Vera Field

Vera-field.com

We're specialized in Process & AI Automation for solar installation and maintenance companies.
Try our free 60-minute workflow assessment where we:

- Identify which manual processes are consuming 10+ hours of your team's time each week
- Show you which automation tools can eliminate each time-waster

AI & automation consultant 

I help product teams move from feature-driven to outcome-driven ways of working.

My methodology in a nutshell:
- Create internal alignment on target audience/ICP
- Align on business & and product goals (OKRs, hierarchy of metrics)
- Set up or improve idea discovery & validation processes (e.g. customer interviews, shadowing, protoyping, usability testing, A/B testing)

Interim Product Lead/ Advisor/ Coach

Wave Terminal

- Definition of ICP & beachhead segment
- Setting up & executing customer discovery and experimentation processes
- Defining a product strategy that helps Wave find a differentiated spot in the market

Interim Product Lead

HiveDesk

HiveDesk offers automatic employee time tracking software with project, task, and cost tracking. I'm advising the HiveDesk team on how to build a product that fits seamlessly to their customers needs. Setting up robust product analytics and continuous product discovery to help improve the HiveDesk onboarding flow - from 'aha' to 'habit' - and corresponding metrics.

Product Advisor

Rodeo Software

- Improving the Rodeo New User Experience (setup & activation)
- Helping the Product team move to outcome-driven ways of working
- Setting up & executing discovery & experimentation processes

Interim Sr. Product Manager

Datawallet

Head of Product

Bridgemaker GmbH

Senior Venture Architect

CrefoPayment GmbH & Co. KG

Head of Product and QA

Hungr

Operations Manager

payleven

Project Manager Operations

Team Leader Operations

Product Leader | I help SaaS startups build products that customers love | sonderhouse.io

Read my Idea Discovery & Validation guide here: https://eu1.hubs.ly/H0dfJkm0

What I do
I help B2B SaaS and service companies move from feature-driven to outcome-driven ways of working.
More recently, I also help teams become AI-fluent and build AI-driven products.

How I do it
- Align leadership on the #1 ICP, business, and product goals
- Set up or strengthen discovery & validation practices (continuous discovery, product analytics, rapid testing)
- Coach teams on integrating AI into product strategy and workflows

My belief
Great products come from:
- Sharp focus on the right target audience
- Continuous discovery and rapid validation
- Teams empowered to drive business outcomes, not just ship features

Background
15+ years in product and strategy roles across startups and scaleups in Berlin. I’ve seen successes—and just as many hard lessons.

Work with me
I work as a coach, advisor, or interim product lead. More of my thinking at elsevanderberg.substack.com

Dr. Else van der Berg

New adventure alert: I've joined SwitchUp as a Product Consultant 🍾 

To the outside world, Switchup is a service that monitors energy contracts in Germany - making sure customers always get a fair deal.

Internally, we're rebuilding the entire company as an AI-first subscription agent.

The goal isn't to automate human customer support out. It's to automate the noise around them, so they can focus on offering top-notch service where it's needed. 

𝗪𝗵𝘆 𝗱𝗶𝗱 𝗜 𝗷𝗼𝗶𝗻?

🔹  A small nimble team of very smart, curious, and humble people. I'm not just saying that to stroke their egos  Arik Meyer Maren Hinzmann Inna Makoterska

🔹  Very fun project, and the stuff of the future! Automation workflows & AI where it actually makes sense, building an internal tool where AI agents and humans can collaborate. Lots of space to play and learn. 

🔹 Arik Meyer puts his money where his mouth is. No "blah blah we're gonna change the world" - besides the for-profit business Switchup he founded a not-for-profit sister company that helps consumers that have been screwed over get justice. 

We're looking for M-shaped people with an engineering and/or product background to join un (Product engineer; Technical product manager; Data architect). Hit me up if you'd like to work with me!

Product management for AI-native startups │ Interim, advisor, coach

I built an AI agent that finds customer quotes for opportunities (pain points, wishes, desires). Here's how it works - and why I still manually compile the end-report:

𝗧𝗵𝗲 𝗽𝗿𝗼𝗯𝗹𝗲𝗺: Teams enthusiastically jump to solutions without really understanding the opportunity.

So before ideation sessions, I send participants a pre-read with:
👉 Opportunity description
👉 Evidence why it matters
👉 Direct quotes from customer interviews
In the pre-AI era, manually hunting through interview transcripts took hours. Thank god for LLMs.  I built an agent (using windsurf + Claude Code) that:

- Analyzes all interview transcripts for relevant quotes
- Optionally searches subreddits where your ICP hangs out

𝗖𝗿𝗶𝘁𝗶𝗰𝗮𝗹 𝗻𝗼𝘁𝗲𝘀:

1. Interview quotes >>> forum posts (you know the context and ICP fit). Forums are full of noise, BUT still nice to get a feel.

