AI Consulting Slovenia: How to Choose an Agency AI svetovanje Slovenija: kako izbrati agencijo

If you are evaluating ai consulting slovenia, the best choice is not the agency with the flashiest demo or the smartest prompt engineer. It is the partner that can connect AI to your real processes, data, systems, compliance requirements, and measurable business outcomes. In 2026, Slovenian companies should expect an AI consulting agency to do far more than build a chatbot. A credible partner should assess workflows, integrate with ERP/CRM/accounting tools, define ROI, deploy securely, and support adoption across the business.

That matters because AI spending is moving from experimentation to operations. Gartner forecast that worldwide generative AI spending would reach $644 billion in 2025, up 76.4% from 2024 Gartner, 2025. At the same time, McKinsey reported that 65% of organizations were regularly using generative AI in at least one business function by 2024 McKinsey, The State of AI, 2024. The implication for Slovenian SMBs is clear: the question is no longer whether to use AI, but how to choose an agency that can implement it properly.

For many companies, the strongest AI partner is one that combines consulting with delivery. That means strategy, workflow design, automation, custom assistants, deployment, training, and ongoing optimization. This is the practical space where M-AI operates, helping organizations move from ideas to usable AI systems rather than isolated experiments.

What AI consulting in Slovenia should include in 2026

In 2026, AI consulting should be treated like digital transformation with a stronger operational and data layer. A good Slovenian AI agency should not start by asking, “Do you want a chatbot?” It should start by asking where time is lost, where decisions are delayed, where data is fragmented, and where staff repeatedly perform manual work that software could support or automate.

A modern AI consulting engagement should usually include:

• Discovery and use-case prioritization: mapping business processes and identifying where AI creates real value.
• Data and systems review: understanding the quality, location, permissions, and structure of your data.
• Integration planning: connecting AI to existing tools such as CRM, ERP, document systems, support platforms, and finance tools.
• Security and compliance design: especially important for personal data, internal knowledge bases, and regulated workflows.
• Prototype and pilot delivery: proving value quickly with a scoped deployment.
• Production deployment: moving from test environment to live use with monitoring and controls.
• Training and adoption support: ensuring managers and teams actually use the solution.
• Measurement: tracking cost savings, cycle-time reduction, lead quality, response speed, or other agreed KPIs.

This broader scope is not optional. According to IBM, 42% of enterprise-scale companies had already actively deployed AI in their business by 2023, while another 40% were exploring or experimenting with it IBM Global AI Adoption Index, 2023. As adoption rises, the difference between useful AI and expensive AI is implementation quality.

For Slovenian SMBs, practical examples may include AI-powered internal knowledge assistants, invoice/document workflows, customer-service copilots, sales support, and tax-related process automation. If your needs involve tax and accounting automation, for example, a specialized tool such as FURS AI may be more valuable than a generic chatbot. If your challenge is product discovery or retail intelligence, a vertical product such as Shelfze may solve a narrower business problem faster than a custom build. A capable AI consultant should be able to advise when to build, when to buy, and when to combine both.

“There is no AI strategy without data strategy.”

This widely repeated principle remains true because most AI projects fail not at the model layer, but at the workflow, governance, and data-access layer. An agency that ignores that is selling demos, not transformation.

7 criteria for choosing an AI agency beyond ChatGPT prompts

Many agencies now market themselves as AI experts because they can create prompts, wrappers, or simple assistants. That may be useful, but it is not enough for a company that needs stable, secure, scalable solutions. Here are seven criteria that matter more.

1. Business process understanding

The agency should understand how your company works before suggesting tools. Can they map your sales funnel, support flow, procurement process, or finance operations? Can they identify bottlenecks and estimate where AI can save hours or reduce errors? Good AI consulting begins with operations, not technology theater.

2. Integration capability

A solution is only as useful as its connection to your existing stack. Ask whether the agency can integrate with Microsoft 365, Google Workspace, HubSpot, Salesforce, custom CRM systems, accounting software, document repositories, e-commerce tools, APIs, and internal databases. AI that lives outside your workflow usually gets ignored.

3. Data governance and privacy maturity

Especially in Slovenia and the wider EU, data handling must be taken seriously. Your agency should clearly explain where data is processed, whether it is stored, how access is controlled, what model providers are used, and how GDPR obligations are considered. If they cannot explain the architecture simply, that is a warning sign.

