# Rafael Soares

Rafael Soares writes about data, artificial intelligence, and digital commerce, with a focus on how businesses can use technology to improve customer experiences and decision-making.

- Profile: https://b2binfluencer.marketing/creators/rafael-soares/
- LinkedIn: https://www.linkedin.com/in/rafael-bsoares
- Topics: artificial-intelligence, ai-adoption, customer-experience, data-quality
- Based on 15 dated posts, latest 2026-09-17

## Summary

Rafael Soares is most useful for explaining how data quality and AI adoption affect business operations.

- Best for: A Brazilian B2B data, governance or martech vendor helping enterprise buyers before an AI rollout.
- Not for: A consumer brand or global product launch seeking reach, fixed deliverables and usage rights.
- Check first: Confirm whether their posts are explanatory or commercial offers before contacting them.

### Takeaways

- AI projects stall when data is scattered across folders, servers, clouds, email and other sources.
- The CIO's job includes selling AI to employees who fear being replaced.
- AI assistants still need empathy and humour, and automated service needs good judgment.

## Main ideas

- **AI projects stall on scattered, unreliable data**: Soares argues that data quality is foundational to artificial intelligence, not a preliminary chore. Information is dispersed across folders, servers, separate clouds, printed documents, email and social media, making data and AI projects hard to assemble into something coherent. Many companies, they add, lack an internal technology team, time or both, while other emergencies take priority. The post ends with an offer for their team to take on that complexity.
- **The CIO’s real job is selling AI to employees**: Responding to a CIO article, Soares says AI and structured data are already changing business and that the shift is one-way. The hard part is cultural: employees have a legitimate fear of being replaced. They recommend that CIOs explain, educate and train staff, showing how AI can make routines more efficient and possibly less stressful. Companies, they argue, should clarify that the tools are not meant to replace human decisions but to improve competitiveness and results.
- **Ad overload drives buyers away, so companies should structure a CDP**: Soares cites an Accenture finding that 74% of consumers may abandon online purchases because of too many ads. They present a Customer Data Platform as the answer, saying it uses machine learning and AI to map consumer and competitor patterns, support decisions, automate and analyse campaign performance, and adjust strategies in real time. The post explains the concept and ends with an invitation to discuss structuring a CDP. The 74% figure comes from the cited research, not Soares’s own measurement.
- **AI assistants still lack empathy and humour**: Soares says companies are investing in making AI attendants and assistants feel more natural, but that something essential is missing: the connection created by empathy, sympathy and humour in service relationships. The design challenge is to keep humour empathetic and adaptable to whoever is asking while balancing it with helpfulness. The post presents this as an unresolved problem.
- **Implementation failure is a process lesson**: Soares describes a CrowdStrike update that caused Microsoft blue-screen failures, cancelling or delaying about 28,000 flights worldwide and causing losses in other sectors. They argue that teams must run all planned tests and confirm every implementation stage before launching data, AI or e-commerce projects. The post also says it is worth following how the companies involved and affected relate to each other.
- **Commerce and AI are read through events, partners and entrants**: Several posts are brief market observations rather than arguments. Soares announces they will attend VTEX Day with the Synerise team and names the stand’s partners: Dunnhumby on personalisation and loyalty, Infracommerce on applied intelligence at scale, and the VTEX Ads Platform on smarter recommendations, ads, search and interactions. A separate post says Temu is entering Brazil with aggressive marketing, lower prices and local brand partnerships, intensifying competition with Mercado Livre, Americanas, Shopee, Shein and Amazon. Another cites IDC data showing that the corporate applications market grew 12% in the previous year, linking demand to digital transformation and efficiency. These posts show the commercial environment and market shifts Soares tracks; unlike the data-quality and CDP posts, they do not end with a service offer.