2. Your transcripts need proper structure/format for the agent to work!

3. I NEVER send AI reports directly to my team. The agent does heavy lifting. I verify every quote, check for hallucinations, and craft the final document.

Result: Better ideation built on genuine customer understanding, not assumptions.

AI amplifies research. Human judgment ensures quality. That's the balance.

PS. this video is cut after 5 minutes, find the full video on Youtube.
PPS. You can fork my agent from Github. I would suggest tailoring it so it can make sense of YOUR interview transcripts, and follows your own process. You can also just start with my prompt (Notion URL) and build your own

Yes, GPT 5.1 and Gemini 3 are AMAZING, I can ship good-looking stuff even faster now. But they're still not helping me decide *what* to build in the first place. 

It feels like the age-old problem of "getting value out of qualitative customer feedback scattered across sources" is still unsolved.

We've got interview transcripts, we've got sales calls in Gong, we've got survey responses in Pendo, etc. etc. but how can we make sense of it all?

Where do we centralize it and structure it in a way that LLMs can work with it? 

I'm imagining:

👉  Feed an opportunity (customer pain point, desire, wish) to the AI agent of my choosing. Output: every relevant customer quote, plus counter-evidence and competitive analysis; OR

👉 Have AI agent proactively surface patterns and opportunities across all qualitative data sources.

What am I missing? Who's solving this? What's the most difficult part?
Help me out here guys.

Specialists built the last generation of tech companies. Generalists will build the next.

Last week at @producttank someone asked the question "How do AI-native companies  build products differently - compared to "legacy" tech scale ups?"

The question often gets narrowed down to "will PMs survive?" or "are engineers replaceable?" That misses the point.

What's actually shifting is 𝗼𝗿𝗴𝗮𝗻𝗶𝘇𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲.

AI-native companies are smaller. Small and nimble is an asset.

They're hiring what I call natural convergers - M-shaped people who cross domains naturally, without friction. Engineers who ask "what evidence do we have that this matters?" PMs who learn technical concepts by prompting their way through problems.

When one person can run a full build-measure-learn cycle, you don't need handoff managers. You don't need as many layers.

👉  𝗧𝗵𝗲 𝗹𝗲𝗴𝗮𝗰𝘆 𝗺𝗼𝗱𝗲𝗹: specialists in silos, product trios trying to bridge gaps, feedback loops spanning multiple roles.

👉  𝗧𝗵𝗲 𝗔𝗜-𝗻𝗮𝘁𝗶𝘃𝗲 𝗺𝗼𝗱𝗲𝗹: generalists using AI to constantly upskill, zero handoffs, everyone spotting problems and shipping solutions.

𝗧𝗵𝗲 𝗿𝗲𝘀𝘂𝗹𝘁? 
1. More eyes on problem discovery, more diverse ideas, therefore better ideas.
2. Much faster build-measure-learn feedback cycles. Learning velocity is the moat. 

I joined SwitchUp a month ago, and I love it. We have a small team of people of convergent thinkers who can't stay in their lane, love to try new things, share learnings, and solve problems.  I'm looking for more of those!

Are you a:
👉  𝗣𝗿𝗼𝗱𝘂𝗰𝘁 𝗲𝗻𝗴𝗶𝗻𝗲𝗲𝗿/𝗽𝗿𝗼𝗱𝘂𝗰𝘁-𝗺𝗶𝗻𝗱𝗲𝗱 𝗳𝘂𝗹𝗹-𝘀𝘁𝗮𝗰𝗸 𝗲𝗻𝗴𝗶𝗻𝗲𝗲𝗿 (you'd be owning a problem,  without a PM by your side)
👉 𝗧𝗲𝗰𝗵𝗻𝗶𝗰𝗮𝗹 𝗣𝗠 (having owned a technical product before, e.g. API or payment gateway; loving BPMN/state machines/workflow orchestration)
👉 𝗗𝗮𝘁𝗮 𝗮𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁

Hit me up!

I built an ideation agent, and you can fork it.
If you throw a problem at an LLM and let it ideate from scratch, you cut off your own creativity. You anchor on its output. You stop thinking.

So I built an agent that forces ME to go first.

It gathers context. It asks what "good" looks like (weighted criteria). It makes me input my own ideas before the LLM generates anything.

The result: a sparring partner that helps me get unstuck - not an outsourced brain.

I love vibe coding these agents with Claude Code. It forces me to think through and lay down my process. It's codified "Else".

Walkthrough in the video. GitHub repo + initial prompt linked in comments - fork it, make it yours.

I've spent years preaching "validate before you build."