4. ROI orientation

AI projects should have measurable outcomes. The agency should help define success before implementation: fewer support tickets handled by humans, faster report generation, shorter sales response time, lower admin workload, higher conversion rates, or reduced compliance risk. Deloitte found that 74% of surveyed organizations said their most advanced generative AI initiative was meeting or exceeding ROI expectations Deloitte, State of Generative AI in the Enterprise, 2024. The key word is initiative, not experiment.

5. Delivery beyond prototype

Some providers are good at pilots but weak at deployment. Ask for examples of systems they have taken into production. What monitoring do they implement? How do they handle prompt/version changes, fallback logic, user permissions, and model updates? Production readiness separates consulting firms from AI hobbyists.

6. Change management and training

Even strong technical solutions can fail if employees do not trust or use them. A serious AI partner includes onboarding, documentation, workshops, and feedback loops. This is particularly important for Slovenian SMBs where teams are lean and every new system must prove itself quickly.

7. Ability to recommend the right path, not the biggest project

The right partner should sometimes advise against custom development. In some cases, a lightweight automation, a retrieval-based assistant, or a vertical tool will deliver value faster and cheaper. Agencies that always recommend a large custom build may be optimizing for project size, not your outcome.

“Most AI value doesn’t come from the model alone. It comes from redesigning workflows so people and machines each do what they do best.”

That is exactly why an agency should be judged on systems thinking, not just prompt fluency.

Questions to ask about integrations, data, ROI, and deployment

If you want to compare providers objectively, ask the same detailed questions in every meeting. This quickly reveals who understands implementation and who only understands AI marketing.

Questions about integrations

• Which of our current systems can you integrate with directly?
• Do you build API-based integrations, middleware automations, or custom connectors?
• Can the solution write back into our CRM, ERP, or document system, not just read from it?
• How do you handle authentication, permissions, and role-based access?
• What happens if one connected system changes or goes offline?

Questions about data

• What data do you need from us to launch a pilot?
• How do you assess data quality and knowledge-source reliability?
• Will our data be used to train third-party models?
• Where is data stored and processed?
• How do you support GDPR-conscious deployments and internal access controls?

Questions about ROI

• What business KPIs should we track before and after implementation?
• Can you estimate expected time savings or revenue impact by use case?
• How long does a typical pilot take to show measurable value?
• What costs should we expect after launch, including model usage, maintenance, and support?
• Can you show examples where your work improved efficiency or reduced workload?

Questions about deployment

• What is your process from discovery to production?
• How do you test accuracy, reliability, and hallucination risk?
• What fallback workflows exist when AI is uncertain?
• Who owns the solution, prompts, logic, and documentation after delivery?
• Do you provide monitoring, iteration, and post-launch support?

The strongest agencies answer these questions directly, in business language, with examples. They do not hide behind jargon. They can usually show a roadmap with phases, responsibilities, estimated effort, and decision points.

If you are speaking with M-AI, these are exactly the kinds of topics worth discussing early: process fit, system integration, practical automation opportunities, and deployment paths that make sense for a Slovenian business environment.

How Slovenian SMBs can shortlist the right AI partner

For Slovenian SMBs, the biggest mistake is trying to evaluate AI agencies only on technical sophistication. The better approach is to shortlist firms based on business relevance, delivery credibility, and fit with your internal capabilities.

Here is a practical four-step shortlist process.

1. Start with one high-value use case

Do not begin with “we need AI.” Begin with a specific problem, such as reducing time spent answering repetitive customer questions, extracting information from documents, supporting FURS-related administrative work, or helping staff find internal knowledge faster. A narrow starting point makes agency evaluation much easier.

2. Look for local and regional context

Slovenian businesses often need support that reflects EU compliance expectations, multilingual environments, local processes, and pragmatic budgets. The right partner should understand that implementation is not just about model performance; it is about making AI useful inside a real company with existing systems and constraints.

3. Ask for relevant examples, not generic case studies

A polished case study about a global brand may tell you very little. Instead, ask for examples of workflow automation, knowledge assistants, document processing, or internal AI tools that resemble your needs. If an agency has built specialized solutions, such as AI support for FURS-related tasks or vertical commerce tools like Shelfze, that may signal practical product thinking in addition to consulting capability.

