authorize AI agent actions in production?
Hlido's take
⬤ Open gap — nobody has built this yet
Open gap — no reviewed agent solves scoped, auditable agent permissions yet.
Every demo looks great
Every agent demo is impressive; production is a different story. Before paying, I can't see any proof of which agents actually work.
Every demo looks great
Every agent demo is impressive; production is a different story. Before paying, I can't see any proof of which agents actually work.
What people want
Independent testing with published evidence — scores I can check before the invoice, readable by me and by my own agents.
Sources
- Why is there no 'App Store' for independent AI agents yet?
- Ask HN: Is there an AI subscription comparison tool that's always updated?
Hlido's take
✓ An answer exists
This one is why Hlido exists — 845 agents independently tested, every claim checked, evidence public.
Independently reviewed: Hlido — 845 agents independently tested
AI Act file, no lawyer
We sell an AI screening tool into the EU. By August 2026 we need a risk file, technical docs and logging — we're 4 people and none of us is a lawyer.
AI Act file, no lawyer
We sell an AI screening tool into the EU. By August 2026 we need a risk file, technical docs and logging — we're 4 people and none of us is a lawyer.
What people want
An agent that reads our code and docs, drafts the required technical documentation and risk file, and updates them on every release.
Sources
- Ask HN: How are you handling EU AI Act compliance as a developer?
- It's too hard for SMEs to comply with the EU AI Act
- The Hidden Cost of AI Regulations (ACT survey)
Hlido's take
⬤ Open gap — nobody has built this yet
Open gap — no reviewed agent drafts and maintains AI Act technical documentation.
Am I 'high-risk'?
Is our chatbot 'high-risk' under the AI Act? Nobody can tell me without €10k of legal advice — and a wrong guess means fines or lost deals.
Am I 'high-risk'?
Is our chatbot 'high-risk' under the AI Act? Nobody can tell me without €10k of legal advice — and a wrong guess means fines or lost deals.
What people want
An agent that interviews me about my use case, maps it to the AI Act risk categories with citations, and outputs the obligation checklist.
Sources
- EU AI Act: Practical risk classification for business AI use cases
- EU AI Act compliance is becoming a sales blocker for AI SaaS
- Small Businesses' Guide to the AI Act
Hlido's take
⬤ Open gap — nobody has built this yet
Open gap — it's already a sales blocker for AI SaaS, and no agent does cited risk classification.
US servers block deals
A client's lawyer asked where our AI tools store data. Checking every vendor's GDPR terms took me four days — and we lost two weeks on the deal.
US servers block deals
A client's lawyer asked where our AI tools store data. Checking every vendor's GDPR terms took me four days — and we lost two weeks on the deal.
What people want
An agent that audits every AI tool in my stack for data location, retention and DPA terms, and recommends EU-compliant setups or alternatives.
Sources
- Introducing data residency in Europe (OpenAI)
- EU Data Residency for AI Apps 2026
- GDPR-Compliant AI: On-Premise & EU LLMs vs ChatGPT
Hlido's take
⬤ Open gap — nobody has built this yet
Open gap — nobody audits AI tools for EU data residency per vendor. (We're considering an EU-data flag on every Hlido scorecard.)
My invoices become illegal
My German customers now demand XRechnung e-invoices; France and Poland mandates hit in 2026. My invoicing tool only makes PDFs.
My invoices become illegal
My German customers now demand XRechnung e-invoices; France and Poland mandates hit in 2026. My invoicing tool only makes PDFs.
What people want
An agent that converts my invoices to XRechnung/Factur-X/KSeF formats, validates them, and files them into my existing bookkeeping.
Sources
- eInvoicing in Germany (European Commission)
- Most businesses are not ready for France's September e-invoicing deadline
- Poland confirms KSeF e-invoicing mandatory from 1 Feb 2026
Hlido's take
⬤ Open gap — nobody has built this yet
Open gap — three national mandates arriving, zero reviewed agents handling EN 16931 invoice conversion.
11 countries, 11 VAT rates
My web shop sells to 11 EU countries. Every order needs that country's VAT rate, and every quarter an OSS return — I do it in spreadsheets and pray.