Now I'm working on a product where we basically just… build things and see what happens.

𝗧𝗵𝗲 𝗼𝗹𝗱 𝗮𝗿𝗴𝘂𝗺𝗲𝗻𝘁𝘀 𝗳𝗼𝗿 "𝘃𝗮𝗹𝗶𝗱𝗮𝘁𝗲 𝗳𝗶𝗿𝘀𝘁" 𝗵𝗮𝘃𝗲 𝘄𝗲𝗮𝗸𝗲𝗻𝗲𝗱.

👉  𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝘄𝗮𝘀 𝗲𝘅𝗽𝗲𝗻𝘀𝗶𝘃𝗲. Not anymore. Features that took months ship in a week with AI-assisted coding.

👉 𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗺𝗲𝗮𝗻𝘁 𝘀𝗹𝗼𝘄 𝗳𝗲𝗲𝗱𝗯𝗮𝗰𝗸 𝗹𝗼𝗼𝗽𝘀. Building is much, much faster thanks to robust AI assisted coding workflows (don't worry, it's not me vibe coding this ;) )

👉 𝗙𝗲𝗮𝘁𝘂𝗿𝗲 𝗯𝗹𝗼𝗮𝘁 𝗺𝗮𝗱𝗲 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝘀 𝘂𝗻𝘂𝘀𝗮𝗯𝗹𝗲. With SaaS products, users had to dig through menus and toggles. But Wave has an integrated AI agent - you just tell it what you want. The surface area problem is different now.

At the same time, pre-build validation hasn't gotten us far. 
Sure, it has helped us get inside users' heads, and make sure we don't ship anything completely incomprehensible/unusable, but it hasn't helped us *validate*. 

I've been running 1:1 shadowing sessions (where new users try Wave), and landing page tests (comparing a variety of different value props and ideas). 
These have been useful for "do people understand this" , but not for "𝘄𝗵𝗮𝘁'𝘀 𝘁𝗵𝗲 𝗺𝗼𝘀𝘁 𝗺𝗮𝗴𝗶𝗰𝗮𝗹 𝗶𝗱𝗲𝗮."

Our ideas about 'how we can frame Wave' change weekly because the landscape moves so fast. No stable baseline to compare against, small sample sizes -> no patterns. 

And with LLM-powered products, so much is in the execution. Does the AI actually deliver? You can't know until you build it.

The most reliable signal is shipping. Put it in the world. Watch what happens.

Join me at the live Product Engineers podcast today at 17.00 where I'll be chatting with my friend Peppe Silletti about product/dev roles converging.

Engineers are getting involved in product decisions that would have traditionally belonged to PMs. Product managers are shipping code. Designers are ditching Figma for vibe-coded prototypes that ship faster than polished mockups ever could.

Everyone's getting accused of shiny new object syndrome. 
Everyone stepping on each other's toes.

But what about cognitive load? Can one person really be an expert on multiple domains (e.g. both Product AND Engineering), or are we creating "masters of none"?

What has AI changed?

Let's dive in!

PS Peppe - thanks for the artwork. 

Posting the link in the comment section.