4. Compare proposals using the same scorecard

Create a shortlist scorecard with categories such as:

• Understanding of our business problem
• Integration capability
• Data/privacy clarity
• Pilot plan and timeline
• Expected ROI logic
• Support and maintenance model
• Total cost over 12 months
• Communication quality

This keeps the decision grounded. It also helps avoid choosing an agency simply because it uses the most impressive technical language.

One more practical point: ask whether the agency can work in stages. For most SMBs, the ideal path is discovery, pilot, deployment, then scaling. That reduces risk and improves learning. It also allows your company to build internal confidence before expanding AI into more sensitive or complex processes.

Common mistakes to avoid when selecting an AI consultancy

• Choosing based on hype: great demos do not guarantee business value.
• Ignoring integrations: standalone AI tools often fail to stick.
• Skipping ROI definitions: if success is vague, disappointment is predictable.
• Underestimating data readiness: weak source material leads to weak outputs.
• Forgetting adoption: employees need training, trust, and clear workflows.
• Assuming all AI providers are equal: strategy firms, automation studios, and product builders offer very different strengths.

The best outcome usually comes from a partner that can combine consulting discipline with delivery capability. That means they can identify the right use case, build or integrate the right solution, and support your team until the system creates measurable value.

Final thoughts: choose the agency that can operationalize AI

The right answer to ai consulting slovenia is simple: choose an agency that can operationalize AI inside your business, not just demonstrate it. In 2026, a credible AI consultant should understand workflows, data, compliance, integrations, deployment, training, and ROI. If they cannot connect those pieces, they are unlikely to deliver lasting value.

For Slovenian SMBs, a smart selection process starts small, asks hard questions, and prioritizes implementation over hype. Whether you need internal AI assistants, process automation, specialized tools, or a roadmap for broader AI adoption, the right partner should help you move from concept to measurable result.

Talk to M-AI about your AI roadmap

If you want to explore how AI can improve operations, automate repetitive work, or support smarter decision-making in your business, talk to M-AI. We can help assess use cases, define a realistic pilot, and recommend whether a custom solution, targeted automation, or a specialized tool is the best fit.

Ready to shortlist the right AI partner or discuss your first implementation? Visit /#contact and start the conversation.

Kratek odgovor: če iščete ai consulting slovenia, ne izbirajte agencije po tem, kako dobro zna napisati ChatGPT poziv, ampak po tem, ali zna AI povezati z vašimi procesi, podatki, sistemi in merljivimi poslovnimi cilji. Dobra AI agencija v Sloveniji v letu 2026 mora znati več kot demonstrirati orodje: mora razumeti integracije, varnost, skladnost, avtomatizacijo, uvedbo v prakso in donosnost naložbe. Prava izbira je partner, ki vam pomaga od ideje do delujoče rešitve v produkciji.

Za slovenska mala in srednja podjetja to pomeni predvsem eno: ne kupujete “AI-ja”, ampak poslovni rezultat. Če želite hitrejšo obdelavo dokumentov, boljšo podporo strankam, manj ročnega dela v financah ali boljše iskanje po internih znanjih, mora agencija pokazati, kako bo to dosegla v vašem konkretnem okolju. Pri tem so pomembni podatki, ERP/CRM povezave, uporabniška izkušnja, nadzor nad stroški in jasen načrt uvedbe.

Po podatkih Eurostata je leta 2024 AI tehnologije uporabljalo 13,5 % podjetij v EU z vsaj 10 zaposlenimi, kar kaže na hitro rast, a tudi na dejstvo, da je večina podjetij še vedno v fazi izbire pravih primerov uporabe in partnerjev Eurostat, Use of artificial intelligence in enterprises, 2025.

Ravno zato je pri izbiri partnerja za ai consulting slovenia smiselno gledati širšo sliko: ali agencija zna svetovati, zgraditi rešitev, jo povezati z vašimi sistemi in jo tudi dejansko uvesti med zaposlene. V podjetju M-AI se prav ta razlika pogosto pokaže kot ključna: podjetja ne potrebujejo še enega “AI eksperimenta”, ampak uporabno rešitev, ki prihrani čas, zmanjša napake in odpre nove prodajne ali operativne možnosti.