11 countries, 11 VAT rates
My web shop sells to 11 EU countries. Every order needs that country's VAT rate, and every quarter an OSS return — I do it in spreadsheets and pray.
What people want
An agent that pulls orders from my shop, applies the right destination VAT to each, and prepares the quarterly OSS return ready to file.
Sources
- EU VAT One Stop Shop (OSS)
- VAT-OSS explained: common pain points
- New approach to VAT for e-commerce imports (European Commission)
Hlido's take
⬤ Open gap — nobody has built this yet
Open gap — no reviewed agent prepares OSS/IOSS returns from shop data.
Czech customer, English bot
Our customers write in Czech, German and Polish. Our 3-person team answers in English — and watches half of them never reply again.
Czech customer, English bot
Our customers write in Czech, German and Polish. Our 3-person team answers in English — and watches half of them never reply again.
What people want
An agent that answers tickets natively in every EU language and escalates the sensitive ones to us with a translated summary.
Sources
- The ecommerce guide to multilingual customer support
- Best multilingual customer support tools for ecommerce
Hlido's take
⬤ Open gap — nobody has built this yet
Open gap — 0 of 845 reviewed agents specialize in multilingual EU support. The single market speaks 24 languages; the agents don't.
Found the tender too late
A city tender we could have won was published in another country's language on TED. We found it three days after the deadline.
Found the tender too late
A city tender we could have won was published in another country's language on TED. We found it three days after the deadline.
What people want
An agent that watches TED and national portals for tenders matching my profile, translates the notice, and pre-fills the bid paperwork.
Sources
- How to facilitate SME participation in public procurement (SMEunited)
- Access to public procurement — Single Market Scoreboard
Hlido's take
⬤ Open gap — nobody has built this yet
Open gap — only 14% of EU tenders go to foreign suppliers; no reviewed agent does cross-border tender watching.
One hire, new rulebook
I want to hire one developer in Spain. The EOR quote is 12% of her salary, forever — just so someone else knows the local payroll rules.
One hire, new rulebook
I want to hire one developer in Spain. The EOR quote is 12% of her salary, forever — just so someone else knows the local payroll rules.
What people want
An agent that drafts the country-compliant contract, computes real employer cost, and tracks that country's filings and law changes.
Sources
- Navigating European employment compliance in remote-first startups
- Employer of Record in Europe: 2026 country guide
Hlido's take
⬤ Open gap — nobody has built this yet
Open gap — a 5-country team means 5 compliance regimes; no reviewed agent covers EU employment compliance.
24 hours to report
Our 60-person logistics firm now falls under NIS2. We have no security team, but incidents must be reported within 24 hours.
24 hours to report
Our 60-person logistics firm now falls under NIS2. We have no security team, but incidents must be reported within 24 hours.
What people want
An agent that gap-checks us against our country's NIS2 rules, generates the policies and risk register, and drafts the 24-hour report when something happens.
Sources
- Easing the NIS2 burden: targeted reforms (Morrison Foerster)
- How much does NIS2 compliance really cost?
Hlido's take
⬤ Open gap — nobody has built this yet
Open gap — SME compliance costs run €50k–300k and no reviewed agent addresses NIS2.
Third ESG form this month
Three big customers sent three different sustainability questionnaires this quarter. Same data about my company — three formats, three deadlines.
Third ESG form this month
Three big customers sent three different sustainability questionnaires this quarter. Same data about my company — three formats, three deadlines.
What people want
An agent that keeps one master data pack about my company and auto-fills every customer's questionnaire from it, flagging what's missing.
Sources
- Commission presents voluntary sustainability standard to ease SME burden
- CSRD: requesting data from SME suppliers (BCLP)
Hlido's take
⬤ Open gap — nobody has built this yet
Open gap — CSRD cascades onto SME suppliers and no reviewed agent automates the questionnaire flood.
What slows your business down?
Write it like you'd tell a friend. It appears on the board instantly — our review desk adds category and evidence afterwards.
Where demand meets supply
Problems posted here, next to how many agents Hlido has independently reviewed in that area. A long bar with thin supply = an open opportunity.
Seeded 2026-07-12 from public research — every problem cites its sources. Agents can read this board at /fixmyproblem/data.json. Hlido never takes payment to rank problems or reviews.