"𝘛𝘩𝘦 [𝘱𝘳𝘰𝘥𝘶𝘤𝘵 𝘥𝘦𝘷𝘦𝘭𝘰𝘱𝘮𝘦𝘯𝘵] 𝘸𝘰𝘳𝘬 𝘪𝘴𝘯’𝘵 𝘤𝘰𝘮𝘱𝘭𝘦𝘹 - 𝘵𝘩𝘦 𝘱𝘳𝘰𝘤𝘦𝘴𝘴 𝘩𝘢𝘴 𝘮𝘢𝘥𝘦 𝘪𝘵 𝘤𝘰𝘮𝘱𝘭𝘦𝘹. 𝘗𝘳𝘰𝘥𝘶𝘤𝘵 𝘥𝘦𝘷𝘦𝘭𝘰𝘱𝘮𝘦𝘯𝘵 𝘪𝘵𝘴𝘦𝘭𝘧 𝘪𝘴 𝘴𝘪𝘮𝘱𝘭𝘦: 𝘳𝘦𝘴𝘦𝘢𝘳𝘤𝘩 𝘢 𝘱𝘳𝘰𝘣𝘭𝘦𝘮, 𝘥𝘦𝘴𝘪𝘨𝘯 𝘪𝘵, 𝘤𝘰𝘥𝘦 𝘪𝘵, 𝘭𝘢𝘶𝘯𝘤𝘩 𝘪𝘵, 𝘢𝘯𝘥 𝘪𝘵𝘦𝘳𝘢𝘵𝘦. 𝘉𝘶𝘵 𝘰𝘷𝘦𝘳 𝘵𝘪𝘮𝘦, 𝘤𝘰𝘮𝘱𝘢𝘯𝘪𝘦𝘴 𝘵𝘶𝘳𝘯𝘦𝘥 𝘦𝘢𝘤𝘩 𝘴𝘵𝘦𝘱 𝘪𝘯𝘵𝘰 𝘥𝘰𝘻𝘦𝘯𝘴 𝘰𝘧 𝘴𝘶𝘣-𝘴𝘵𝘦𝘱𝘴 𝘳𝘦𝘲𝘶𝘪𝘳𝘪𝘯𝘨 𝘮𝘶𝘭𝘵𝘪𝘱𝘭𝘦 𝘵𝘦𝘢𝘮𝘴, 𝘳𝘦𝘷𝘪𝘦𝘸𝘴, 𝘢𝘯𝘥 𝘧𝘶𝘯𝘤𝘵𝘪𝘰𝘯𝘴. 𝘞𝘩𝘢𝘵 𝘶𝘴𝘦𝘥 𝘵𝘰 𝘳𝘦𝘲𝘶𝘪𝘳𝘦 𝘰𝘯𝘦 𝘣𝘶𝘪𝘭𝘥𝘦𝘳 𝘯𝘰𝘸 𝘵𝘢𝘬𝘦𝘴 𝘴𝘪𝘹 𝘮𝘰𝘯𝘵𝘩𝘴 𝘢𝘯𝘥 𝟣𝟢 𝘵𝘰 𝟣𝟧 𝘥𝘪𝘧𝘧𝘦𝘳𝘦𝘯𝘵 𝘵𝘦𝘢𝘮𝘴 𝘫𝘶𝘴𝘵 𝘵𝘰 𝘴𝘩𝘪𝘱 𝘢 𝘴𝘮𝘢𝘭𝘭 𝘧𝘦𝘢𝘵𝘶𝘳𝘦."

Just had a listen to Lenny Rachitsky's conversation with Tomer Cohen, CPO @ LinkedIn (link to podcast in comments). Tomer explains his new "Full Stack Builder" program, that "teaches coding, design, and PM skills together." The idea is to enable anyone to single-handedly go from idea to launch. 

Very much in line with the role convergence I've been seeing, especially in AI-native companies (think Lovable, Perplexity, Linear, Cursor, and Elevenlabs) (see link to my latest Substack article "Why Your Best Engineers Have Product Sense and Your Best PMs Ship Code" in the comment).

Companies that want to *drive* - not follow - change in the AI era are making drastic org. chart changes to kill the "handover tax". Flat hierarchies. Small teams. M-shaped people that can identify and tackle problems single-handedly.

Everyone's trying to buy word of mouth right now.

I get it. AI-native tools need an extremely powerful WOM machine just to get picked up once by early adopters. Eeeveryone is suffering from AI tool fatigue. 

People will only give you a try if you've been recommended by MULTIPLE people they trust.
So naturally, everyone's trying to buy their way in. 

As a "mini influencer" I get the pitch regularly: cash and "advisory council" spots to plug products.

I don't think it works.

A couple of paid influencer posts aren't enough to create the omnipresence an AI startup needs to cut through the noise.

I say this from experience.  I've taken the "advisory council for a post" offer twice - only for products I genuinely believed in. Both posts performed terribly. I'm probably not doing it again. It costs authenticity. Not worth it.

But when I write about Vistaly or Claude Code? Those posts fly. Because I'm a genuine power user sharing how I actually work. They're chockfull of"in the trenches" authenticity.

So if you can't buy WOM, what's the answer?

𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝘀𝗼𝗺𝗲𝘁𝗵𝗶𝗻𝗴 𝗽𝗲𝗼𝗽𝗹𝗲 𝗴𝗲𝗻𝘂𝗶𝗻𝗲𝗹𝘆 𝘄𝗮𝗻𝘁 𝘁𝗼 𝘁𝗮𝗹𝗸 𝗮𝗯𝗼𝘂𝘁.

We're trying hard Wave Terminal. Last week a fan posted about us in China. 155K views. Got picked up by the GitHub community. 60K more views. A wave of installs.

We couldn't have bought that. It only happened because someone found it remarkable enough to share unprompted.

The toughest part: "remarkable" keeps shifting. AI capabilities upgrade constantly. Today's magic is tomorrow's baseline.

So you ship fast. You find the wow before the window closes.

We're going with: 𝗦𝗽𝗲𝗲𝗱 𝗮𝗻𝗱 𝗰𝗼𝗻𝘃𝗶𝗰𝘁𝗶𝗼𝗻. 

(PS I described my experiences GTM & Product discovery at Wave Terminal in my latest Substack article, link in comment)

taplio