What AI consulting in Slovenia should include in 2026

AI svetovanje v Sloveniji v letu 2026 mora biti bistveno bolj operativno in poslovno usmerjeno kot v prvem valu navdušenja nad generativno umetno inteligenco. Dober ponudnik ne prodaja generičnega “AI workshopa”, ampak strukturiran pristop.

Takšno svetovanje bi moralo vključevati vsaj naslednje elemente:

• identifikacijo primerov uporabe, ki imajo dejanski poslovni učinek,
• oceno pripravljenosti podatkov in informacijskih sistemov,
• izbiro arhitekture med orodji, modeli in integracijami,
• izdelavo prototipa ali pilotnega projekta,
• merjenje ROI in opredelitev KPI-jev,
• uvajanje v produkcijo z vzdrževanjem,
• upravljanje tveganj, vključno z varnostjo, pravnimi vprašanji in kakovostjo odgovorov.

McKinsey ugotavlja, da organizacije že poročajo o merljivih učinkih generativne AI v več poslovnih funkcijah, vendar največjo razliko ustvarjajo podjetja, ki AI vgradijo v procese, ne pa tista, ki ostanejo pri izoliranih eksperimentih McKinsey, The state of AI, 2024. To je tudi najboljša definicija kakovostnega AI svetovanja: od eksperimenta do procesa.

Za slovenska podjetja je dodatno pomembno, da svetovalec razume lokalni kontekst. To vključuje delo z dokumenti v slovenščini, davčnimi in administrativnimi tokovi, internimi bazami znanja ter pogostimi sistemi, ki jih podjetja že uporabljajo. Primer dobre prakse je avtomatizacija dokumentnih tokov in finančnih opravil, kjer so specializirane rešitve pogosto bolj smiselne kot generični chatboti. Če podjetje potrebuje pomoč pri davčnih ali administrativnih procesih, je lahko relevanten tudi specializiran produkt, kot je FURS AI pomočnik, medtem ko je za produktivnost in delo z digitalnimi policami oziroma organizacijo vsebin lahko smiseln pogled na Shelfze.

“There is no AI strategy without a data strategy.”

Ta pogosto citirana industrijska ugotovitev je za izbiro agencije zelo uporabna: če vam ponudnik govori samo o modelih, ne pa tudi o podatkih, dostopih, kakovosti virov in governance, gledate nepopolno sliko.

7 criteria for choosing an AI agency beyond ChatGPT prompts

Spodaj je sedem kriterijev, po katerih lahko precej bolje ocenite agencijo kot pa po všečni predstavitvi ali obljubi, da vam bodo “implementirali AI”.

1. Razumevanje poslovnega problema, ne samo orodja

Najprej preverite, ali agencija zna prevesti vašo težavo v merljiv primer uporabe. Dober partner bo vprašal, kje nastaja izguba časa, kje so ozka grla, koliko stane ročno delo, kje prihaja do napak in kako se trenutno meri uspeh. Slab partner bo preskočil poslovni kontekst in šel naravnost v demo.

2. Sposobnost integracij z vašimi sistemi

Večina vrednosti nastane šele, ko je AI povezan z realnimi viri podatkov in delovnimi tokovi. Zato vprašajte po izkušnjah z ERP, CRM, dokumentnimi sistemi, e-pošto, bazami znanja, SharePointom, računovodskimi orodji in internimi API-ji. Če agencija ne obvlada integracij, bo vaš AI ostal izoliran pripomoček.

3. Pristop k podatkom, varnosti in skladnosti

Podatki niso samo tehnično vprašanje. So poslovno, pravno in reputacijsko vprašanje. IBM poroča, da je povprečni globalni strošek kršitve varnosti podatkov leta 2024 dosegel 4,88 milijona USD, kar je rekordna vrednost IBM, Cost of a Data Breach Report 2024. Čeprav to ni slovenski podatek, dobro pokaže, zakaj mora AI partner resno obravnavati dostopne pravice, hrambo podatkov, anonimizacijo, revizijske sledi in model governance.

4. Jasna metodologija od pilota do produkcije

Veliko projektov obtiči v pilotni fazi. Vprašajte, kako agencija vodi projekt od delavnice do produkcijskega okolja. Ali definirajo uspeh pilota? Ali imajo načrt za testiranje, spremljanje kakovosti odgovorov, fallback mehanizme in podporo po uvedbi? Prava agencija vas ne pusti pri “proof of concept” brez naslednjih korakov.

