Tim Norris-Wiles retweeted
Neuralift is now available in the AI you already use.
Our new MCP server brings your frontier customer models into Claude, Copilot Studio, agents and other compatible AI tools. Give your teams and agents access to learned discovery on your customer data, use cases and KPIs.
Neuralift MCP democratizes the intelligence in your first-party data, allowing you and your agents to explore behavioral patterns, customer affinities, differences between customers, discovered segments and ranked actions to lift your KPIs.
This is where learning matters. With Neuralift learning happens before the conversation. It anchors discoveries, analytics, benchmarks, values, stats and incrementality with evidence you can inspect and trustfully reason on.
“We tried out various LLMs on the MIT data warehouse and we got accuracy of zero. Not low—zero.”
Michael Stonebraker, Turing Award winner, MIT
Neuralift addresses a fundamental gap in marketing and advertising AI. Governed descriptions help an LLM or RAG system understand what a field means. The behavioral relationships across those fields still have to be learned.
Neuralift models ground customer data so that reasoning tools can see patterns across time for purchases, visits, engagement and loyalty - all tied to the use cases you want to solve and the KPIs you want to lift.
Ask questions in your own words. Compare segments. Explore valuable subsegments and lookalike seeds. Follow the evidence into the next question. The language model explains and reasons over findings computed from your customer data.
Which customers should we be looking for more of?
What distinguishes our most valuable customers?
Which audiences belong in our next acquisition campaign?
The 30-second video below shows that last example.
I've written a lot about AI working horizontally across the business. I'm excited because this is a practical step in that direction. Marketing, Analytics, CRM, leadership and agency partners can ask their own questions of the same customer model. People and agents can put that understanding to work wherever they need it. Shared customer intelligence, grounded in your warehouse data, with evidence your teams can inspect.
If you lead Marketing or Analytics at a consumer brand, message me with a customer question your team struggles to answer. We'll use it as the starting point for a Neuralift demo.
I laughed at first, and then realised just how depressing it is. Hopefully people are not actually putting children in front of this kind of service. Anyone know which company is behind it??
Excited to share the 2026 Words of the Year, AI edition:
1. Substrate
2. Scaffolding
3. Unpack
4. Honestly
5. Sharp
📰 BREAKING 📰: @ScuderiaFerrari lodge appeal in relation to @LewisHamilton’s podium finish in Malaysia’s Bahrain Grand Prix
Be very interesting to see how this plays out 🍿
#f1 #FormulaOne #ferrari #hamilton44
🤖 Made with AI
Tim Norris-Wiles retweeted
🔥 Our hype reel for @G2Eshows 💰
One thing about this space. The amount of data and analytics related to marketing is profound. Every transaction and every behavior is captured. This is exactly why we are having so much success here.
Add columns to a table and LLM predictions start to fall apart.
@LTMpredict found they drop to majority-guess levels by ~30 columns, while classic methods improve.
Customer data is FAR wider than this.
Research on what this means for marketing coming soon from @neuraliftoff
Have to say it was a big surprise to see Mr Tumble on #granddesigns this evening. And that he has a wife!
My Opus 5.5 initial thoughts:
1) It's quick. Feels a bit like 4.6 did which is high praise.
2) Less verbose in responses. Finally.
3) Prose generation is a massive improvement. Opus 5 was AWFUL at writing and it feels like this fixes it.
4) Efficiency has gone through the roof.
Interesting to see news of Harvey's margin collapse. On surface an incredible product + team, + getting to 15bn value is laudable in a number of ways. Just surprised to read that lawyers don't like using it + their UX sucks... I assumed UX / harness would be their moat 🤷♂️
Find out why the addition of a neural-net layer to your customer data stack is becoming the norm for performant brands. Download our white paper with SAS here.
neuralift.ai/white-papers/wa…
Tim Norris-Wiles retweeted
Too many AI personal assistants, too little time to assess all of them. Nice work here: assistantbenchmark.com
Tim Norris-Wiles retweeted
We've collaborated with @SASsoftware Customer Intelligence 360!
Discovery>Decision>Action is a closed-loop framework for Marketing AI that begins at the data warehouse, touches the customer in real-time and gets smarter over time.
Interested in SOA Marketing AI? Read this 👇
Imitation is the most sincere form of flattery, they say.
Nice to see Apple jump on the clay bandwagon with @neuraliftoff this week! 😏
Huge. Nemotron is going to get serious distribution control out of this, and it feels like the timing for N4 probably key.
techcrunch.com/2026/08/26/nv…
There's upside hiding in your data, it's just hard for marketers to find without help!
Using the same old tactics won't deliver anything new, and waiting for the data team to rescore the base means you're in Q4 already 😫
Thankfully there's another way: neuralift.ai
Get a lot of people asking me to explain "Lift Potential" mode in @neuraliftoff, so here goes.
Lift Potential works with transactional data in your warehouse to identify where to prioritise... It changes our segment discovery view to be ranked by $£€ upside, not population...
This matters b/c historically the way we approached marketing was to spend money first, read the results, and then optimize towards the best performing group or groups.
Lift Potential allows you to see where the likely growth lives BEFORE you have spent a penny 👀
You can even game out upside scenarios using the built-in modeller, which allows you to assess worst / good / best cases, with everything is based on your 1P data; no presuppositions, biases or guesswork.
Combines beautifully w/ our new Actions feature! nitter.gq/jonathanmendez/status/…
We've been working all summer on turning customer 360 tables in @Snowflake @databricks @GoogleCloudTech and CDPs into "Ranked Actions."
This builds off the outputs of our Neural Network Segmentation we do with @nvidia GPUs.
More to do but we are getting close to release.





