## Work to read

- [Data quality as the foundation of AI work](https://www.linkedin.com/posts/rafael-bsoares_falta-de-dados-pode-ser-um-problema-para-activity-7192140548407775233-jUiz)
- [The CIO's job becomes selling AI to employees](https://www.linkedin.com/posts/rafael-bsoares_ia-inovaaexaeto-transformaaexaetodigital-activity-7191059133616304128-Niss)
- [Ad overload, abandoned carts and the case for a CDP](https://www.linkedin.com/posts/rafael-bsoares_excesso-de-an%C3%BAncios-faz-consumidores-desistirem-activity-7193952860747771905-ZVOa)
- [Empathy and humour in AI customer service](https://www.linkedin.com/posts/rafael-bsoares_desafio-do-humor-empresas-de-tecnologia-activity-7214592535212523520-ZZoT)
- [Synerise at VTEX Day, with named partners on the stand](https://www.linkedin.com/posts/rafael-bsoares_synerise-vtex-vtexday-activity-7450894598132023296-Q_Dy)
- [A CrowdStrike outage read as a process lesson](https://www.linkedin.com/posts/rafael-bsoares_crowdstrike-telaazul-microsoft-activity-7221491440890073088-KGzg)

## Brand relationships

- **Synerise** (disclosed_partnership)

## Partnerships

The posts do not document a partnership programme for Rafael Soares: they include no media kit, rate card, sponsored-content disclosure, or published deliverable and usage terms. They do show a working pattern around vendors, events, and direct business conversations. In an April 2026 post, they described attending VTEX Day with the Synerise team, showing client results from behavioural and predictive AI, and naming Dunnhumby, Infracommerce, and the VTEX Ads Platform as partners at that stand. Several other posts invite readers to talk directly about data quality, CDP setup, or AI adoption. Those are business-development conversations, not sponsorship packages; the posts state no payment, exclusivity, or licensing conditions for any collaboration. Brands should ask whether they accept sponsorships or paid work, and what a collaboration would include and cost.

## Alternatives

- Ocaoa Amorim: A better choice when the vendor's product serves logistics and supply-chain operations, and the buyer's task is getting operations leaders to fix processes before buying systems. The posts make that argument head-on: companies spend millions on technology before correcting their processes, and the choice between WMS, TMS and DMS should follow the operation's main pain. They back it with a plain breakdown of what each system does, who it suits and which indicators to watch (OTIF, logistics cost per order, returns rate, vehicle occupancy). A second post on cross-docking offers a repeatable format: definition, benefit list, KPIs and named users such as Amazon, Mercado Livre and Walmart. The delivery is operational and structured, with an explicit warning that no system repairs a broken process on its own.
- Ana Luiza Paes Araújo das Neves: A better choice when the vendor sells supplier quality, inspection or procurement software to industrial buyers, and the collaboration needs to argue that quality belongs in the sourcing decision, not in pre-audit paperwork. The post starts with a concrete buyer question: the supplier delivered on time and at the negotiated price, but the material arrived out of specification; was it still a good purchase? It answers with a five-step sequence (qualify, specify, measure, handle deviations, develop) and notes that OTIF alone does not cover rejections, non-conformities, repeat failures, rework or corrective-action effectiveness. That gives a quality-management or supplier-risk vendor a usable campaign spine, and the post closes by asking readers where quality sits in their own process.
- Fernanda Oi: A better choice when the need is to evaluate AI tools and research methods, and the vendor wants a practitioner to show how a tool performs rather than explain a category. The post compares OpenAI's Deep Research with Gemini firsthand on the same prompt, shows the full prompt, spells out the criteria (period respected, source quality, analytical depth, cross-report synthesis) and names the shortcomings: Gemini stalls at times and produces text that runs long. It closes by arguing that the assumption ChatGPT is always the best option is worth revisiting. A vendor selling a research or AI tool to marketing and UX teams gets a documented test and an honest verdict, which is harder to commission than a category explainer.

---
AI-assisted summary of public posts. Audience demographics are not available. Engagement appears only where we checked recent posts. Method: https://b2binfluencer.marketing/methodology/