5. Merjenje ROI in poslovnih učinkov

Če ni načina merjenja, ni resne poslovne odločitve. Deloitte ugotavlja, da se številne organizacije pri generativni AI premikajo iz faze eksperimentiranja v fazo osredotočenosti na merljivo vrednost Deloitte, State of Generative AI in the Enterprise, 2024. Zato mora agencija znati opredeliti prihranjene ure, skrajšanje časa obdelave, zmanjšanje napak, povečanje konverzij ali večjo odzivnost podpore.

6. Sposobnost prilagoditve slovenskemu jeziku in lokalnim procesom

To je v Sloveniji pogosto podcenjen kriterij. Veliko rešitev izgleda dobro v angleščini, precej slabše pa deluje pri slovenskih dokumentih, internem žargonu, računovodskih postopkih ali pravnih obrazcih. Agencija mora znati pokazati, kako bo rešitev delovala v realnem slovenskem okolju, ne samo v demo primeru.

7. Prenos znanja na vašo ekipo

Dober AI partner vas ne želi narediti trajno odvisne od sebe za vsako malenkost. Del kakovostnega svetovanja so tudi izobraževanje, dokumentacija, governance smernice in usposabljanje ključnih uporabnikov. To je posebej pomembno za SMB podjetja, kjer ista ekipa pogosto skrbi za procese, IT in operativo.

“Most AI projects fail not because the models are weak, but because deployment, change management, and business alignment are weak.”

Ta misel dobro povzame realnost trga: tehnologija sama po sebi ni več glavna ovira. Ovira je izvedba.

Questions to ask about integrations, data, ROI, and deployment

Ko se pogovarjate z agencijo, ne ostanite pri splošnih vprašanjih. Spodaj so vprašanja, ki hitro razkrijejo, ali je partner resen.

Vprašanja o integracijah

• Katere sisteme ste že povezovali: ERP, CRM, e-pošta, DMS, baze znanja?
• Ali znate delati prek API-jev, webhookov in varnih konektorjev?
• Kako boste rešili dostopne pravice in ločili občutljive podatke?
• Kaj se zgodi, če zunanji sistem ni na voljo?

Vprašanja o podatkih

• Katere podatke potrebujemo za uspešen pilot?
• Kako ocenite kakovost in popolnost podatkov?
• Ali se naši podatki uporabljajo za treniranje tujih modelov?
• Kako je urejena hramba, anonimizacija in revizijska sled?

Vprašanja o ROI

• Kateri KPI-ji so smiselni za naš primer uporabe?
• Kako hitro lahko pričakujemo prvi merljiv učinek?
• Kako boste ločili dejanski poslovni učinek od “wow efekta” pri demu?
• Kakšen je okvirni TCO: razvoj, licence, uporaba modelov, vzdrževanje?

Vprašanja o uvedbi in produkciji

• Kaj je vaš tipičen načrt od discovery faze do produkcije?
• Kdo pri nas mora sodelovati in koliko časa bo to zahtevalo?
• Kako spremljate kakovost odgovorov po uvedbi?
• Kaj vključuje podpora po lansiranju?

Če agencija na ta vprašanja odgovarja konkretno, z realnimi primeri in jasno metodologijo, ste verjetno na dobri poti. Če dobite predvsem splošne odgovore o “inovativnosti”, “prompt engineeringu” in “AI revoluciji”, brez omembe sistemov, podatkov in merjenja učinkov, bodite previdni.

Gartner je v zadnjih letih večkrat izpostavil, da uspeh digitalnih in AI pobud ni odvisen le od tehnologije, temveč tudi od sprememb procesov, upravljanja in sprejetja pri uporabnikih Gartner research summaries on AI adoption and governance, 2024. To pomeni, da mora biti dober svetovalec hkrati tehničen, poslovno usmerjen in dovolj pragmatičen za uvedbo v realno okolje.

How Slovenian SMBs can shortlist the right AI partner

Za slovenska mala in srednja podjetja je najpametnejši pristop kratek, strukturiran izborni postopek. Ni treba pripravljati dolgega razpisa, je pa smiselno imeti jasna merila.

1. Najprej opredelite 1 do 3 prioritetne primere uporabe

Namesto splošnega vprašanja “kaj lahko AI naredi za nas?” raje pripravite konkretne izzive. Na primer: avtomatska obdelava vhodnih dokumentov, AI pomočnik za podporo strankam, iskanje po internih pravilnikih, pomoč pri pripravi prodajnih ponudb ali avtomatizacija administracije.

2. Pripravite kratek opis obstoječega okolja

Zapišite, katere sisteme uporabljate, kje so podatki, kdo so uporabniki in kje so največje omejitve. Tako boste hitreje dobili primerljive predloge različnih ponudnikov.

3. Izberite 3 do 5 agencij in jih ocenite po enotnih kriterijih

Uporabite preprosto ocenjevalno tabelo: razumevanje problema, integracije, podatki in varnost, ROI, hitrost izvedbe, reference, podpora po uvedbi. Tako se izognete temu, da zmaga najbolj všečen nastop namesto najboljšega partnerja.

4. Zahtevajte predlog pilotnega projekta

Ne zahtevajte samo ponudbe. Zahtevajte predlog pilota s cilji, KPI-ji, časovnico, potrebnimi podatki in oceno tveganj. To hitro pokaže, kdo razume izvedbo.

5. Preverite reference in podobne primere

Ni nujno, da je agencija delala točno v vaši panogi, mora pa znati pokazati podobno kompleksnost: povezovanje sistemov, delo z dokumenti, interno znanje, večjezičnost, uvedba med zaposlene. Če ponudnik omenja konkretne produkte in rešitve, je to pogosto dober znak. Na primer, podjetje, ki poleg svetovanja razvija tudi specializirane AI produkte, kot so rešitve na M-AI, pogosto bolje razume razliko med prototipom in produkcijskim sistemom.

6. Ocenite kulturni fit in način sodelovanja

Za SMB podjetja je to pomembnejše, kot se zdi. Potrebujete partnerja, ki komunicira jasno, ne komplicira po nepotrebnem in zna predlagati pragmatične korake. AI projekt ni samo tehnična implementacija; je sodelovanje med vodstvom, operativo in pogosto tudi zunanjimi partnerji.

7. Začnite dovolj majhno, da bo projekt izvedljiv, a dovolj veliko, da bo učinek viden

Najboljši prvi projekt ni nujno največji. Običajno je najbolj smiseln primer uporabe z veliko ponavljajočega se dela, dobro dostopnimi podatki in jasno metriko uspeha. Tako lahko hitro preverite vrednost in se nato širite naprej.

Prav tu lahko dober partner naredi največjo razliko. Namesto splošnega svetovanja pomaga določiti realistično pot: od hitre analize priložnosti do pilota, nato do produkcije in širitve na druge procese. To je pristop, ki ga slovenska podjetja vse bolj potrebujejo, ker na trgu ni več glavno vprašanje, ali AI deluje, ampak kje deluje najbolje in kako ga uvesti brez nepotrebnega tveganja.

Zaključek: izberite partnerja, ki bo dostavil rezultat, ne le demonstracije

Če povzamemo: najboljša izbira za ai consulting slovenia ni agencija z najbolj glasnim marketingom, ampak tista, ki zna povezati strategijo, podatke, integracije, avtomatizacijo in produkcijsko uvedbo. Leta 2026 bo razlika med uspešnimi in neuspešnimi AI projekti še manj tehnološka in še bolj izvedbena.

Za slovenska SMB podjetja je prava pot jasna: izberite konkreten primer uporabe, preverite pripravljenost podatkov in sistemov, postavite KPI-je in izberite partnerja, ki zna pokazati, kako bo rešitev delovala v vašem okolju. Če iščete pragmatičen pristop, ki združuje svetovanje, razvoj in realne poslovne primere uporabe, si oglejte storitve na m-ai.info ter preverite tudi specializirane rešitve, kot sta FURS AI in Shelfze.

CTA: se želite pogovoriti o AI priložnostih v vašem podjetju?

Če želite oceniti, kateri AI primer uporabe ima za vaše podjetje največji potencial, in dobiti konkreten predlog za pilot ali implementacijo, stopite v stik z ekipo M-AI. Na uvodnem pogovoru lahko hitro razjasnite, kaj je smiselno avtomatizirati, kako pristopiti k integracijam in kako meriti ROI.

Kontaktirajte M-AI tukaj: https://m-ai.info/#contact